Volume 124, Number 4. Jul/Aug 2021 MIT Technology Review

Citation preview

The change issue Volume 124 Number 4

Jul/Aug 2021

USD CAD

$9.99 $10.99

Technology, trends, and the new rules of business

JOIN

At EmTech MIT, immerse yourself in bold thinking and visionary innovation on the most significant technologies and global trends that will propel business and shape our world.

US

LIVE

ONLINE

September 28-30, 2021

S

MEET

T E

MIT Technology Review’s flagship event on emerging technology and trends

H

PL

U

Innovators Under 35 OF

AT

EMTECH

MIT,

WE’LL

The widespread effect of intersecting technologies such as AI, IoT, biotech, and cybersecurity

SUBSCRIBERS

202

1

EXPLORE:

The global impacts of technology consolidation, misinformation, energy, and climate

SAVE

10%

WITH

CODE

The new rules of business for a world that mandates corporate social responsibility

PRINTJA21

EmTechMIT.com/RegisterNow

AT

02

From the editor

The power of us

SIMON SIMARD

T

hough we’ve called it the “Change” issue, this edition of the magazine is really about two things: reflection and empowerment. For far too many of us, the pandemic has been a study in feeling powerless, and we’ve had little time to reflect as we focus on keeping ourselves and our loved ones safe, employed, and as mentally sound as possible. We’ve been forced to cope almost constantly with the twisting, morphing uncertainties that life has thrown at us. And yet in this unprecedented environment incredible stories of hope and empowerment have emerged. We see people finding ways to respond to suffering and injustice with positive change. Take the stories Abby Ohlheiser has collected (page 9), including those of Carlisa Johnson, who turned a Google Doc into a nexus of power for the Black Lives Matter movement, and Fiona Lowenstein, who nurtured an online community of thousands into a place where those suffering from covid-19 can get vital information. Sarah Jaffe writes (page 74) that a failed vote to unionize Amazon workers at a facility in Alabama may be discouraging, but around the US, workers in the increasingly expansive tech sector are waking up to their power to organize, and to demand dignity. In an essay on the arc of progress (page Michael Reilly dangerously misleading, if for no other is executive 16), Sheila Jasanoff harks back to West reason than that they can be interpreted editor of to justify individualism at all costs. In the Bengal in India, where she was born, and MIT Technology Review tells how under British rule the region’s US, this attitude has been corrosive to supthriving industry of woven textiles was port for government funding of important crushed by the Industrial Revolution. The high-tech industries like chip fabrication, lesson isn’t that technological advancement is bad—it’s that we which, as Jeremy Hsu writes, is one reason America is racing to must take care not to assume that all such change is for the best, catch up to manufacturers overseas (page 54). We have similar or that it comes without costs. work to do in the rapidly evolving field of clean energy, where— As Jasanoff writes, the good news is that we are not bystanders as Gernot Wagner writes—the price of solar panels has tumbled in the process. We are the ones who create technology, after all; over the last few decades (page 82). With a bit of a boost from we have the power to choose what gets built and how it is used. further R&D funding and favorable policies, solar stands a real Nowhere is this agency on fuller display than in this year’s list chance of helping decarbonize the planet. Reflection can lead to positive change, but it needs to come of 35 Innovators Under 35 (page 23). I hope you’ll take time to sit with this list. I find it impossible not to come away inspired with empowerment. As Karen Hao notes in her feature (page by their accomplishments—from swarms of French-toast size 48), a group of marginalized AI workers suffered indignities satellites to new research into fusion power to a pair of budding for years in the white-male-dominated field before finding one companies racing to bring optical computing to market. These other. With that sense of community, they realized they had the innovators are literally creating the future before our eyes. power to challenge the biggest companies in the world to be As we know, each of them stands on the achievements of better and more inclusive. If nothing else, I hope when you’re through with this issue, those who have come before. And yet the tech world is replete with narratives about single-minded mavericks bucking ortho- you’ll have given yourself time to reflect, feel your own power, doxy to realize their vision of the future. Those stories can be and just maybe find that you’re a little bit changed.

“First Republic took the time to get to know me — that real human connection is everything.” BLAIR HOLBROOK

Operations Strategy

(855) 886-4824

| Ŕrstrepublic.com | New York Stock Exchange symbol: FRC MEMBER FDIC AND EQUAL HOUSING LENDER

04

Contents

THE CHANGE ISSUE Report

The list

Features

Review

9

23

48

74

Activists’ online toolbox

35 Innovators under 35

The fight to reclaim AI

These organizers used apps and social media to energize others during the pandemic. By Abby Ohlheiser

Our annual list of 35 rising young innovators is an opportunity to take a look not just at where technology is now but where it’s going and who’s taking it there.

How a small group of researchers at the margins built a movement to wrest the world’s most powerful emerging technology from Big Tech’s control. By Karen Hao

Books, culture, and policy in perspective

14

The era of autonomy Zoox CEO Aicha Evans talks about leading the self-drivingcar company and the power of opportunity. By Anthony Green 16

Four arguments Sheila Jasanoff discusses the dangerous appeal of tech. Joe Garcia asks how prisoners can reenter a world transformed by digital tech. Jennifer Neda John reflects on the ways online influencers can mislead.

+ + + + +

Inventors . . . . . . . . . . . . . . . . . . .24 Humanitarians . . . . . . . . . . . 30 Visionaries . . . . . . . . . . . . . . . . .34 Pioneers . . . . . . . . . . . . . . . . . . . .38 Entrepreneurs . . . . . . . . . . . . .44

54

The great chip divide A shortage of microchips is threatening to slow digital innovation inspired by the promise of ever faster, cheaper computing power. By Jeremy Hsu

Sarah Jaffe on how to have a labor movement in the age of Big Tech. James Surowiecki on the task facing modern-day trustbusters. Gernot Wagner on what cheap solar power means for climate change. Wudan Yan on the toll of online harassment.

The back page

60

88

Those who fall behind get beaten up

The change chronicles

Can science transcend nationalism? By Yangyang Cheng 66

Kiara Royer ponders whether online activism can really lead to change.

Fast time A series of photographs taken around the world show how climate change is warping geological time. Photos by Ian van Coller Cover illustration by Sophy Hollington

But wait, there’s more. Lots more. You’re already a subscriber. Activate your account and start enjoying:

technologyreview.com/subonly

• Unlimited web access • Exclusive digital stories • The Algorithm newsletter • Access to 120+ years of publication archives

06

Masthead

Editorial

Corporate

Consumer marketing

Executive editor

Chief executive officer and publisher

Michael Reilly

Elizabeth Bramson-Boudreau

Senior vice president, marketing and consumer revenue

Doreen Adger

Editor at large

David Rotman News editor

Niall Firth Managing editor

Timothy Maher Commissioning editors

Bobbie Johnson Konstantin Kakaes Amy Nordrum

Assistant to the CEO

Katie McLean Human resources manager

James Wall Manager of information technology

Director of digital marketing

Emily Baillieu Director of event marketing

Brenda Noiseux

MIT Technology Review Insights and international Vice president, Insights and international

Nicola Crepaldi Director of custom content, US

Laurel Ruma Senior project manager

Martha Leibs

Colby Wheeler

Growth marketing manager Em Okrepkie

Content manager

Office manager

Assistant consumer marketing manager

Senior manager of licensing

Linda Cardinal

Caroline da Cunha Circulation and print production manager

Jason Sparapani Ted Hu Director of custom content, international

Senior editor, MIT News

Product development

Alice Dragoon

Chief technology officer

Email marketing manager

Senior editor, biomedicine

Drake Martinet

Tuong-Chau Cai

Antonio Regalado

Director of software engineering

Senior editor, climate and energy

Molly Frey

Advertising sales

James Temple

Board of directors

Head of product

Senior editor, digital culture

Mariya Sitnova

Vice president, sales and brand partnerships

Martin A. Schmidt, Chair Peter J. Caruso II, Esq. Whitney Espich Jerome I. Friedman David Schmittlein Glen Shor Alan Spoon

Abby Ohlheiser

Senior project manager

Senior editor, cybersecurity

Allison Chase

Patrick Howell O’Neill

Software engineer

Senior editors, AI

Jack Burns

Karen Hao Will Douglas Heaven Senior editor, computing

Siobhan Roberts Senior editor, podcasts and live journalism

Events Senior vice president, events and strategic partnerships

Amy Lammers

Jennifer Strong

Director of event content and experiences

Podcast producer

Brian Bryson

Anthony Green Editor, Pandemic Technology Project

Lindsay Muscato Senior reporters

Tanya Basu (humans and technology) Eileen Guo (technology policy and ethics) Reporters

Charlotte Jee (news) Neel Patel (space) Tate Ryan-Mosley (data and audio)

Senior director, brand partnerships

MIT Records (alums only)

617-253-8270

Nicole Silva

Ian Keller [email protected] 203-858-3396

Event operations manager

Senior director, brand partnerships

Elana Wilner

[email protected] 877-652-5295

Miles Weiner [email protected] 917-275-2048

Licensing and permissions

Event content producer

Erin Underwood Associate director of events

Events associate

Bo Richardson

Finance director

Madison Umina

Enejda Xheblati

Proofreader

General ledger manager

Barbara Wallraff

Olivia Male

Photo editor

Stephanie Arnett

877-479-6505

Web

Kristen Kavanaugh

Audience engagement associate

Kyle Thomas Hemingway

National

[email protected]

Finance

Marketing and events designer

Executive director, brand partnerships

Customer service and subscription inquiries

Debbie Hanley [email protected] 214-282-2727

Senior event content producer

Abby Ivory-Ganja

Emily Luong

Marii Sebahar [email protected] 415-416-9140

Senior director, brand partnerships

Engagement editor

Art director

Executive director, brand partnerships

Marcy Rizzo

Madeleine Frasca

Eric Mongeon

Caitlin Bergmann [email protected]

Marcus Ulvne

Head of international and custom events

Linda Lowenthal

Chief creative officer

Executive director, integrated marketing

Director of business development, Asia

International

Event partnership coordinator

Design

Andrew Hendler [email protected] 646-520-6981

Francesca Fanshawe

Kristin Ingram [email protected] 415-509-1910

Copy chief

Social media editor Benji Rosen

Tim Borton

Accountant

Anduela Tabaku

847-559-7313 Email

www.technologyreview.com/ customerservice

Reprints

[email protected]

Digital sales strategy manager

Casey Sullivan [email protected] 617-475-8066 Media kit

www.technologyreview.com/media

MIT Technology Review One Main Street 13th Floor Cambridge, MA 02142 617-475-8000 The mission of MIT Technology Review is to make technology a greater force for good by bringing about better-informed, more conscious technology decisions through authoritative, influential, and trustworthy journalism. Technology Review, Inc., is an independent nonprofit 501(c)(3) corporation wholly owned by MIT; the views expressed in our publications and at our events are not always shared by the Institute.

CANCER DOESN’T STOP. NEITHER DO WE. FOR ONE NIGHT. WE STAND TOGETHER.

SATURDAY 8/21 8 ET/7 CENTRAL

American Lung Association’s LUNG FORCE, Amgen, Cless Family Foundation, Fanconi Anemia Research Fund, Farrah Fawcett Foundation, Jazz Pharmaceuticals, Laura Ziskin Family Trust, Legacy Circle, LUNGevity Foundation, Mirati Therapeutics, Pancreatic Cancer Canada, Sara Schottenstein Foundation, Society for Immunotherapy of Cancer, Lew, Jean, and Kari Wolff Stand Up To Cancer is a division of the Entertainment Industry Foundation (EIF), a 501(c)(3) charitable organization.

A DV E R T I S E M E N T

The Green Future Index 2021 ENERGY TR ANSITION

KEY O

Green leaders

O

Greening middle

O

Climate laggards

O

Climate abstainers

Top 5 Rank

Bottom 5 Country

Score

Rank

Country

Score

1. . . . . . . ........... Ethiopia ....... . . . . . . . . . . . . . . . . . .

8.1

76 . . . . . . . . . . . . . . . Qatar ..............................

1.8

2 . . . . . . ............ Angola ......... . . . . . . . . . . . . . . . . . .

7.4

75 . . . . . . . . . . . . . . . Ukraine ..........................

1.9

3 . . . . . . ............ Uganda ........ . . . . . . . . . . . . . . . . . .

7.4

74. . . . . . . . . . . . . . . . Russia ............................

2

4 . . . . . . . ........... Cameroon . . . . . . . . . . . . . . . . . . . .

7.4

73 . . . . . . . . . . . . . . . Iran ..................................

2

5 . . . . . . ........... Nigeria ......... . . . . . . . . . . . . . . . . . .

7.3

72 . . . . . . . . . . . . . . . Hong Kong....................

2.1

The Green Future Index is a ranking of 76 countries and territories building a low-carbon future. It measures how economies are pivoting toward clean energy, industry, agriculture, and society through investment in renewables, innovation, and green finance.

Experience the interactive index, view the data, and download the full report at:

Q

The energy transition pillar measures the degree to which each country or territory is promoting renewable energy.

Q

Seven of the top 10 countries are in Africa, leading the way for their recent rate of change.

Q

The political will to implement an effective energy transition has been absent in Hong Kong, and the territory is highly dependent on coal-fired power.

The index ranks the green performance of countries and territories across five pillars: • Carbon emissions • Energy transition • Green society

• Clean innovation • Climate policy

technologyreview.com/gfi

The Green Future Index was produced in association with Salesforce, Citrix, and Morgan Stanley.

RE

We wanted to think about the root cause of this, not as singular incidents ... but as the product of structural, systemic racism. —PAGE 12

VO I ERT W P LearnBooks, how people are using arts, andtechnology culture to bring about social and in perspective political change.

CARLISA JOHNSON, 29 GEORGIA

Johnson created a widely shared Google Doc with links to educational material, contact details for public officials, ways to take action, and information on the Black Lives Matter movement in the days after George Floyd was murdered by a Minneapolis police officer in May 2020. The document became a resource for activists, particularly those new to the movement.

ACTIVISTS’ ONLINE TOOLBOX GRACE HEEJUNG KIM; COURTESY PHOTO

How four organizers used apps and social media to energize others during the pandemic. By Abby Ohlheiser For the past year, people in the United States were supposed to live at a distance and keep to themselves to prevent the spread of a deadly virus. At the same time, there was a nationwide reckoning on systemic racism and inequality, a contentious election, and a rise in racist violence. The activism that met those crises was often organized wholly or partially online, sometimes by young people experimenting with creative new ways to build social movements. We spoke to four people in their 20s who, through the internet, became viral voices, key organizers, and vital resources over the past year. Their stories are shared here in their own words and have been edited and condensed for clarity.

uring last summer’s iteration of Black Lives D Matter, a lot of people were entering activism for the first time. On the personal side, we were in the midst of a pandemic and I live with someone who is very immune compromised, so I couldn’t go to protests, which is what I usually do. And I felt very helpless. So I needed to figure out a way that I could feel like I was actually having some sort of impact. The people who are directly around me, those who I interact with most on social media— they are ingrained in the world of activism in the same way that I am. There weren’t necessarily private conversations where people were having these eye-opening moments or reckonings, where they were like, “Oh my gosh,

09

something needs to be done. I don’t know what to do.” The people I’m in community with—they know what to do. This document was something, I felt, for my friends to share out with their family members and their friends. That specifically speaks to me as a Black person. Most Black people didn’t need to have these conversations because they already know this. And a lot of my community is Black people. I have a lot of academics as friends. I created my post, and my call to actions, on my personal Facebook and made it public, and then my friends shared it. That’s how it snowballed. It went from campus to campus. I have no clue how this happened, but it started to go to celebrities as well. So the cast of [the TV series] Riverdale started sharing it. I noticed that there were a lot of teenagers asking if it was okay to share it, which is a demographic that I have no access to. It’s kind of cliché now, but this activism, working toward correcting inequities—you have to operate as a chorus. When one voice goes out, there are others who are still sustaining. And so even though my document was live during that specific time period when so much was happening, it was happening so fast, and I just couldn’t sustain it toward the end. But there are so many other documents that exist that have created this network that is still flourishing today. Q

I needed to figure out a way that I could feel like I was actually having some sort of impact.

Change

FIONA LOWENSTEIN, 27 Fiona Lowenstein’s Slack group for people with covid-19 now has

10,000 MEMBERS and gets about 50 to 100 new members each week.

NEW YORK

Lowenstein is the founder and editor in chief of Body Politic, a media organization and wellness collective based in New York that hosts a Slack support group for people with covid-19, including those with long-term symptoms. It now has more than 10,000 members and gets about 50 to 100 new members each week. The group was the meeting place for the Patient-Led Research Collaborative, a global collective of covid-19 patients who are recording and sharing data about their own symptoms, and which has begun to publish research on long-term covid. got sick very early on when the pandemic hit I the United States. My first symptoms were on March 13, 2020. I was sick before there was a comprehensive symptom

list from the CDC, before there was any information about longterm recoveries or young people getting severely sick. Those first couple of weeks that I was sick and in the hospital, I just lacked a lot of information. The support group started as an emotional support group. People with covid and people with long covid really needed a place to talk to each other. But then it quickly became an info-sharing group, because we were lacking information from our doctors and from health agencies. We were just sort of talking to each other and trying to figure it out ourselves. The group was on Instagram actually at the time, as a DM chat. It had maybe, like, 25 to 30 people in it. A lot of people with myalgic encephalomyelitis, or chronic fatigue syndrome, reached out to us kind of early on in the pandemic. They’re, for the most part, people who have long-term symptoms following

Julie 8:09 PM Today I walked 30 minutes without getting tired, hugged my partner for the first time since mid-March, frosted the birthday cake he made me, and just found myself in the kitchen ... of all things ... whistling! Day 53

GRACE HEEJUNG KIM; COURTESY PHOTOS

10

COURTESY IMAGE

Report

viral infections. They reached out to provide guidance on how to navigate an illness like this but also on health, advocacy, and how to interact with health agencies. I think that really informed the way that we moved forward, and made sure that we were always contextualizing long covid within these broader longterm illnesses and chronic illnesses. This is not entirely new. There’s actually a whole history of this stuff. Our support group is really just for covid patients, but we do have a special advocacy channel where we have some users from other chronic illness and disability and health justice organizations that talk about some of these more intersectional issues. For those experiencing or recovering from covid-19, we have channels for almost every system of the body: reproductive, neurological, muscular, circulatory, gastrointestinal. That’s where we’ll discuss very specific symptoms. We also have channels for specific communities. We have three private channels that you join by request: a BIPOC channel, an LGBTQ channel, and a channel for medical professionals. And then there are channels that are a little bit more geared toward people’s specific mental-health needs. We have a Victories channel. That’s where you post everything from “I took a shower for the first time in a week” to “I reunited with my family after six months.” We also have a channel called Need to Vent, which is kind of the opposite—it’s where you go when you just need to really spill your guts on how you’re feeling and how things are going. Q

11

ERYNN CHAMBERS, 28 NORTH CAROLINA

Chambers saw some bad statistics about crime in Black neighborhoods being used to support and spread racist narratives on TikTok—and, having started building a following on the app, decided to debunk them in song: Black neighborhoods are overpoliced, so of course they have higher rates of crime. And white perpetrators are undercharged, so of course they have lower rates of crime. Erynn Chambers wrote a song to debunk statistics about crime in Black neighborhoods that eventually had

TWO MILLION VIEWS on TikTok.

And all those stupid stats you keep using are operating off a small sample size. So shut up! Shut up, shut up, shut up, shut up, shut up, shut up! The song rocketed around the Internet and was shared widely beyond TikTok last summer, eventually earning more than 2 million views. Chambers— @Rynnstar on TikTok—has continued to be active and popular on the app, with 720,000 followers. was no great inspiration I T or anything like that. I was just standing on my porch one day last summer and I sang a random tune that popped into my head and hit Post. The next day, it had really taken off, and I was stunned. And then another creator, Alex Engelberg, made a remix, a choral remix of it—basically a barbershop quartet. That really made it take off. Maybe it was just the right place, right time, you know? People found it snappy, especially with the barbershop remix. Very few people plan to go viral, but everything that I’ve done that has gone viral was completely unexpected.

I use TikTok knowing that it’s not the ideal platform. It’s where I have the most followers. But it’s been pretty good. I’ve met a lot of people that are friends now through TikTok. I made a lot of connections, which has been really good, really encouraging. Honestly, that’s probably the biggest reason why I still do it, because a lot of times it’s emotionally exhausting. TikTok kind of pigeonholes a lot of Black creators into “You can only talk about this one thing.” It’s frustrating how TikTok seems to pick and choose what they let be seen on the platform. I really try to promote and push when I do other things, because I want people to know that I’m more than just political videos they share with their aunts. I’ve got a lot of interests. I want to be able to talk about my random hobbies as much as any white creator. I also enjoy talking about linguistics. I like talking about musical theater. I like talking about history. And I don’t want people to think that if they follow me for one thing, they’re never going to see anything else. Q

SUNNIE LIU, 22 TEXAS

In the summer of 2020, Sunnie Liu, an undergraduate student at Yale, was part of a small group of young ChineseAmericans who wanted to find ways to address anti-Blackness within their own community. So they co-founded the WeChat Project. WeChat, a Chinese app that is kind of like a social network, messaging service, and sharing app in one, is extremely popular with the Chinese diaspora in the US. The project creates content that seeks to counteract what is often overwhelmingly right-wing discourse, news, and misinformation shared there. So far, the project has published more than 25 bilingual articles that have reached hundreds of thousands of readers across social media. here’s not that many young people of the T Chinese diaspora on WeChat. And there’s very few progressive voices on there. So even though conservatives are a minority among ChineseAmericans and Asian-Americans in general, right-wing discourse and political information really dominate the WeChat platform. The way that WeChat works is that political news usually spreads from media accounts into group chats. These group chats often just end up being echo chambers for spreading sensationalism, conspiracy theories, and unfortunately lots of right-wing rhetoric. The vast majority of the active WeChat users in the States are members of the Chinese diaspora who are first-generation

Change

immigrants. And so they’re largely older than us. They’re probably part of our parents’ or grandparents’ generations. For the Chinese diaspora, there’s not only this generational divide; there’s often also a language and a cultural divide. At the dinner table, you have parents who are most comfortable speaking Mandarin, and the kids are most comfortable speaking English. And so the kid could be trying to express some complicated political idea in English, but then the parents have no idea what’s going on, and vice versa when the parents are speaking Mandarin. They have very different traumas that they’ve experienced in either generation, too, with the kids growing up and experiencing racism, perhaps since they were five and entered into American schools, versus the parents growing up in China, where there’s not really the same racism. But they might have lived through the Cultural Revolution. We’re trying to bridge that gap by starting these conversations between the generations online in a language that meets the parents and grandparents of the Chinese diaspora where they are. The main source of news on WeChat comes from media accounts. WeChat calls them “official accounts.” They’re essentially microblogs that publish articles people can read and share and discuss, and like and comment on. We’re out there publishing with some of the rare progressive media accounts on WeChat. Because we’re a rare voice of the children’s and grandchildren’s generation, a lot of people in the older generations are curious because of these cultural and language barriers. Unfortunately, there’s a lot of miscommunication between the kids and their

The WeChat Project is trying to

BRIDGE THE GAP by starting conversations between the generations online.

parents, and the grandparents and the grandchildren. They’re curious what their children are thinking. In response to anti-Asian violence, for us, especially on WeChat but also in general—honestly, throughout the news media, we felt like most coverage of the topic was missing class analysis and gender analysis. Especially after Atlanta, this is something that all Asian-Americans need to worry about. This is an attack on our lives, on our safety. But if you look at who was actually attacked, the vast majority of these people were the most vulnerable in our communities. They are low income. They are elderly. Often, they’re immigrants who do not speak English. And these are very different from the people who are most visible and vocal in response. And so we wanted to also think about the root cause of this, not as singular incidents by racist people, but as the product of structural, systemic racism and classism. Q

GRACE HEEJUNG KIM; COURTESY PHOTO

12

Active Molecule Unlocked

Clinical studies show that the natural molecule, Urolithin A, improves mitochondrial function and boosts leg muscle strength*. But, it’s not as simple as eating more pomegranates. Most people’s microbiomes don’t produce enough Urolithin A by eating the precursors found in pomegranates. Our scientists in Switzerland have spent more than a decade researching and developing the first clinically tested, highly . pure form of Urolithin A, unlocking its benefits for everyone. We called it A precise dose of 500mg Mitopure delivers six times more Urolithin A than a glass of pomegranate juice.*

timelinenutrition.com/sa

*See disclaimers on website.

As we age, our ability to generate vital cellular energy declines. This can lead to a decline in muscle function, metabolism, energy levels and resiliency. Aging is an inevitable part of life, but scientists have discovered a molecule that rejuvenates the power sources inside our cells – our mitochondria.

14

Change

T R : Q + A

THE ERA OF AUTONOMY Zoox CEO Aicha Evans on leading the self-driving-car company and the power of opportunity. By Anthony Green | Portrait by Winni Wintermeyer

The past year has been a wild ride for Aicha Evans. Zoox, the autonomous-vehicle company Evans leads as CEO, was acquired by Amazon last summer for a reported $1.2 billion. And when the company unveiled its vehicle in December, the car represented a significant departure from the automobile as we know it. Meant to serve as a self-driving taxi, it looks more like a high-tech carriage than a car. Sliding glass doors welcome passengers from either side, and each of the vehicle’s outer corners houses a “sensor pod” with multiple lidar and radar modules and cameras to help it navigate. Beneath the floorboard is an electric motor that can whisk passengers to their destination at up to 75 miles per hour. Zoox’s vehicle is one of the few driverless rigs built from the ground up, and Evans says it’s also one you’ll never get to own. Instead, Zoox plans to launch an app-based ride-hailing service in cities including San Francisco and Las Vegas, where the vehicle is being tested. I spoke with Evans about what it’s like to try to change an industry and transform the way we move about our cities.

Report

Q: How do you define Zoox? Is it an AI company? A robotics company? A: It’s a transportation com-

pany that takes advantage of AI, robotics—all of the new techniques around electric vehicles and software—and blends all of that to basically open up a new area of transportation. A city like San Francisco that has housing issues and is worried about business flight has 30% of its real estate footprint dedicated to parking. So if people were using Zoox to go from point A to point B, those buildings could be replaced, reclaimed for businesses or for housing and parks. The other thing that’s really important from a Zoox standpoint is taking advantage of sensors and computing to make all of this happen. One of the questions we get all the time is: Why are you building a vehicle? Well, because the passenger car of today was architected and designed for human drivers. Rearchitecting and redesigning the vehicle to make it easiest and safest for AI to drive is what we’re all about. Q: How might autonomous vehicles affect our lives when we’re not on the road? A: The world 30, 40 years

from now will look very different. We talk about autonomy as really the beginning of a wave. Sort of like also what happened with the internet and then the PC and then wireless, and then the smartphone. I mean, the smartphone is not that old, right? Sometimes

I’m like, how did we even operate without these things? Well, we did. And I think autonomy will allow a lot of things like that— around goods, around services. I think there are a lot of things we physically go to a place for today that in the future will come to us through autonomy. Q: Many driverless cars are primarily trained in Western, urban environments. How well will these systems work in other places? A: Mathematical algorithms

are not biased. However data can be. Not because the data is bad but because of where you collect the data. In terms of Zoox, what I can tell you is that we will not go somewhere without training on local data sets. If you don’t do that and you make assumptions, life could get quite complicated. I think that as an industry, we understand the science, and it’s important to understand from an input standpoint what could be problematic. You’re also more likely to do that if there are people in the room who don’t all look the same and think the same. Q: What’s your approach to leadership? A: It’s changed over the years.

You’re first an engineer, and you get noticed because you’re one of the best in the room on the project. And you try to improve and learn more and have more impact, but very quickly you sort of understand the math around one versus many. And as a

single person, you can only do so much. Leadership is not command and control. It’s really about: How do you get people together? How do you get them sold and bought in on a mission? And then how do you work together to accomplish it? Q: When it comes to building teams, what’s your approach to diversity? A: We’re building a consumer

product at the end of the day, and consumers come in all shapes, all races, all genders, all … everything. And so it’d be bad to build a consumer product without having people who look like and think like the consumers. Another element of it is access and equity. Like, for an African-American woman with short hair, getting my hair cut is a big deal. Not a lot of people can cut my hair—let me just put it that way. You start noticing any time you move to a new city: “Oh, I just have to find MLK Boulevard. And on either side of it, all the barbers will know how to cut my hair.” And that was true in DC. That was true in Austin. That was true in Portland. And so I look forward to our vehicle picking up people in whatever neighborhood they are from, and basically giving them the opportunity to be transported to where the economic access is. Q: You’re a Black woman in power, in tech. That is not common. How do you deal with that? A: I don’t actually wake up

every day thinking I’m a

15

Black woman with power in tech. Occasionally when I’m getting pulled over or when somebody is doing something stupid, I am reminded, but I don’t want to be an angry person all the time. I don’t think that’s very productive. What I do think about is, I represent opportunity. I represent that it is possible. I’ve spent a lot of time asking myself: What is embedded systematically that makes this rare, and how can we break that down? I remember going to a Lego robotics competition with my son for the first time in the Bay Area. And my son said to me, “Wow, there’s only one of me here.” I’m like, “No, no, no, you’re confused.” And I looked around, and yes, there was only one of him and there were very few girls’ teams. And the lack of opportunity is already right in front of you. And then you bring that to Zoox and you say, “Hey, crew, we’re going to sponsor Lego robotics. Are some of you interested in being mentors?” I look at opportunity. I look at how do you make a meaningful, positive difference? This is our problem as a nation. These are problems that have not been solved for centuries. Our nation was built and started a certain way. So a little bit of honey and less vinegar might be the right way to go. This interview has been condensed and edited for clarity. Anthony Green produces podcasts at MIT Technology Review.

16

Change

ARGUMENTS

THE DANGEROUS APPEAL OF TECH

Sheila Jasanoff is a professor of science and technology studies at the Harvard Kennedy School.

Technology doesn’t rule us. We direct it, but often by inaction.

ove it or hate it, technology enthralls us with the promise of change. Sometimes it’s the presumed benefits that grab our attention: curing disease, replacing fossil fuels, increasing food supplies, unlocking the secrets of the deep sea, colonizing Mars, or ending the ravages of old age. Other times the risks loom larger. What if we unleash a killer virus, set in motion a nuclear doomsday, block out harmful solar radiation with chemicals that prove toxic, or build computers that decide humans are dispensable? The battle between light and dark in the way we imagine technological change is ancient. In Greek mythology, Prometheus suffered agonies for bringing fire to Earth, and Daedalus lost his son to the urge to fly to freedom. But the most optimistic and

L

most pessimistic views of technology both rely on a common misconception: that a technological pathway, once embarked upon, leads to inevitable social consequences, whether utopian or dystopian. This view, known as technological determinism, is historically flawed, politically dangerous, and ethically questionable. To achieve progress, societies like ours need a more dynamic understanding of why technology changes, how we change with it, and how we might govern our powerful, marvelous machines. Technology is not an autonomous force independent of society, nor are the directions of technological change fixed by nature. Technology at its most basic is toolmaking. Insisting that technological advances are inevitable keeps us from acknowledging the disparities of wealth and power that drive innovation for good or ill. Technology is always a collective venture. It is what it is because many people

imagined it, labored for it, took risks with it, standardized and regulated it, vanquished competitors, and made markets to advance their visions. If we treat technology as self-directed, we overlook all these interlocking contributions, and we risk distributing the rewards of invention unfairly. Today, an executive officer of a successful biotech company can sell stock worth millions of dollars, while those who clean the lab or volunteer for clinical trials gain very little. Ignoring the unequal social arrangements that produced inventions tends to reproduce those same inequalities in the distribution of benefits. Throughout human history, the desire for economic gain has underwritten the search for new tools and instruments—in fields like mining, fishing, agriculture, and recently gene prospecting. These tools open up new markets and new ways to extract resources, but what the innovator sees as progress often brings unwanted change

MARTHA STEWART

By Sheila Jasanoff

SELMAN DESIGN

Report

to communities colonized by imported technologies and their makers’ ambitions. For example, in West Bengal, where I was born, weavers lost such skills as making the intricate narrative motifs of the Baluchari sari during 200 years of British rule. Indeed, Britain’s first industrial revolution, which introduced the power loom in cities like Lancaster but adopted punitive tariffs to keep out hand-loomed cloth from India, was also a story about dismantling Bengal’s once-flourishing textile industry. Lost arts had to be regained after the British left. The cost of a radical break with a nation’s own economic and cultural heritage is incalculable. The desire for military advantage is another driver of technological change that can, in some instances, benefit civil society—but “dual use” technologies often retain ties to forces that prompted their development. Nuclear energy, a spinoff from the pursuit of the atomic bomb,

was sold to the world by US President Dwight Eisenhower as “atoms for peace.” Yet nuclear power remains closely tied to the threat of nuclear weapons proliferation. Similarly, the internet and world wide web, which revolutionized how much of the world lives today, owe much to the US Defense Department’s vision of a network of computers. First celebrated as a space for emancipation, the digital world has slowly revealed its antidemocratic features: constant surveillance, cybersecurity threats, the lawlessness of the dark web, and the spread of misinformation. More public awareness of the internet’s origins might have led to a more accountable cyberworld than the one designed by hotshot technologists. The story of the internet shows that modern societies are often better at imagining the upsides of technology than its downsides. But the trajectory of innovation is also guided by more subtle cultural preferences, often with profound consequences.

17

In US biomedicine, for example, energy, attention, and money tend to be directed to high-impact, silver-bullet solutions, or “moonshots,” rather than to messier changes in the social infrastructures that give rise to many health problems. This inclination is reflected in Congress’s decision to authorize $10 billion for Operation Warp Speed to bring a covid-19 vaccine quickly to market. Moderna owes much of its success as a vaccine manufacturer to that massive public spending, and both Moderna and Pfizer have benefited hugely from lucrative supply contracts with the US government. At the same time, about a third of all US deaths from the pandemic occurred in nursing homes, a result of decades of underinvestment in the unglamorous social practices of elder care. Collectively, we chose to ignore the plight of the vulnerable elderly, and spent big on technology only when everyone was at risk. Change may not be inevitable, but economists have a point when they talk about “path dependency,” or the notion that once an engine gets going it’s bound to follow an existing track. Sunk costs— foundations laid, machinery ordered, workforces trained—cannot be recovered. It often seems easier to go where the flows of materials and social practices have already cut deep channels. It’s not surprising, then, that defense spending has proved to be one of the prime motivators of innovation, even though such investments perpetuate power imbalances and seldom respect cultural or ethical sensitivities. In his famous poem “The Road Not Taken,” Robert Frost reflects on how the human mind constructs narratives of inevitability. We come to a fork in the road, we choose a path, and then as memory plays its tricks we come to see that choice as shaping all that came after. Faced with mounting problems of inequality, diminishing resources, and a looming climate calamity, we must learn to recognize the flaws in such linear storytelling, and to imagine the future along as-yet-untraveled pathways of change. Q

18

Change

Joe Garcia is a correspondent for the Prison Journalism Project and a staff writer for San Quentin News at San Quentin State Prison, where he is incarcerated.

LEFT BEHIND

How are prisoners supposed to reenter society when technology has moved on? By Joe Garcia

alifornia recently promised to provide free computer tablets to all state prisoners by the end of 2021, allowing prisoners like me to email our loved ones through a highly restricted prison messaging service and download content like movies and books. It’s a great first step, but without more open and frequent internet access, there’s no way we’ll ever truly keep pace with the changing world outside our prison walls. I’ve been locked up since 2003. Back then Apple had barely launched iTunes, and I was still in awe of the so-called highspeed connection I’d paid Time Warner to install in my apartment. In all the years since then, I haven’t logged a single second of internet activity. My frames of reference for what it means to be online now come from network television and print media. When my first parole opportunity arrives, in 2023, I’ll be 53 years old. As

C

a convicted murderer, I will need to convince the parole board not only that I have been rehabilitated, but also that I am able to be a productive and employable citizen. It can be hard to keep up with changes in technology even when you’re experiencing them firsthand. When you’re locked away, it’s virtually impossible. I’m hoping to pursue a career path in journalism when I get out of prison, and I worry every day about reentering a global economy with my grossly outdated tech skill set. I know the job market will expect internet fluency more than ever in the postpandemic world, since much of America’s media workforce has gone remote. Roughly 2.3 million people are incarcerated in the US. Even though the internet is a given in the rest of society, access in prison is so restricted it’s almost nonexistent. Prisoners are only allowed to use a tiny number of programs that might offer Zoom classes with outside teachers, or to browse

an extremely limited set of whitelisted sites through intranets that are carefully cordoned off from the public internet. At San Quentin State Prison, where I reside, the computers that prisoners can access feature preloaded interactive programs that offer the same basic experience as reading from a textbook. My only experience using a search engine has been through LexisNexis, which the prison library licenses to allow us to study case law. In San Quentin’s vaunted coding program, Code 7370, offered through the prison education organization The Last Mile, hand-selected prisoners build and sell actual websites for commercial use. But even they do not have internet access. As a working journalist in prison, I know I have First Amendment rights, but I’m deprived of a key technology that the United Nations identified a decade ago as a means of exercising one’s freedom of expression. I understand why government officials, prison administrators, and the public might fear that giving convicted felons real-time internet access would open a Pandora’s box of suspicious activity. Yet if elementary school kids can surf the web safely with parental locks and controls, how hard would it be for someone to design a system that provides incarcerated people with more meaningful access? In Belgium, for example, an innovative platform called PrisonCloud has offered limited and controlled internet access to prisoners for years. In Finland and Denmark, open prisons, which have minimal security and some of the lowest recidivism rates, also allow limited internet access. Almost all incarcerated people in US prisons today will be released back into their communities in the future. That includes many of my peers who have been incarcerated since before the advent of the internet. President Joe Biden has said that the majority of incarcerated Americans deserve a bona fide second chance at life—but we need some form of internet access in order to have a genuine chance of successfully reentering today’s tech-driven society. Q

PHOTO: EDDIE HERENA; ILLUSTRATION: DANIEL ZENDER

ARGUMENTS

Report

The rise of influencers has made it easier for young people to fall for misinformation. By Jennifer Neda John

IN SOCIAL MEDIA WE TRUST teenage girl peers gravely at the camera, the frame wobbling as she angles her phone at her face. A caption superimposed on her hoodie shares an ominous warning: If Joe Biden is elected president of the United States, “trumpies” will commit mass murder of LGBT individuals and people of color. A second caption announces, “this really is ww3.” That video was posted to TikTok on November 2, 2020, and liked more than 20,000 times. Around that time, dozens of other young people shared similar warnings across social media, and their posts drew hundreds of thousands of views, likes, and comments. Clearly, the claims were false. Why, then, did so many members of Generation Z—a label applied to people aged roughly 9 to 24, who are presumably more digitally savvy than their predecessors—fall for such flagrant misinformation?

COURTESY PHOTO

A

19

I’ve worked as a research assistant at LGBT, and their past posts discussed familthe Stanford Internet Observatory since iar topics like family conflict and struglast summer, analyzing the spread of gles in math class. This sense of shared online misinformation. I’ve studied for- experience made them easy to believe, eign influence campaigns on social media even though they offered no evidence for and examined how misinformation about their claims. Making matters worse was the inforthe 2020 election and covid-19 vaccines went viral. And I’ve found that young mation overload many people experience people are more likely to believe and pass on social media, which can lead us to trust on misinformation if they feel a sense of and share lower-quality information. The common identity with the person who election rumor appeared among dozens shared it in the first place. of other posts in teenagers’ TikTok feeds, Offline, when deciding whose claims leaving them with little time to think should be trusted and whose should be critically about each claim. Any efforts ignored or doubted, teenagers are likely to challenge the rumor were relegated to to draw on the context that their commu- the comments. As young people participate in more nities provide. Social connections and individual reputations developed through political discussions online, we can expect years of shared experiences inform which those who have successfully cultivated family members, friends, and classmates this identity-based credibility to become teenagers rely on to form their de facto community leaders, opinions and receive updates attracting like-minded people on events. In this setting, a and steering the conversation. community’s collective knowlWhile that has the potential edge about whom to trust on to empower marginalized which topics contributes more groups, it also exacerbates to credibility than the identity the threat of misinformation. Jennifer Neda of the person making a claim, People united by identity will John is a sopheven if that identity is one the find themselves vulnerable omore at Stanford University young person shares. to misleading narratives that majoring in Social media, however, target precisely what brings human biology. She researches promotes credibility based on them together. online misinidentity rather than commuWho, then, has a role to play formation at the nity. And when trust is built in promoting accountability? Stanford Internet Observatory. on identity, authority shifts to Social media platforms can influencers. Thanks to looking implement recommendation and sounding like their followers, influ- algorithms that prioritize a diversity of encers become trusted messengers on voices and value discourse over clickbait. topics in which they have no expertise. Journalists must acknowledge that many According to a survey from Common readers get their news from social media Sense Media, 60% of teenagers who use posts viewed through the lens of idenYouTube to follow current events turn to tity—and present information accordingly. influencers rather than news organiza- Policymakers must regulate social media tions. Creators who have built credibility platforms and pass laws to address online see their claims elevated to the status of misinformation. And educators can teach facts while subject matter experts strug- students to assess the credibility of sources gle to gain traction. and their claims. Shifting the dynamics of online diaThis, in large part, is how the rumor of plans for post-election violence went viral. logue will not be easy, but the dangers The individuals who shared the warning misinformation can fuel—and the promwere deeply relatable to their audience. ise of better conversations—compel us Many were people of color and openly to try. Q

20

Change

VIRTUAL POLITICS

consulted in social or political matters, even though digital platforms have provided us with a voice and a way of expressing it earlier in life (an estimated 81% of teenagers 13 to 17 are now active on at least one social media site). That may stem from a feeling that our voices don’t matter because we cannot vote until we turn 18. But most of us will be able to by the next presidential election in 2024, if not sooner. Digital platforms have the potential to redefine civic engagement and allow the opinions of both young and old people to play a deeper role in policymaking. As my generation speaks out online, the lawmakers who are shaping our future will need to figure out how best to listen to those of us who will live in it. Otherwise, young people’s enthusiasm for politics could dry up. At a time when our trust in government is nearing historic lows, the future of political participation is at stake.

Young people are speaking out online. Will it lead to real-world action?

Digital democracy

By Kiara Royer

The idea that some combination of technology and a new generation is redefining politics is not new—the same thing happened with the radio, and later with television. But social media, in particular, has brought unique changes. That means my generation has a special role to play in figuring out how these platforms get used. The ways young people use such tools are already changing the look of political campaigns and grassroots organizing. Many nonprofits and other groups are now recruiting more and more young people to play larger roles within their organizations. The key to making sure young people stay engaged is including them in more political conversations, says Beth Simone Noveck, director of New York University’s Governance Lab and New Jersey’s first chief innovation officer. Noveck leads a project called CrowdLaw, which studies ways lawmakers can use technology to incorporate the opinions of citizens, especially young ones, into the legislative process. She also heads a GovLab program called ReinventED, which centers on using technology to engage students, educators, and caregivers, especially from

ast summer, my friend Jessica Rosberger texted me with an idea. “I think I may have something,” she began. We were about to graduate from high school and had spent the last three months of senior year taking classes at home because of the covid-19 pandemic, and lately we’d been following the news of racial justice protests around the US in the wake of George Floyd’s murder. An hour and a half later, we published Jessica’s idea as an online petition. In it, we argued that former attorney general William Barr, who graduated from our high school and was given the Distinguished Alumni Award in 2011, had violated the school’s core values with his involvement in the violent removal of protesters from Lafayette Square in Washington, DC, on June 1, 2020. We hoped our petition would encourage the school’s alumni council to rethink Barr’s award. Jessica and I coordinated over Google Docs, talked with reporters and alumni over Zoom, and shared the petition on Instagram, Facebook, and Twitter. By July, it had more than 8,700 signatures, was cited in an op-ed in the Washington Post, and propelled us to a virtual meeting with the alumni council. It was my first taste of the power of using the internet and social media as political tools. Unfortunately, it’s a feeling that’s still too rare, even for my generation—young citizens’ opinions are rarely

L

Kiara Royer is a freshman at Williams College majoring in history and political science. She is the news section editor for the Williams Record.

YUKO ROYER

ARGUMENTS

Report

21

presidential campaign last year. “I think that there certainly are people who will just post about something on social media and that’s the end of the chain, but lots of those people are people who wouldn’t have done anything at all,” he says.

ALLIE SULLBERG

Staying engaged

marginalized communities, in efforts to solve education issues. Exercises completed by ReinventED show that students’ priorities even in the midst of a pandemic lean toward solving real-world problems and improving nontraditional academic subjects. Policymakers, on the other hand, are more concerned with public health and school reopening plans. “The people who are most expert in education—mainly students and teachers, and to a lesser extent the parents of those students—are rarely, if ever, consulted in how we design our schools,” Noveck says. “My hope is that by using tools like this, by laying bare what people really care about, that can help to change the direction of what we’re focusing on.” Digital platforms, however, may be a double-edged sword. Participating in online movements may not translate into offline engagement—some experts warn it could have the opposite effect. “On social media,

you can get a burst of interest, sometimes a burst of activity, because it’s so easy to feel like you’ve participated just by clicking a link or retweeting something or using a hashtag,” says Nicholas Carr, a sociology professor at Williams College. “What’s unclear is whether social media will help or hurt the ability of activists to sustain interests in a long-term campaign of change.” Instead, the result may be “slacktivism,” a term coined during the rise of the internet for the practice of publicly supporting a cause in ways that take little effort, often to make yourself look good. “That can diminish or even demean the seriousness of political discourse in a way that can kind of hinder our ability to solve big problems,” says Carr. People who engage in this performative activism are still spreading political messages, though, says William Golub, a junior at Stanford University who volunteered with the texting team on Joe Biden’s

After we met with the alumni council last July, months passed, and Jessica and I hadn’t received any updates on Barr’s award. Frustrated, we published an open letter to the council on Medium in early September. The council responded two days later with a public update stating that it would share its decision once its written report was complete. “Ultimately,” read the report released two months later, “we would not recommend revoking the Award bestowed on then-former Attorney General Barr in 2011.” (Barr held the post from 1991 to 1993, and again in 2019–2020.) The council said this decision was based on community feedback, the “complex” process of revocation and precedent, and the lack of “undisputed information available” regarding Barr’s involvement at Lafayette Square. It was devastating. I felt as though the council, whose youngest member graduated from high school in 2002, had dismissed our efforts. And yet, I can see now that our work wasn’t in vain. Our school’s student-run newspaper published an in-depth analysis of the report, admonishing the council for its decision. Jessica and I received emails from our former teachers, who said our petition had sparked classroom discussions about topics ranging from Barr’s actions to political engagement more broadly. And the council contacted Jessica and me directly, thanking us “for taking an active role in alumni affairs and for your early dedication, as alumnae, to the legacy of the school.” Even though the final decision was not what I had hoped for, the experience taught me that my voice is just as important as the voices of people much older, and that technology can help make it heard. But people must be willing to listen. Q

Artificial intelligence, demystified The Algorithm newsletter

Exclusively for MIT Technology Review subscribers.

technologyreview.com/algorithm

23

I N N O V A T O R S U N D E R The 35 Innovators Under 35 is our yearly opportunity to take a look at not just where technology is now, but where it’s going and who’s taking it there. More than 500 people are nominated every year, and from this group the editors pick the most promising 100 to move on to the semifinalist round. Their work is

then evaluated by our panel of judges (you can see who they are on page 47), who have expertise in such areas as artificial intelligence, biotechnology, software, energy, and materials. With the insight gained from these rankings, the editors pick the final list of 35. —The editors

GUTTER CREDIT HERE

24 Change

35 Innovators Under 35

Their innovations point the way toward light-based chips, better gene editing, and skinlike electronics.

RYAN BABBUSH AGE

AFFILIATION

32

Google

IN N O V A T I O N

RYAN YOUNG

Efficient quantum simulation algorithms might help find novel, powerful materials. Molecules are complicated. Forget the grade-school picture of electrons orbiting a nucleus like planets around the sun. Electrons can be shared among many atomic nuclei. They interact with one another in ways described by the equations of quantum mechanics. It’s these complex interactions, which grow exponentially with the number of electrons, that largely govern chemical reactions and the properties of molecules. Simulating these electrons with perfect precision might take a conventional computer millions of years. But algorithms running on quantum computers might be able to perform precise computations in days or even

hours. This would provide clues on how to precisely design molecules with desired properties and tailor their reactions with amazing control. Sufficiently precise quantum simulation might allow chemists to create new compounds like better high-temperature superconductors, catalysts that could take nitrogen or carbon dioxide out of the air, new drugs, more efficient solar cells, strong lightweight materials for airplanes, and so forth. It would be a way to quickly figure out how a new substance would behave without actually having to synthesize it. It might herald a new age of materials science. Between 2014 and 2020, Ryan Babbush published dozens of papers—together with collaborators at Google and elsewhere—that outlined dramatically more efficient quantum simulation algorithms. The upshot is that some quantum simulation calculations could, in principle, be done in hours, on a sufficiently powerful quantum computer. Take the case of nitrogenase, an enzyme that some bacteria use to

25

remove nitrogen from the air and create ammonia, a compound of nitrogen and hydrogen. This process, known as nitrogen fixation, is essential for agriculture, which is why nitrogen-based fertilizers are a linchpin of the world’s food system. Nitrogenase is a big molecule that includes a catalytic site known as FeMoco. Currently, an energy-intensive technique known as the Haber-Bosch process produces most fertilizers, accounting for about 2% of humanity’s total energy usage. “If we could figure out how that enzyme [nitrogenase] is doing this, then we might be able to design an industrially viable alternative for producing fertilizer, which could scale and save a huge amount of energy,” Babbush says. He and his collaborators have found a potential way to use a quantum computer to analyze FeMoco and shed light on the mechanism by which it first breaks the bonds between nitrogen atoms that are bound together in nitrogen gas and then succeeds in combining the nitrogen with hydrogen. (Babbush acknowledges that competing approaches using clever approximations to simulate molecules on classical computers might get there first.) Another line of research that Babbush has advanced aims to figure out how quantum computers can calculate the behavior of electrons in metals and crystals. Potential applications could include finding better superconductors or making more efficient solar cells. In these materials, the repeating pattern of the atoms creates very complicated behavior among the interdependent electrons. And Babbush is figuring out how quantum computers can be used to make sense of these interactions. If quantum computers succeed in remaking our material world, Babbush’s work will be one reason why. —SIOBHAN ROBERTS

Change

AMAY BANDODKAR AG E

A FFILIATION

33

North Carolina State University

VIRGINIA SMITH

AGE

AFFILIATION

31

Carnegie Mellon University

INNOVATION IN N OVATION

His lightweight sensors could make wearable tech more useful and practical.

Wearable technology can provide real-time information about a person’s health and fitness, but creating sensors that can collect data without a cumbersome and impractical system of staying powered has proved difficult. Amay Bandodkar thinks he’s hit on a new way of creating “self-powered” biochemical sensors through unconventional technologies, making wearable tech lighter and less cumbersome. It’s about four times smaller and 20 times lighter than similar devices produced two years ago, he says. The key to shrinking the sensor was overhauling how it’s powered. “All the groups that were working on this were using these really bulky batteries, and the sensor was around 3% of the total size and weight,” he says. So he built a sensor that doesn’t require a battery: it harnesses the catalytic properties of enzymes to generate signals without the need for power sources. While this concept can be used to develop self-powered sensors for some chemicals, for other kinds of sensors that still need a power source, Bandodkar has developed a lightweight battery that runs on sweat. It’s made of a magnesium anode and a cathode made of silver and silver chloride, separated by a dry cellulose membrane. When a person wearing it starts to perspire, the cellulose membrane absorbs the sweat and acts as an electrolyte, effectively turning the battery on and powering the sensor. Bandodkar has successfully tested out a heart-rate sensor running on his battery, opening the gateway for heart-monitoring wearables. —NEEL V. PATEL

Her AI techniques are efficient and accurate while preserving fairness and privacy.

hen Virginia Smith began her PhD in artificial intelligence, she had a question: How do you train a neural network on data that is stored across multiple machines? Her attempts to answer it have made her a leader in the field of federated learning, which seeks to handle data spread across hundreds, or even millions, of remote sources. Google researchers first introduced federated learning in 2017 to use with the company’s mobile devices. The method they devised involved training millions of neural networks locally before sending them to a company

W

server to be merged together in a master model. It allowed the master model to train on data from every device without making it necessary to centralize that data. This not only reduced latency in the mobile experience but could also improve each user’s data privacy. But combining millions of AI models also risks creating a central model that performs well on average but poorly for outliers—for example, voice recognition software that fails when the speaker has an unfamiliar accent. So Smith proposed a new technique for more “personalized” federated learning. Rather than merge a million localized models into one, it merges the most similar localized models into a few—the more heterogeneous the data, the greater the number of final models. Each model still learns from many devices but is also tailored to specific subsets of the user population. Smith also works to overcome other challenges in federated learning, such as accounting for different power and memory constraints on different devices. To encourage more research, she co-created an open-source tool that lets researchers test their federated techniques on more realistic data sets and in more realistic environments. —KAREN HAO

JESSICA JOHNSON (BANDODKAR); ILLUSTRATION SOURCE: COURTESY PHOTO

26

35 Innovators Under 35

27

JIE XU A GE

AFFILIATIO N

33

Argonne National Laboratory

I NNOVATION

WES AGRESTA (XU); ILLUSTRATION SOURCE: COURTESY PHOTO

She makes durable, easy-to-manufacture polymer semiconductors for skin-like electronics.

Jie Xu has made printable, stretchable electronics viable for mass production. Her multiple breakthroughs could be used in future wearable technology, advanced robotics, and human-computer interfaces with sensors connected to the skin. The key for Xu was inventing polymer circuits that kept working despite being flexed, stretched, and repeatedly moved. That had been a challenge for researchers until 2016, when Xu engineered a two-polymer coating applied to a rubbery surface that could be stretched to twice its size and still conduct electricity. In 2019, she refined the technology so that her stretchable semiconductors could be mass-produced using roll-to-roll manufacturing, a common industrial fabrication process used to print anything from textiles to plastics on large rollers. It was the first time anyone had achieved such a feat at scale. In the short term, Xu’s materials and manufacturing inventions can make flexible displays and skin-worn medical sensors much more practical and easy to make. Samsung Electronics has already patented two methods Xu helped define during collaboration with the company. Xu’s materials could also aid in the design of prosthetics with functional skin-like outer coverings. Wary of adding yet more plastics into the world, Xu is searching for versions of the polymer semiconducting materials that are recyclable or biodegradable. “I think that kind of idea should be integrated from the very beginning of any commercial material,” she says. —RUSS JUSKALIAN

XIAO SUN

AGE

AFFILIATION

34

IBM

INNOVATION

He designs imprecise—but energy-efficient—AI hardware and software.

rtificial-intelligence systems often require a vast amount of computation. That’s why in recent years, AI hardware researchers have been striving to achieve lower precision, which is good enough to produce a correct answer but avoids the use of calculations that require keeping track of lots of digits. Deep learning relies on networks that might have dozens of layers, and millions, or even billions, of parameters that must be adjusted to the correct values, a process called training the network. This often takes days

A

or weeks of computations using hundreds of specialized chips. Xiao Sun is part of a research group at IBM that has been finding ways to perform those computations using three-digit, or even just two-digit, numbers (in contrast, a modern laptop or cell phone uses 20 digits to make calculations, while most dedicated machine-learning chips use five). The real trick is in finding techniques that allow for small numbers to be used throughout the computation. You might still have to do many trillions of computations, but each one will be far simpler. This saves both time and energy—using two-digit numbers is more than 20 times more energy efficient than doing the same calculations using numbers in the billions, according to a paper by Sun and colleagues at IBM. In February, IBM announced a new chip, based in part on Sun’s work, that trains neural networks using computations involving mostly three-digit numbers. The company hopes to use it not only to train large neural networks in cloud computing centers but also in mobile phones that could train on local data. —PATRICK HOWELL O’NEILL

YICHEN SHEN

Change

AGE

AF F I L I A T I O N

32

Lightelligence

NICHOLAS HARRIS

AGE

AFFILIATION

33

Lightmatter

I N NOVATION

INNOVATION

Optical chips that can make calculations for neural networks are poised to become big business.

Shining light through optical chips might be the fastest way for neural networks to make decisions.

here are two basic types of computations involving neural networks. First, the networks must be trained, which usually involves showing them lots of data, causing them to adjust the strength of the connections between their numerous “neurons.” Next, those existing connections are used to make decisions. It’s the difference between learning to drive and driving. The difference is crucial. If a neural network takes weeks to learn how to recognize images, that’s not necessarily a problem. But if it is driving an autonomous car, it needs to be able to make life-or-death inferences in fractions of a second. That’s where optical computers come in. Despite decades of research, they’ve never worked that well. It’s harder to manipulate photons than electrons. But for certain types of computations—like those commonly needed when using an existing neural network to make inferences—photons are just the thing. In 2017, Yichen Shen and Nicholas Harris (at right) published a widely cited paper on the use of optical circuits for machine-learning tasks including speech and image recognition. Their design, one review article notes, “represents a truly parallel implementation of one of the most crucial building blocks of neural networks using light, and modern foundries could easily mass-fabricate this type of photonic system.” This means that optical computers on a chip could become a huge business, with one in every device that uses a neural network to make decisions. Shen and Harris now run competing startups. Shen’s firm, Lightelligence, released a prototype optical AI chip in 2019, and Shen says they have secured over $100 million in funding. —KONSTANTIN KAKAES

or decades physicists and engineers have dreamed of making optical chips that use photons, not electrons, to do computing. Such circuits could be lightning fast and energy efficient. But making them work has been difficult. In 2017, Nicholas Harris, together with Yichen Shen (at left) and other colleagues at MIT, published a widely cited paper describing a design that allowed them to calculate the outputs of neural networks that had been conventionally trained. The paper describes a circuit of 56 programmable interferometers—devices that carefully break apart and recombine light waves. The circuits they created solved a simplified problem of recognizing vowels correctly—distinguishing about three quarters of the 180 cases they tried. This wasn’t as good as a conventional computer, which got over 90% of them right. Shortly thereafter, Shen and Harris launched competing startups. Once a given neural network has been trained and implemented on an optical chip, performing inferences— figuring out which vowel corresponds to which sound, or how an autonomous car should react if a pedestrian steps into the street—can be almost as simple as shining light through it. This has the advantage of being both fast and energy efficient. In March 2021, Lightmatter announced it would soon start selling a “machine-learning accelerator” chip. “It’s just a completely different kind of computer,” says Harris. “Right now we’re at about a factor of 20 times more efficient than the most advanced node in digital computers.” Lightmatter closed a second round of funding in May, bringing its total investment to $113 million. —KONSTANTIN KAKAES

T

F

COURTESY PHOTO (SHEN); BOB O’CONNOR (HARRIS)

28

35 Innovators Under 35

JONATHAN GOOTENBERG A GE

30

AFFILIATIO N

MIT

SHELLEY ACKERMAN

AGE

AFFILIATION

29

Bolt Biotherapeutics

INNOVATION I NNOVATION

JUSTIN KNIGHT (GOOTENBERG); JORDAN ENGLE (ACKERMAN)

Expanding the capabilities of gene editing.

The gene-editing tool CRISPR uses a protein called Cas9 to snip out a targeted part of the genome. It does amazing work but has downsides. It can cause unintended edits to other places in the genome, and if you want to make only a temporary tweak, Cas9 can’t do it. Jonathan Gootenberg is creating other editing tools to get around these shortcomings and add to the capabilities of CRISPR. Gootenberg has used Cas12—a more compact protein than Cas9— to edit many genes at a time. This capability could be used to edit a patient’s immune cells so that they fight cancer. Then there’s Cas13. Gootenberg and his colleague Omar Abudayyeh (a 2020 Innovator Under 35) demonstrated that the protein could target RNA instead of DNA—an intriguing finding. Many viruses have RNA as their genetic material, and bacteria have both DNA and RNA, so the researchers reasoned that you could use Cas13 to find genetic material from pathogens in human cells. That led to repurposing the gene-editing tool as a paper-based diagnostic test, and in 2019 Gootenberg and Abudayyeh cofounded Sherlock Biosciences to commercialize the technology. —EMILY MULLIN

She co-invented a novel immunotherapy for difficult-to-treat cancers.

sing the body’s own immune system to fight cancer has shown promise against several types of tumors, but it’s not always effective. “There’s a whole subset of patients that it really doesn’t work well in,” says Shelley Ackerman. Tumors have to be “hot,” or inflamed, for immunotherapy drugs to work well. Hot tumors are characterized by the presence of a type of immune cell called T cells. Immunotherapy drugs give those T cells a boost, making them better cancer fighters. But many tumors are “cold” and thus evade the immune system. Without any T cells to

U

29

work with, immunotherapy drugs fail against these tumors. As a graduate student at Stanford, Ackerman worked with Edgar Engleman, a professor of medicine and pathology, to develop a therapy aimed at turning cold tumors into hot ones. The approach uses a tumor-targeting antibody chemically attached to an immune-stimulating small-molecule drug that prompts the immune system to recognize and attack the tumor, transforming it into a hot one invaded by tumor-killing T cells. Engleman founded a biotech company, Bolt Biotherapeutics, in 2015 to commercialize the approach; Ackerman joined Bolt in 2018. As a child, Ackerman lost her uncle and a close friend to metastatic cancer within a year, and that experience made her want to keep working on the therapy in hopes that it would one day be used to treat patients. Last year, Bolt began testing its approach in patients with breast, gastric, and other tumors that express a protein known as HER2. The company, which has raised $438 million in funding, is also developing drugs for colorectal, lung, and pancreatic cancers. —EMILY MULLIN

GUTTER CREDIT HERE

30 Change

35 Innovators Under 35

Their creative approaches to innovation are making the world a more equitable place.

EMMA PIERSON AGE

AFFILIATION

30

Cornell University

IN N O V A T I O N

BOB O’CONNOR

She employs AI to get to the roots of health disparities across race, gender, and class. Cornell University computer scientist Emma Pierson uses AI and emerging data science models to reveal how health disparities arise between sexes, races, socioeconomic groups, and other demographic categories. “These are fancy ways of saying I use math to find patterns in large data sets, and the specific types of patterns I’m looking for are attempting to answer sort of old questions in health and social sciences,” she says. The “old questions” she’s investigating range widely in their specifics, but she focuses on uncovering how systemic inequalities in public health come to be, and pointing at

ways to dismantle them. For example, by analyzing mobile-phone data, she recently showed that particular “superspreader” locations were primarily responsible for transmitting covid-19 across populations, and that low-income and minority communities suffered greater risk of exposure. Beyond the pandemic, Pierson’s research team recently examined nearly a decade’s worth of data to show the extent of racial disparities in traffic stops made by police across the US. She analyzed menstrual health data from millions of women in 109 countries to demonstrate how effects on mood and behavior are experienced universally, seeking to destigmatize discussions around women’s health. And she used deep learning to study data on knee pain, revealing that the problem was often poorly measured and even exacerbated in patients from racially underserved groups and lower economic backgrounds. Pierson has made it her mission to see this work break out of the

31

confines of academia. She’s a fairly regular contributor to the New York Times and the Atlantic, offering up a layperson’s account of her work to a large audience. And she engages directly with organizations that can pressure policymakers. Her work on racial disparities in traffic stops ultimately led the Los Angeles Police Department to announce that it would reduce the number of random stops it conducted, and state departments of health leaned on her covid-19 findings to determine how to safely reopen businesses. A self-professed math nerd, Pierson simultaneously earned a bachelor’s degree in physics and a master’s in computer science at Stanford before moving to Oxford as a Rhodes scholar, where she earned a master’s in statistics, afterward earning a doctorate in computer science at Stanford. “I wanted to work on problems that were very concretely tied to people’s lives,” she says. “I think this sense was particularly driven by my own family’s medical history.” In December 2011, Pierson learned she was carrying a genetic mutation that increases her risk of breast and ovarian cancer, and it drove her to focus on work that could make an impact in health care and medicine. Industries like health care deal with insanely large data sets that can really be understood only with the kinds of analytical techniques Pierson has mastered. The data might be the genomes of thousands of people, containing millions upon millions of data points, or medical images from many different patients, representing terabytes of information. AI tools can sort through this data and look for patterns that no human could readily identify. “Computational methods are not optional here,” Pierson says. “They’re the only solution.” —NEEL V. PATEL

SHRIYA SRINIVASAN

Change

AGE

AFFILIATION

27

MIT

I N NOVATION

Her surgical techniques provide a sense of touch to people with prosthetic limbs.

s a child, Shriya Srinivasan, now a postdoctoral researcher in biomedical engineering at MIT, witnessed the challenges of living with prosthetic limbs. A friend had been born with missing limbs and used prosthetics; like amputees, whose nerves have been severed, her brain lacked the important neural signals that enable most people to feel objects, maintain balance, and sense their body’s position in space. Srinivasan has invented two new types of surgical techniques that could soon help people using prosthetic limbs regain their sense of touch.

A

Her first innovation, which she developed as an MIT doctoral student, involves grafting small segments of muscle onto the residual limb; the goal is to enhance the mind’s awareness of limb position and movement. Patients who underwent a version of this procedure in clinical trials have exhibited far greater control over their prostheses— and less pain—than those with traditional amputations. Her second procedure has shown early promise in re-creating touch. It works by fitting a person’s residual limb with flaps of skin from the fingertips or feet, encased by a muscle graft and electrode. A prosthetic arm or hand is then equipped with sensors and a wireless transmitter; when it touches an object, it conveys that sensation to the natural sensors on the grafted skin—which relay it on to the brain. Both techniques can be performed either as part of an amputation procedure or in patients with previous amputations. Ultimately, Srinivasan hopes her work will help make using a prosthetic limb far more like the real thing—while initiating a broader shift in our approach to amputation from a form of salvage to a method of restoring mobility. —JONATHAN W. ROSEN

SRIRAM CHANDRASEKARAN AGE

AFFILIATION

34

University of Michigan

INNOVATION

His AI systems identify better treatments for tuberculosis. Before covid-19, tuberculosis was the most dangerous infection in the world, killing more than 1.5 million people annually. The problem prompted Sriram Chandrasekaran to build AI tools to identify potent drug combinations to treat it. His goal is to boost the effectiveness of existing antibiotics to combat drug resistance among TB patients. Drug-resistant infections occur when people don’t finish their course of treatment or are treated incorrectly. They can also occur when people come in contact with a patient infected with drug-resistant bacteria. While a typical TB treatment regimen lasts six to nine months, a drug-resistant case takes 18 to 24 months to treat. Chandrasekaran wants to drastically reduce this timeline. Curing patients faster could also save thousands of dollars in treatment costs. Chandrasekaran’s systems predict the effectiveness of various drug combinations for TB. “We’ve found some really surprising ones,” he says, including an antipsychotic drug that would enhance the potency of existing antibiotics. He and his team confirmed the results against the TB bacterium in the lab. Many drugs work in the lab but aren’t effective in the body, and Chandrasekaran wanted to make sure his algorithms take this into account. One system he built simulates characteristics of the infection site—for example, how much oxygen it gets or whether amino acids are present, which can affect a drug’s effectiveness. Chandrasekaran’s lab is now identifying promising drug combinations for use in clinical trials of treatment against drug-resistant TB. —EMILY MULLIN

ILLUSTRATION SOURCE: JIMMY DAY (SRINIVASAN); AKHIL KANTIPULY / MICHIGAN ENGINEERING (CHANDRASEKARAN)

32

33

AADEEL AKHTAR AGE

AFFILIATION

34

Psyonic

INNOVATION

His bionic hands combine sensitivity with affordability.

GLEN GYSSLER (DEVICE); SKOT WIEDMANN (AKHTAR)

photo and calligraphy TK

Aadeel Akhtar has developed algorithms that make upper-limb prosthetics much more functional to use. Some send electrical currents to stimulate the nerves so that users can “feel” what their prosthetics are touching; others record the electrical currents caused by muscle contractions, making it possible to control movement. Akhtar has been doing this work for over 10 years, first as a doctoral researcher at the University of Illinois at UrbanaChampaign and then, starting in 2015, as the founder of the roboticlimb startup Psyonic. Akhtar holds four patents on advances in prosthetics that have all gone into Psyonic’s first product, the Ability Hand. The Ability Hand was designed to be controlled by both muscle sensors and Bluetooth (yes, there’s an app!) and provide tactile sensory data to its user, all while withstanding the normal stresses of everyday life (like getting knocked against a table) without cracking. Akhtar’s team of 20 designed with affordability in mind, he says, and built a hand inexpensive enough to be covered by Medicare. This means far more people in the US will be able to afford it. Previously, Akhtar explains, the only insurance that covered bionic hands was associated with veterans’ benefits and worker’s compensation claims, which he estimates cover only about 10% of the need in the United States. Participation from Medicare would make bionic hands available to 75% of individuals in the US who need them. “If Medicare covers it, then other insurers usually follow suit,” Akhtar says. —EILEEN GUO

34

Change

KAYLA LEE AG E

AFFILIATION

30

IBM

I NNOVATION

She’s working to build a more diverse future for quantum computing. n 2018 Kayla Lee joined the enterprise consulting group at IBM, where part of her job is to persuade clients they should be interested in quantum computing. For each client, she says, she needs to figure out the same thing: “How do you make this new technology that is a little bit complicated, and sounds kind of like a science project, relevant to them?” There are parallels between that work and her other project: leading the launch of the IBM-HBCU Quantum Center, a partnership between the company and 23 historically black colleges or universities, which aims to make quantum computing more accessible to Black students and faculty. Lee wants to give Black STEM students and scholars the foundation to excel in this emerging field.

I

Through the partnership, HBCUs have access to IBM’s cloud-based quantum computing service, which undergraduates, graduates, and faculty can use for research. The partnership not only supports Black faculty working on quantum projects but provides funding to “seed these research projects,” says Lee. In one example, IBM recently partnered with the International Society for Optics and Photonics to create a faculty award in quantum optics and photonics specifically for IBM-HBCU Quantum Center members. Lee sees the project as a way to support Black students in an area where they’re grossly underrepresented. In 2017, Black students were awarded just 3% of all bachelor’s degrees in physics in the United States, and only 2% of physics PhDs. What’s more, according to the National Science Foundation, a third of all Black students who have earned doctoral degrees got their bachelor’s degrees at HBCUs, but to date few HBCUs have offered opportunities for students to study or conduct research in quantum information. Lee aims to change that. She wants the Quantum Center to create “clear

opportunities for engagement” and simply show students “what quantum scientists look like.” This is especially important, she says, because quantum computing is such a young field. “We really are at the start of a new model of computation, in the same way that we were at the start of a new model ... back in the ’60s,” she says. “So the questions we’re asking today are:

PETER GARRITANO

They’re looking to the future of quantum computing, energy policy, robotics, and more.

35 Innovators Under 35

35

VARUN SIVARAM AGE

AFFILIATION

32

Biden-Harris administration

INNOVATION

COURTESY PHOTO

Designing new public policies to promote energy innovation.

What do the qubit implementations look like? How do we make less noisy qubits? What does that architecture look like?” But for Lee there’s a further question about quantum computers: “I’m more focused on who gets to use them.” The question of who gets the opportunities to work on this cutting-edge technology will shape the way the field

develops. She points to artificial intelligence, which is already known to be afflicted by problems with racial bias. She says this problem could be exponentially worse in quantum computing, both because of the complexity and inscrutability of the machines and because “there are even fewer representative people” in the field. —EILEEN GUO

Varun Sivaram earned his doctorate researching novel solar materials, but when he graduated in 2013, it wasn’t clear where he could apply those skills in the private sector. Very few startups working on advanced approaches had survived the clean-tech bust of the early 2010s. Commodity silicon solar panels, mostly made in China, dominated the business. That experience prompted him to begin exploring what changes to the innovation system would be required in order to develop better and cheaper clean energy technologies. In studies and books, Sivaram argued that governments must provide far more funding and early policy support for crucial technologies. He also concluded that solar power would still require significant advances to generate an ever larger share of electricity. He worked on these issues directly as chief technology officer at ReNew Power, a large Indian renewable energy company. Now he’s joined the Biden administration as a senior advisor for energy innovation to John Kerry, the US climate czar. Sivaram traveled to India with Kerry, who negotiated a partnership to help that nation achieve its 2030 climate goals. Those include reaching 450 gigawatts of renewable capacity. Sivaram believes that innovation is the most powerful lever the US has to help the rest of the world raise its climate ambitions. Driving down the cost of carbon-free technologies makes it cheaper, easier, and more politically palatable to accelerate the shift to emissions-free energy. Sivaram adds that this is particularly crucial for poorer nations, which often can’t afford to sacrifice economic growth. Without such advances, emissions in emerging economies will soar in coming decades, he warns. —JAMES TEMPLE

A FFILIATION

30

Google

IN N OVATION

Her work helps ensure that fancy AI tools perform in the real world. Emma Beede has an unorthodox claim to technological fame: a study she ran showed that one of her employer’s new technologies was unfit for actual use in the real world. Beede’s study tested a deep-learning algorithm created by Google Health to screen eye images for diabetic retinopathy, a condition caused by high blood sugar that damages the retina and

SARA BERGER AG E

AFFILIATION

33

IBM Research

I NNOVATION

Employing machine learning to make pain management more accessible. Developing smart technology to help patients assess and manage pain is a deeply personal pursuit for Sara Berger, who spent years watching her parents cope with chronic pain and struggle to navigate the medical system. “A lot of the suffering from having chronic pain is about no longer having control over your body and your body’s sensations,” says Berger. “Being able to use digital technologies provides a sense of control and creates more informed conversations with physicians.” A neuroscientist at IBM’s T.J. Watson Research Center, Berger

—NEEL V. PATEL

employs machine learning to quantify long-term pain and help predict ways to relieve it. With wearables and environmental sensors, she can capture metrics including heart rate, sleep patterns, and even the acoustic properties of a patient’s speech, all of which provide data about the person’s pain experience. Those metrics can then be analyzed using machine learning, taking into consideration other factors such as the emotional toll that often results from chronic discomfort, decreased mobility, or lost time with loved ones. What results is a far more holistic and informed assessment and treatment plan than those informed by traditional pain scales, which are prone to bias and oversimplification. “Pain isn’t linear,” says Berger. “Our assessment of it shouldn’t be either.” Many people with chronic conditions, especially women and people of color, feel marginalized by the health care system and experience bias when they seek treatment for pain. “I’m on a mission to transform pain management into an accessible, personalized, and trusted experience for individuals across different socioeconomic backgrounds,” says Berger.  —KATHRYN MILES

PRIYA DONTI AGE

AFFILIATION

28

Carnegie Mellon University

INNOVATION

Finding climatechange solutions via computer science and public policy. Priya Donti knows that a problem as complex and pervasive as climate change won’t be solved by one discipline alone. That’s why she cofounded Climate Change AI, an interdisciplinary organization that brings together academics and industry experts to demonstrate how machine learning can help. Donti’s work combines computer science, engineering, and

AÄRON VAN DEN OORD AGE

AFFILIATION

33

DeepMind

INNOVATION

His AI system creates artificial voices that sound remarkably human. In 2016, Aäron van den Oord had just won an award for his research in image generation when he was struck by an idea. If his technique could learn to predict a two-dimensional sequence of pixels, could it also learn to predict a waveform and thus generate realistic voices? The idea was intriguing but seemed like a long shot. His manager at DeepMind, an AI research subsidiary of Google, gave him two weeks to try it out, saying that if it didn’t work, he should move on to something else.

public policy, and her research focuses on how electric grids can more reliably integrate renewable energy. In 2019, Donti was also a lead author on an influential paper titled “Tackling Climate Change with Machine Learning.” “The tremendous response we received from that paper demonstrated just how many people felt a moral obligation to work on climate change but who also felt like they lacked the necessary community to do that work,” says Donti. Donti, a second-generation Indian American says she’s well aware of the immense burden already felt by some of the planet’s most vulnerable people and recognizes that climate change will only exacerbate those burdens. “We know that the world’s most disadvantaged populations are going to be disproportionately affected by climate change,” says Donti. “Climate Change AI wants to help mitigate that.” —KATHRYN MILES

The results beat everyone’s expectations. Within two weeks, van den Oord had a prototype. Within three months, it was generating more realistic voices than any existing systems. Within another year, Google had begun using WaveNet, as the system came to be called, to generate voices for Google Assistant. WaveNet now powers 51 voices as well as Google’s newest voice assistant, which calls salons and restaurants on behalf of users to book appointments or reserve tables. The results are startlingly realistic. When Google CEO Sundar Pichai first demoed Duplex in 2018, with all its human-like “umms” and “ahs,” it set a new bar for what can be possible when people communicate with machines. While voice assistants need to do more than just generate a synthetic voice—they also need to be able to recognize when someone is talking and understand what’s being said, each of which is a challenge unto itself—researchers have long sought to create the right artificial voice for achieving natural and engaging conversations. “There’s a lot of meaning in a voice,” says van den Oord. —KAREN HAO

KRELL INSTITUTE (DONTI); COURTESY PHOTOS

EMMA BEEDE AG E

makes it difficult to sense light. Beede found that the algorithm, which had performed with over 90% accuracy in the lab, bombed in real-world tests across 11 clinics in Thailand. She found that this was because the algorithm was trained on high-quality eye scans, and when the quality of images taken in the clinic suffered because of factors like poor lighting, the scans were rendered useless. More than 20% of retinal scans were rejected, leaving frustrated patients and their health-care providers looking for more conventional alternatives. Beede thinks such unsatisfying results are a critical example of the need to ensure that AI-powered tools for humans are put through rigorous and meticulous testing before being deployed. “Humans in the real world are complicated, and we should account for that,” she says. “We need to be doing our due diligence to study those downstream effects so that we can mitigate any risk for harm.”

35 Innovators Under 35

KAITLYN SADTLER

AGE

AFFILIATION

31

National Institutes of Health

INNOVATION

Her test was among the first to determine how many people had been infected with covid-19. In early 2020, Kaitlyn Sadtler envisioned a long, slow season getting her lab up and running. Then covid-19 happened. Within weeks, she and her team were among the first to develop an effective antibody assay

NIH/CHIA-CHI CHARLIE CHANG (SADTLER); SINEAD DUBEAU (ELLIS); HECTOR GARCIA-MOLINA (SADIGH)

DORSA SADIGH A GE

AFFILIATION

30

Stanford University

I NNOVATION

She uses simulated environments to teach robots to be better collaborators with people. By developing new ways for computers to anticipate people’s actions, Dorsa Sadigh wants to help pave the way for a future in which human and robots do things like share the roads. In one widely cited paper from 2016, she and her colleagues considered the idealized case of two cars, one driven by a person and another by a computer program. She first had real people drive a car in a video-game-like simulation with several autonomous counterparts that followed preplanned routes. On the basis of people’s behavior in the simulation, she developed a model for how humans drive, which the robot driver then used to

capable of determining how many people had been infected with covid-19, whether they’d shown any symptoms or not. Antibodies tag viruses for destruction and help the body mount an immune response. Those antibodies can linger for months. Existing tests didn’t pinpoint the unique antibodies for the covid-19 virus, leading to false positives among people who had previously been exposed to other coronaviruses. Sadtler and her team at NIH made a highly sensitive antibody test, which uses six different assays to more accurately identify the presence of covid-19 antibodies. Early results published in January confirmed that about 16.8 million Americans had been infected with covid-19 but hadn’t been diagnosed. (Sadtler will update those findings this fall and estimates that as many as one-third of all Americans have been infected with the virus.) The blood test is sensitive enough to determine whether an individual has antibodies from the virus itself or in response to a vaccine, and it can distinguish between variants of the virus as well. It’s simple and cheap to use, making it practical in both rich and poor countries. “This is a global pandemic,” says Sadtler, “which means we need to think globally.” —KATHRYN MILES

devise new strategies for interacting with them. Without ever being explicitly told to do so, it did things like slowly backing up at an intersection, encouraging the “human” to go first. It also developed an attitude, learning how to cut human drivers off or force them to change lanes by swerving toward them. More recently Sadigh and Dylan Losey, at the time her postdoctoral student, taught robots in a simulated setting how to trick humans in a game that involves negotiating who will do more work in carrying plates to a table. “This robot is capable of bringing two plates, but misleads the human to believe that it can only carry one in order to reduce its overall effort,” they wrote in a paper on the work. Teaching robots to be lazy might not sound particularly worthwhile. But Sadigh and Losey are thinking of future applications in which robots might be called upon to help stroke patients in their recovery, for example. Robots, they say, “need to make intelligent decisions that motivate user participation.” —WILL DOUGLAS HEAVEN

37

LEAH ELLIS AGE

AFFILIATION

31

MIT, Sublime Systems

INNOVATION

A new, climate-friendly way to make cement.

Making cement is one of the single largest drivers of climate change, accounting for almost a tenth of global carbon dioxide emissions. Ground-up limestone is typically cooked together with sand, clay, and other materials in kilns that are heated to around 1,500 ˚C (2,700 ˚F). The limestone releases carbon dioxide as it breaks down, as do the fossil fuels that are burned to achieve those temperatures. For every resulting pound of cement, roughly a pound of carbon dioxide escapes into the atmosphere. Leah Ellis came up with a better way. Sublime Systems, a startup she cofounded in March 2020, dissolves pulverized limestone in water and then applies an electric current to trigger a series of chemical reactions. The general idea of using electricity rather than heat to break down limestone has been around for a while, though earlier attempts worked at higher temperatures. Sublime’s apparatus operates at room temperature. Lots of carbon dioxide is still released from the limestone, but it’s much easier to capture and reuse—the gas comes out of one end of the device, mixed with oxygen, while hydrogen gas is released from the other end. This electrochemical reaction produces pure lime, a white powder made of calcium, oxygen, and hydrogen. It can then be cleanly cooked in a kiln with silicon and oxygen to make cement. Ellis and her colleagues are still considering a variety of potential business models. Because they can rely on increasingly cheap electricity from solar or wind farms, Ellis says, they’ll be able to match the prices of standard cement. —JAMES TEMPLE

Change

They’re making crucial advances in fusion power, computing, biosensors, and robotics.

DAVID ROLNICK AG E

AFFILIATION

30

McGill University

I NNOVATION

He’s employing artificial intelligence in the fight against climate change. In 2019, as a postdoctoral researcher at the University of Pennsylvania, David Rolnick was lead author of an influential report that described various ways machine learning could reduce greenhouse-gas emissions and help society adapt to climate change, from predicting energy needs to managing forests to modeling planet-scale weather systems. His coauthors included DeepMind cofounder Demis Hassabis and Turing Award winner Yoshua Bengio. That year, Rolnick was a lead organizer for the first workshops on climate change at three leading AI conferences, and lead organizer of an event on AI at the United Nations Climate Change Conference. “David Rolnick has been hugely influential in convening AI practitioners to

work on climate change,” says Andrew Ng, a cofounder of Google Brain and former chief scientist at Baidu. “By helping shape a vision of how AI could help climate change and tirelessly organizing a community around it, he has catalyzed a significant amount of activity on this important topic.” Rolnick now leads a group at McGill University that uses different AI techniques to attack problems related to climate. For example, data relevant to climate change—records of infrastructure spending or greenhouse-gas emissions or simply weather patterns—varies enormously between countries. And yet climate needs to be understood at a global level. “In the Global South there can be less information on infrastructure,” says Rolnick. “So policymakers may have less to go on when it comes to making decisions about energy requirements or managing coastal flood risk.” Countries also have different regulations about what does and does not get recorded. Germany gathers information on where its solar panels are, for example, but the US does not, so researchers are using machine learning to identify solar panels in the US from

satellite imagery. Machine learning can also be used to forecast energy demand more accurately than is possible with existing techniques, Rolnick says. This allows energy providers to manage their electricity grids more efficiently. Rolnick and his colleagues are trying to come up with new machine-learning techniques that could be applied to the study of climate change as well. For instance, they are building algorithms for transfer learning, which involves training an AI on one set of examples and then transferring what it’s learned to new situations. They are also researching meta-learning, a set of techniques that make AI better at learning from small or incomplete data sets. Rolnick thinks these methods are especially useful for modeling biodiversity because sources of real-world data are so patchy. Rolnick is also involved in projects that combine machine learning with climate models to simulate complex physical and atmospheric processes like cloud formation. The precise means by which clouds form, and how much they reflect or absorb sunlight, is one of the largest sources of uncertainty in existing climate models— partly because simulating clouds in climate models is computationally intensive. Using machine learning to find patterns in when and where clouds form and how reflective they tend to be—without trying to understand the underlying atmospheric chemistry—allows scientists to run models more quickly. Rolnick and his collaborators are convinced AI will be a crucial tool in fighting climate change. All the same, there are growing concerns that machine learning itself is part of the problem. He acknowledges that training today’s largest AI models consumes large amounts of energy, but he points out that this contributes a tiny fraction of global emissions—and that the real climate risks from AI arguably have more to do with its uses in areas such as oil and gas exploration. “I’m much more worried about negative applications of machine learning than I am about its energy use,” he says. —WILL DOUGLAS HEAVEN

GUILLAUME SIMONEAU

38

GUTTER CREDIT HERE

35 Innovators Under 35

39

Change

MOSES NAMARA AG E

29

A FFILIATION

Clemson University

ANNA GOLDIE

AGE

AFFILIATION

27

Google Brain / Stanford University

INNOVATION IN N OVATION

Working to break down the barriers keeping young Black people from careers in AI.

Moses Namara knew two fundamental truths: first, that misuses of AI disproportionately harm Black communities around the world, and second, that Black people are underrepresented in university AI programs. Just 1.8% of students enrolled in computer science PhD programs in the United States were Black in the 2018-2019 school year, and the numbers were only marginally better for master’s students. Namara knew something else, too: that the barriers to entry are often rooted in resources, and that some of those resources were things a mentorship network could provide. “One is just information,” he says. For example: applicants need to know which research opportunities to pursue as undergrads, which university programs and professors best suit their interests, and what resources might be out there to help with the expensive process of actually applying. “If you don’t know where to look for the information, then that’s the number one step that you’re going to fail,” he says. So in 2018 Namara co-created Black in Artificial Intelligence to help students applying to graduate school. Black in AI has mentored 400 applicants, 200 of whom have been accepted to competitive AI programs. It provides an array of resources: mentorship from current PhD students and professors, CV evaluations, and advice on where to apply. Namara now sees the mentorship system evolving to the next logical step: helping Black PhD and master’s students find that first job. —ABBY OHLHEISER

She uses AI to design microchips much more quickly than humans can.

nna Goldie designs computer chips using reinforcement learning, an AI technique that works by repeatedly generating solutions from an artificial neural network. The system then provides feedback to the network, “reinforcing” pathways that lead to successful outcomes and weakening pathways that don’t. Building on this branch of machine learning, which also underlies the most successful methods for teaching computers to play games like chess or Go, has allowed Goldie and her team to speed up the process of chip design.

A

Modern chips are composed of millions or even billions of components. Some perform computations; others store data in short-term memory. Figuring out the best way to place all the components in a chip’s layout can take engineers weeks or even months—they must try to minimize power consumption and area but also maximize performance, all while making sure that traffic between components doesn’t get too congested. Goldie’s AI can, in under six hours, come up with solutions that match—or even outperform—the ones that people were able to develop. In early 2021, Goldie collaborated with Google engineers to produce physical versions of her layouts for Google’s latest artificial-intelligence chip. By using AI to design better hardware faster, she hopes to pave the way for AI advances that further improve and accelerate hardware design, creating a symbiotic loop between hardware and artificial intelligence. “It generates these very strange, alien-looking layouts,” she says. “The chip designers were like: What if it goes wrong?” It didn’t. —WILL DOUGLAS HEAVEN

COURTESY PHOTO (NAMARA); PRIYANKA SHARMA (GOLDIE)

40

35 Innovators Under 35

41

GEORGE BOATENG A GE

AFFILIATIO N

28

SuaCode.ai

I NNOVATION

RAENG/GGIMAGES/FRANCISKOKOROKO (BOATENG); COURTESY PHOTO (SHULAKER)

He built a smartphonebased platform to teach young people to code— and tackle Africa’s IT skills gap in the process. George Boateng’s venture, SuaCode.ai, emerged largely by accident. In 2013, as an undergraduate at Dartmouth College, he’d teamed up with a group of friends to launch a summer innovation boot camp for high school students in their native Ghana. When the donated laptops they’d gotten for the course broke down a few years later, they were in a fix: only a quarter of the students had laptops of their own, and buying more would overwhelm their budget. All the students, however, had smartphones—so Boateng and his colleagues redesigned the coding module to fit a five-inch screen. The experience went so well that it hatched a spinoff: in 2018, Boateng and cofounder Victor Kumbol ran their first pilot of SuaCode, an eight-week smartphone-based course. The course, which teaches Processing, a Java-based language, now has more than 600 graduates from two dozen countries. Boateng, currently a doctoral candidate in applied machine learning at ETH Zurich, also engineered an English- and French-speaking AI-powered teaching assistant named Kwame—a nod to Ghana’s first president, Kwame Nkrumah. “His pan-Africanist vision resonates with our goal of empowering youth across the continent,” Boateng says. Boateng’s hope is that the automated nature of the course will help it reach far more students—providing early exposure to coding that will serve as a bedrock for further education and ultimately help them land well-paying jobs in tech. —JONATHAN W. ROSEN

MAX SHULAKER

AGE

AFFILIATION

33

MIT

INNOVATION

His work with carbon nanotubes could lead to the next generation of computers.

Max Shulaker has built the world’s first functional computer using carbon nanotubes, and he has also designed systems that combine computing, memory, and sensing directly on top of one another on a single chip. Together, these new technologies could increase energy efficiency in computers up to 1,000-fold and make possible a whole world of new devices like low-cost medical sensors. Carbon nanotubes are “basically a straw that is one carbon atom thin,” says Shulaker. For 20 years, researchers have talked about using them to replace traditional silicon

chips. But turning carbon-nanotube transistors and wires into actual devices has proved difficult, and Shulaker has solved several problems to make them work. He developed a way to remove poorly formed carbon nanotubes during production, devised new processes to create wafers of nanotube-based transistors using regular industrial fabrication plants, and invented a new design ensuring that chips built with a certain number of defective tubes are guaranteed to work. These breakthroughs represent a significant step toward next-generation computer systems far more energy efficient than anything built to date. Shulaker’s drive led him to work on another feat: monolithic, three-dimensional nanosystems. These fuse microprocessor, memory, and additional functional layers directly on top of one another using carbon nanotubes. Traditional designs have the microchip and memory on separate chips connected by wires. But moving massive quantities of data between those chips leads to slowdowns and wasted energy—a problem known in the industry as “the memory wall.” Shulaker’s 3D nanosystems solve it. —RUSS JUSKALIAN

JINXING ZHENG

Change

AGE

AFFI L I A T I O N

34

Institute of Plasma Physics, Chinese Academy of Sciences

MARC MISKIN

AGE

AFFILIATION

34

University of Pennsylvania

I N NOVATION

INNOVATION

He created new physics models for controlling fusion reactions and hot plasma.

He figured out how to give motion to microscopic robots.

inxing Zheng has devised better ways to model the use of powerful magnets for controlling plasma at extreme temperatures, a major advance for fusion-based energy. Zheng’s work is helping China leapfrog the rest of the world and design the largest fusion reactor to date, called the China Fusion Engineering Test Reactor. CFETR is expected to finish construction and go online before 2035, though it may take five to 10 years to reach full power. Fusion reactors, based on the energy released when atoms are combined, have great potential for creating clean energy and are inherently safer than existing nuclear power based on fission reactions. But no one has built a practical one, in part because it’s so challenging to contain the necessary plasma, which can reach temperatures of hundreds of millions of degrees Celsius. Zheng’s innovation amounts to having discovered new theoretical models for understanding how multiple large superconducting magnets can rapidly change their magnetic fields to keep plasma in one place while fusion reactions occur. In 2018, with the help of Zheng’s models, a fusion reactor in Hefei, China, called the Experimental Advanced Superconducting Tokamak—nicknamed “the artificial sun”—controlled plasma at a record temperature of 50 million ˚C for 102 seconds. China’s future CFETR is intended to operate at over 1 gigawatt of power sometime in the 2030s. That’s double the power of ITER, a fusion reactor currently being completed in the south of France with cooperation from countries around the world.

arc Miskin has given life to a technology that’s eluded the world’s top nanoscientists for decades: robots too small to see. Miskin’s tiny bots piggyback on more than 50 years of electronics innovation, making it possible to build silicon chips smaller than the width of a human hair. The challenge was getting these circuits, which function as the robots’ brains, to move: previous approaches to connecting them to a pair of microscopic legs required too much voltage to work at such a tiny scale. His technique fabricates legs from sheets of platinum a dozen or so atoms thick, capped on one side with an even smaller layer of titanium. When activated with a current—generated by solar cells attached to the robot brain—the platinum bends, causing the bot to march forward. Miskin’s initial prototype, which he developed as a postdoctoral researcher at Cornell University, requires only one-fifth of a volt to move and measures just 40 by 40 microns—smaller than many single-celled microorganisms. It’s recognized by Guinness World Records as the smallest ever walking robot, and a million of them at a time can be fabricated on a single 10-centimeter wafer. For now, Miskin’s robot does little more than prance under a microscope, but his lab at the University of Pennsylvania, where he’s a professor of electrical and systems engineering, is fabricating limbs for a “smart bot” with programmable memory, developed with researchers at the University of Michigan. In the longer term, Miskin envisions tiny bots being used to engineer new materials, rid crops of pests, or even act as microscopic surgeons, programmed to eliminate cancer cells one by one. —JONATHAN W. ROSEN

J

—RUSS JUSKALIAN

M

COURTESY PHOTO (ZHENG); UPENN (MISKIN)

42

35 Innovators Under 35

ADNAN MEHONIC A GE

AFFILIATIO N

34

University College London and Intrinsic

NAKO NAKATSUKA

AGE

AFFILIATION

31

ETH Zurich

INNOVATION I NNOVATION

COURTESY PHOTO (NEHONIC); ALEXANDER TANNO (NAKATSUKA)

Memristors can be a new and more efficient building block of modern computers.

Memristors are a novel type of electric circuit element that were first theorized to exist in 1971. In 2008, researchers at Hewlett-Packard identified them for the first time, in nanodevices made from titanium dioxide, but the technology has not replaced flash memory as initially predicted. Resistors are elements of a circuit that control the flow of electric current. A memristor, as its name suggests, is like an adjustable resistor with memory. Turn the power off, and a memristor “remembers” the most recent resistance it had. That holds the promise of faster, more efficient chips that integrate memory with logic. Adnan Mehonic is developing memristors out of silicon oxide, the material most commonly used in computer chips. His most straightforward goal is to make dense, lowpower, high-speed memory. More ambitiously, he is using the physics of memristors to implement in-memory computing and brain-like functionalities for future neuromorphic systems. Among other applications, memristors could greatly improve the energy efficiency of AI systems, proponents say. “Crossbar arrays” of memristors, says Mehonic, could perform deep-learning tasks using one-500th as much energy as current hardware. A startup he cofounded concluded a $1.9 million financing round in March. —PATRICK HOWELL O’NEILL

Her miniature biosensors could give scientists better insight into depression and dementia.

Nako Nakatsuka is building tiny sensors that can detect chemical changes in the brain and other parts of the body more precisely than ever before. Scientists can use such information to help them understand and treat conditions like depression and dementia. Compared with earlier sensors, Nakatsuka’s are better at differentiating between structurally similar chemicals, like neurotransmitters and their precursors and metabolites. For now, her sensors are used to take measurements on samples in the lab, but the technology is being refined to work directly in the body and on a wider range of chemicals.

43

Nakatsuka built her sensors using molecules called aptamers, which can be designed to have strong affinity for specific targets. She first used an aptamer constructed from DNA that changes its shape in the presence of serotonin, a neurotransmitter that plays an important role in bodily functions like sleep and appetite, and in conditions like depression and obsessive-compulsive disorder. Later she developed a way to attach the aptamer to the opening of a tiny pipette, just 10 nanometers in diameter, hooked up to an electrical circuit. As the aptamer changes its shape in the presence of serotonin, it alters the electrical current. The sensor can measure samples in brain fluid or tissue, or potentially directly next to individual neurons in a lab dish or in the brain. “This might help us to better understand Parkinson’s and other diseases,” Nakatsuka says. Her sensors could be used to monitor how neurons in or from a patient with such a disease function in real time. And since aptamers can be used in all sorts of tests, Nakatsuka’s technology could lead to faster, cheaper, and more accurate detection for all kinds of medical conditions and infections. —RUSS JUSKALIAN

GUTTER CREDIT HERE

44 Change

35 Innovators Under 35

They’re building companies aimed at solving some of the world’s most pressing problems.

JANICE CHEN AGE

AFFILIATION

30

Mammoth Biosciences

IN N O V A T I O N

CHRISTIE HEMM KLOK

She’s using CRISPR to make new diagnostic tests. Janice Chen was jumping into an Uber, cramming in equipment the size of a microwave. At the time a PhD student at the University of California, Berkeley, Chen had been invited to a lab to look for the human papillomavirus in hospital medical samples using a new technique she had created. Soon enough, bingo. Her test, which uses the gene-editing tool CRISPR, was able to spot the virus nearly every time, offering a new way to test for germs. She and several other students, along with Jennifer Doudna, the co-discoverer of CRISPR, cofounded a company with plans to develop a new generation of testing instruments. They called it Mammoth Biosciences.

The diagnostics business isn’t easy to break into: a few companies with well-established technologies dominate. Chen is now in charge of a team of 40 as the chief technology officer of Mammoth. She says she leans on her experience playing chess competitively as a teenager, when she learned how to build a position move by move, make meaningful sacrifices, and get inside competitors’ minds. Chen grew up in Salt Lake City. Her parents were immigrants from China. Her brother is a world champion and Olympic medalist in figure skating. When she was growing up, she says, her parents urged her and her siblings to “find your passion and do your best to move it forward in a significant way.”

45

Chen crammed for years studying chess moves but ended up finding her real interest while moonlighting at her father’s biotech supply company. That’s where she first copied genes and engineered a bacterium. Then, at Princeton University, she had a chance to help out on a large ongoing project to assemble the entire genome of a yeast cell from DNA parts. As an undergraduate, she did menial lab tasks. Still, here was life being engineered from the ground up. And she was part of it. For her PhD, Chen landed a spot in Doudna’s Berkeley lab, where CRISPR editing had been co-developed in 2012. Chen joined a fast-paced hunt to discover and understand even more types of DNA editors and harness them for new uses. She demonstrated a way that a particular gene-editing enzyme could be used as a diagnostic test. Her test could find a specific sequence of viral DNA in a sample, cut it, and unleash a fluorescent signal that would report the result. That looked useful enough for infectious-disease testing to try to commercialize it, which is what led her to cofound Mammoth in 2017. Then came covid-19. When the rollout of the standard tests stumbled in the spring of 2020, the US Food and Drug Administration gave Mammoth and dozens of other smaller companies an emergency green light to sell their tests for the virus. It was a crisis, and that meant the purse strings were loosened too. Mammoth has won $30 million in government funding since the pandemic began. As of May 2021, Mammoth was preparing to commercialize the company’s first product, kits that public health labs can use to run 1,500 simultaneous covid-19 tests with less human intervention than existing ones require. —ANTONIO REGALADO

JACOB BECRAFT

Change

AGE

AFFILIATION

30

Strand Therapeutics

I N NOVATION

He runs a company that’s figuring out the next steps for messenger RNA.

afe and effective covid-19 vaccines have finally provided an exit to the pandemic. The most innovative of these vaccines use messenger RNA—strings of nucleic acids— to instruct cells to make a protein found in the virus, causing the body to produce antibodies against it. Now scientists are eyeing all sorts of other potential uses for this underlying technology. “With covid-19, we have gone from mRNA being potentially useful to saying we know it works in humans,” says Jacob Becraft. He runs a startup called

S

Strand Therapeutics, which is working on the next step for mRNA—ways to “program” the molecules to do additional useful tricks, like turn on only in specific cell types, at specific times, or automatically copy themselves so as to strengthen their effects. Though mRNA’s effects are temporary (because it’s an unstable molecule), using it is in many ways simpler, safer, and faster than trying to change the genome of a cell. One idea the company is pursuing is to use injections of mRNA to instruct the body’s immune cells to attack cancers of the skin and breast. Becraft grew up in a small farming community in central Illinois, famous as home to a federal lab that discovered how to mass-produce penicillin during World War II. In high school, he says, he didn’t have patience for pictures of cells, with their labeled parts. “It wasn’t until college that I got exposed to biology as a machine, not just a list of things to memorize,” he says. “But when someone tells me how a system works, I get it. I can imagine it.” —ANTONIO REGALADO

TAMMY HSU AGE

AFFILIATION

30

Huue

INNOVATION

Her new dye can make one of the world’s most common types of clothing more environmentally friendly. Many consumers don’t realize that indigo, the signature color of denim, requires synthetic chemicals like formaldehyde and cyanide, which can be harmful to workers and can sometimes contaminate local water sources. Given that jeans are one of the most ubiquitous clothing items in the world, this is a huge environmental problem. Tammy Hsu, the chief scientific officer of Huue, worked with colleagues to study how color is made in nature and program microbes to enzymatically produce the shade they wanted. The result is a sustainable solution that doesn’t rely on harmful processes or chemicals. Now the challenge is to make the natural dye as cheap to use as the synthetics the industry relies upon. “The chemical industry has had 100 years to hone their process and make it cost efficient,” Hsu says. “We were founded two years ago. We’re trying to catch up with that. That’s one of our biggest goals, to drive down the price of our process.” Huue is on track to release its indigo dye next year. Next up for Hsu is figuring out how to coax microbes to produce a range of different dyes. “We’re trying to provide the fashion industry with an alternative way,” she says. —TANYA BASU

ILLUSTRATION SOURCE: TAMMY BECRAFT; ABOLFAZL AGHANOURI

46

35 Innovators Under 35

Judges Pieter Abbeel Professor, UC Berkeley; Director, Berkeley Robot Learning Lab

SARA SPANGELO

AGE

AFFILIATION

34

Swarm Technologies

I NNOVATION

Her tiny satellites could bring connectivity to the remotest places on Earth.

Animashree Anandkumar Bren Professor, Caltech, and Director of AI Research, Nvidia David Berry CEO, Valo Health; General Partner, Flagship Pioneering Ed Boyden Y. Eva Tan Professor in Neurotechnology, MIT/HHMI

ILLUSTRATION SOURCE: COURTESY PHOTO

Meredith Broussard Associate Professor, NYU

Sara Spangelo didn’t quite make it as an astronaut. But four years after an unsuccessful tryout with Canada’s space agency, she’s achieved her own space milestone: unveiling the world’s lowest-cost always-available satellite communications network. Spangelo, who holds a PhD in aerospace engineering from the University of Michigan, is CEO of Swarm Technologies, which seeks to provide affordable data services for devices anywhere on Earth. Today, nearly 90% of the planet’s surface, including oceans, deserts, and polar regions, lacks internet access. Connecting via satellite has long been cost-prohibitive, because satellite networks typically cost billions of dollars to deploy and maintain. The key to lowering costs was to bring down size: Swarm’s satellites, roughly the size of a slice of French toast, are the smallest two-way communication devices in orbit today. Because they’re so compact, they can hitch rides on commercial rockets for bargain prices: total launch costs for Swarm’s full constellation of 150 satellites, which the company will finish placing in low Earth orbit by the end of 2021, will run less than $3 million. Swarm’s data connection, which uses the VHF radio spectrum, won’t enable seafarers to stream Netflix: its current transfer rate of 1 kilobit per second is similar to 1990s dial-up. Swarm’s niche, rather, is giving customers the ability to transmit small yet highly useful packets of information from the world’s most far-flung places. This enables them to remotely monitor water supplies, detect leaks in pipelines, measure soil contents, track wildlife, or guarantee the temperature of vaccines in cold-chain transport. —JONATHAN W. ROSEN

Yet-Ming Chiang Kyocera Professor of Materials Science and Engineering, MIT James Collins Termeer Professor, MIT John Dabiri Centennial Professor of Aeronautics and Mechanical Engineering, Caltech Gozde Durmus Assistant Professor, Stanford University Oren Etzioni CEO, Allen Institute for AI; Professor Emeritus, Computer Science, University of Washington David Fattal Founder and CEO, Leia, Inc. Javier Garcia Martinez Professor of Inorganic Chemistry, University of Alicante, Spain; President-elect, IUPAC Julia R. Greer Mettler Professor of Materials, Mechanics, and Medical Engineering, Caltech Zhen Gu Professor, Zhejiang University Ilan Gur Founder and CEO, Activate.org Marc Lajoie CEO, Outpace Bio

47

Hao Li CEO & Cofounder, Pinscreen; Distinguished Fellow, UC Berkeley Zackary Lipton Assistant Professor of Operations Research and Machine Learning, Carnegie Mellon University Zlatko Minev Quantum physicist, IBM Quantum; Founder, Open Labs Andrew Ng Founder, DeepLearning.AI; CEO, Landing AI; General Partner, AI Fund Nicole Paulk Assistant Professor, University of California San Francisco Deb Raji Fellow, Mozilla Foundation John Rogers Simpson/Querrey Professor of Materials Science and Engineering, Biomedical Engineering, and Neurological Surgery, Northwestern University Nabiha Saklayen CEO and Cofounder, Cellino Rachel Sheinbein Venture Partner, Gratitude Railroad Mona Sloane Sociologist, New York University Cyrus Wadia Head of WW Product Sustainability, Amazon Jennifer West Dean of Engineering and Applied Sciences, University of Virginia Minmin Yen CEO and Cofounder, PhagePro Jackie Ying A*STAR Senior Fellow and Director, NanoBio Lab Alice Zhang CEO, Verge Genomics Ben Zhao Neubauer Professor of Computer Science, University of Chicago

48

Change

THE FIGHT TO RECLAIM AI

How a small group of researchers at the margins built a movement to wrest the world’s most powerful emerging technology from Big Tech’s control.

By Karen Hao

ILLUSTRATION BY RICARDO SANTOS

49

Radical AI

T

Clockwise from top left: Raphael Gontijo Lopes, Deborah Raji, Rediet Abebe, Timnit Gebru, Joy Buolamwini, William Agnew.

Timnit Gebru never thought a scientific paper would cause her so much trouble. In 2020, as the co-lead of Google’s ethical AI team, Gebru had reached out to Emily Bender, a linguistics professor at the University of Washington, and asked to collaborate on research about the troubling direction of artificial intelligence. Gebru wanted to identify the risks posed by large language models, one of the most stunning recent breakthroughs in AI research. The models are algorithms trained on staggering amounts of text. Under the right conditions, they can compose what look like convincing passages of prose. For a few years, tech companies had been racing to build bigger versions and integrate them into consumer products. Google, which invented the technique, was already using one to improve the relevance of search results. OpenAI announced the largest one, called GPT-3, in June 2020 and licensed it exclusively to Microsoft a few months later. Gebru worried about how fast the technology was being deployed. In the paper she wound up writing with Bender and five others, she detailed the possible dangers. The models were enormously costly to create—both environmentally (they require huge amounts of computational power) and financially; they were often trained on the toxic and abusive language of the internet; and they’d come to dominate research in language AI, elbowing out promising alternatives.

50 Like other existing AI techniques, the models don’t actually understand language. But because they can manipulate it to retrieve text-based information for users or generate natural conversation, they can be packaged into products and services that make tech companies lots of money. That November, Gebru submitted the paper to a conference. Soon after, Google executives asked her to retract it, and when she refused, they fired her. Two months later, they also fired her coauthor Margaret Mitchell, the other leader of the ethical AI team. The dismantling of that team sparked one of the largest controversies within the AI world in recent memory. Defenders of Google argued that the company has the right to supervise its own researchers. But for many others, it solidified fears about the degree of control that tech giants now have over the field. Big Tech is now the primary employer and funder of AI researchers, including, somewhat ironically, many of those who assess its social impacts. Among the world’s richest and most powerful companies, Google, Facebook, Amazon, Microsoft, and Apple have made AI core parts of their business. Advances over the last decade, particularly in an AI technique called deep learning, have allowed them to monitor users’ behavior; recommend news, information, and products to them; and most of all, target them with ads. Last year Google’s advertising apparatus generated over $140 billion in revenue. Facebook’s generated $84 billion. T h e c o m p a n i e s h ave invested heavily in the technology that has brought them such vast wealth. Google’s parent company, Alphabet, acquired the London-based AI lab DeepMind for $600 million

Change in 2014 and spends hundreds of millions a year to support its research. Microsoft signed a $1 billion deal with OpenAI in 2019 for commercialization rights to its algorithms. At the same time, tech giants have become large investors in university-based AI research, heavily influencing its scientific priorities. Over the years, more and more ambitious scientists have transitioned to working for tech giants full time or adopted a dual affiliation. From 2018 to 2019, 58% of the most cited papers at the top two AI conferences had at least one author affiliated with a tech giant, compared with only 11% a decade earlier, according to a study by researchers in the Radical AI Network, a group that seeks to challenge power dynamics in AI. The problem is that the corporate agenda for AI has focused on techniques with commercial potential, largely ignoring research that could help address challenges like economic inequality and climate change. In fact, it has made these challenges worse. The drive to automate tasks has cost jobs and led to the rise of tedious labor like data cleaning and content moderation. The push to create ever larger models has caused AI’s energy consumption to explode. Deep learning has also created a culture in which our data is constantly scraped, often without consent, to train products like facial recognition systems. And recommendation algorithms have exacerbated political polarization, while large language models have failed to clean up misinformation. It’s this situation that Gebru and a growing movement of like-minded scholars want to change. Over the last five years, they’ve sought to shift the field’s priorities away from simply enriching tech companies, by expanding who gets

to participate in developing the technology. Their goal is not only to mitigate the harms caused by existing systems but to create a new, more equitable and democratic AI.

“HELLO FROM TIMNIT” In December 2015, Gebru sat down to pen an open letter. Halfway through her PhD at Stanford, she’d attended the Neural Information Processing Systems conference, the largest annual AI research gathering. Of the more than 3,700 researchers there, Gebru counted only five who were Black. Once a small meeting about a niche academic subject, NeurIPS (as it’s now known) was quickly becoming the biggest annual AI job bonanza. The world’s wealthiest companies were coming to show off demos, throw extravagant parties, and write hefty checks for the rarest people in Silicon Valley: skillful AI researchers. That year Elon Musk arrived to announce the nonprofit venture OpenAI. He, Y Combinator’s then president Sam Altman, and PayPal cofounder Peter Thiel had put up $1 billion to solve what they believed to be an existential problem: the prospect that a superintelligence could one day take over the world. Their solution: build an even better superintelligence. Of the 14 advisors or technical team members he anointed, 11 were white men. While Musk was being lionized, Gebru was dealing with humiliation and harassment. At a conference party, a group of drunk guys in Google Research T-shirts circled her and subjected her to unwanted hugs, a kiss on the cheek, and a photo. Gebru typed out a scathing critique of what she had observed: the spectacle, the cult-like worship of AI celebrities, and most of all, the overwhelming homogeneity. This

boy’s club culture, she wrote, had already pushed talented women out of the field. It was also leading the entire community toward a dangerously narrow conception of artificial intelligence and its impact on the world. Google had already deployed a computer-vision algorithm that classified Black people as gorillas, she noted. And the increasing sophistication of unmanned drones was putting the US military on a path toward lethal autonomous weapons. But there was no mention of these issues in Musk’s grand plan to stop AI from taking over the world in some theoretical future scenario. “We don’t have to project into the future to see AI’s potential adverse effects,” Gebru wrote. “It is already happening.” Gebru never published her reflection. But she realized that something needed to change. On January 28, 2016, she sent an email with the subject line “Hello from Timnit” to five other Black AI researchers. “I’ve always been sad by the lack of color in AI,” she wrote. “But now I have seen 5 of you :) and thought that it would be cool if we started a black in AI group or at least know of each other.” The email prompted a discussion. What was it about being Black that informed their research? For Gebru, her work was very much a product of her identity; for others, it was not. But after meeting they agreed: If AI was going to play a bigger role in society, they needed more Black researchers. Otherwise, the field would produce weaker science—and its adverse consequences could get far worse.

A PROFIT-DRIVEN AGENDA As Black in AI was just beginning to coalesce, AI was hitting its commercial stride. That year, 2016, tech giants spent an estimated $20 to $30 billion

Radical AI

“WE DON’T HAVE TO PROJECT INTO THE FUTURE TO SEE AI’S POTENTIAL ADVERSE EFFECTS.” on developing the technology, according to the McKinsey Global Institute. He a t e d by c o r p o ra t e investment, the field warped. Thousands more researchers began studying AI, but they mostly wanted to work on deep-learning algorithms, such as the ones behind large language models. “As a young PhD student who wants to get a job at a tech company, you realize that tech companies are all about deep learning,” says Suresh Venkatasubramanian, a computer science professor who now serves at the White House Office of Science and Technology Policy. “So you shift all your research to deep learning. Then the next PhD student coming in looks around and says, ‘Everyone’s doing deep learning. I should probably do it too.’” But deep learning isn’t the only technique in the field. Before its boom, there was a different AI approach known as symbolic reasoning. Whereas deep learning uses massive amounts of data to teach algorithms about meaningful relationships in information, symbolic reasoning focuses on explicitly encoding knowledge and logic based on human expertise. Some researchers now believe those techniques

should be combined. The hybrid approach would make AI more efficient in its use of data and energy, and give it the knowledge and reasoning abilities of an expert as well as the capacity to update itself with new information. But companies have little incentive to explore alternative approaches when the surest way to maximize their profits is to build ever bigger models. In their paper, Gebru and Bender alluded to a basic cost of this tendency to stick with deep learning: the more advanced AI systems we need are not being developed, and similar problems keep recurring. Facebook, for example, relies heavily on large language models for automated content moderation. But without really understanding the meaning behind text, those models often fail. They regularly take down innocuous posts while giving hate speech and misinformation a pass. AI-based facial recognition systems suffer from the same issue. They’re trained on massive amounts of data but see only pixel patterns—they do not have a grasp of visual concepts like eyes, mouths, and noses. That can trip these systems up when they’re used on individuals with a different skin tone from the people

51

they were shown during training. Nonetheless, Amazon and other companies have sold these systems to law enforcement. In the US, they have caused three known cases of police jailing the wrong person—all Black men—in the last year. For years, many in the AI community largely acquiesced to Big Tech’s role in shaping the development and impact of these technologies. While some expressed discomfort with the corporate takeover, many more welcomed the industry’s deep well of funding. But as the shortcomings of today’s AI have become more evident—both its failure to solve social problems and the mounting examples that it can exacerbate them—faith in Big Tech has weakened. Google’s ousting of Gebru and Mitchell further stoked the discussion by revealing just how much companies will prioritize profit over self-policing. In the immediate aftermath, over 2,600 Google employees and 4,300 others signed a petition denouncing Gebru’s dismissal as “unprecedented research censorship.” Half a year later, research groups are still rejecting the company’s funding, researchers refuse to participate in its conference workshops, and employees are leaving in protest. Unlike five years ago, when Gebru began raising these questions, there’s now a well-established movement questioning what AI should be and who it should serve. This isn’t a coincidence. It’s very much a product of Gebru’s own initiative, which began with the simple act of inviting more Black researchers into the field.

IT TAKES A CONFERENCE In December 2017, the new Black in AI group hosted its first workshop at NeurIPS. While

organizing the workshop, Gebru approached Joy Buolamwini, an MIT Media Lab researcher who was studying commercial facial recognition systems for possible bias. Buolamwini had begun testing these systems after one failed to detect her own face unless she donned a white mask. She submitted her preliminary results to the workshop. Deborah Raji, then an undergraduate researcher, was another early participant. Raji was appalled by the culture she’d observed at NeurIPS. The workshop became her respite. “To go from four or five days of that to a full day of people that look like me talking about succeeding in this space—it was such important encouragement for me,” she says. Buolamwini, Raji, and Gebru would go on to work together on a pair of groundbreaking studies about discriminatory computer-vision systems. Buolamwini and Gebru coauthored Gender Shades, which showed that the facial recognition systems sold by Microsoft, IBM, and Chinese tech giant Megvii had remarkably high failure rates on Black women despite near-perfect performance on white men. Raji and Buolamwini then collaborated on a follow-up called Actionable Auditing, which found the same to be true for Amazon’s Rekognition. In 2020, Amazon would agree to a one-year moratorium on police sales of its product, in part because of that work. At the very first Black in AI workshop, though, these successes were distant possibilities. There was no agenda other than to build community and produce research based on their sorely lacking perspectives. Many onlookers didn’t understand why such a group needed to exist.

52

Change

“THESE CHANGES THAT WE’RE FIGHTING FOR—IT’S NOT JUST FOR MARGINALIZED GROUPS.”

recognition, and Massachusetts now requires police to get a judge’s permission to use it. Both the US and the European Commission have proposed additional regulation. “First we had to just be there,” says Gebru. “And at some point, what Black in AI says starts to become important. And what all of these groups together say becomes important. You have to listen to us now.”

FOLLOW THE MONEY Gebru remembers dismissive comments from some in the AI community. But for others, Black in AI pointed a new way forward. This was true for William Agnew and Raphael Gontijo Lopes, both queer men conducting research in computer science, who realized they could form a Queer in AI group. (Other groups that took shape include Latinx in AI, {Dis}Ability in AI, and Muslim in ML.) For Agnew, in particular, having such a community felt like an urgent need. “It was hard to even imagine myself having a happy life,” he says, reflecting on the lack of queer role models in the field. “There’s Turing, but he committed suicide. So that’s depressing. And the queer part of him is just ignored.” Not all affinity group members see a connection between their identity and their research. Still, each group has established particular expertise. Black in AI has become the intellectual center for exposing algorithmic discrimination, critiquing surveillance, and developing data-efficient AI techniques. Queer in AI has become a center for contesting the ways algorithms infringe on people’s privacy and classify them into bounded categories by default.

Venkatasubramanian and Gebru also helped create the Fairness, Accountability, and Transparency (FAccT) conference to create a forum for research on the social and political implications of AI. Ideas and draft papers discussed at NeurIPS affinity group workshops often become the basis for papers published at FAccT, which then showcases that research to broader audiences. It was after Buolamwini presented at the first Black in AI workshop, for example, that FAccT published Gender Shades. Along with Actionable Auditing, it then fueled several major education and advocacy campaigns to limit government use of facial recognition. When Amazon attempted to undermine the legitimacy of Buolamwini’s and Raji’s research, dozens of AI researchers and civil society organizations banded together to defend them, foreshadowing what they would later do for Gebru. Those efforts eventually contributed to Amazon’s moratorium, which in May the company announced it would extend indefinitely. The research also set off a cascade of regulation. More than a dozen cities have banned police use of facial

After Gebru and Mitchell’s firing, the field is grappling anew with an age-old question: Is it possible to change the status quo while working from within? Gebru still believes working with tech giants is the best way to identify the problems. But she also believes that corporate researchers need stronger legal protections. If they see risky practices, they should be able to publicly share their observations without jeopardizing their careers. Then there’s the question of funding. Many researchers want more investment from the US government to support work that is critical of commercial AI development and advances the public welfare. Last year, it committed a measly $1 billion to non-defenserelated AI research. The Biden administration is now asking Congress to invest an additional $180 billion in emerging technologies, with AI as a top priority. Such funding could help people like Rediet Abebe, an assistant professor of computer science at the University of California, Berkeley. Abebe came into AI with ideas of using it to advance social equity. But when she started her PhD at Cornell, no one was focused on doing such research. In the fall of 2016, as a PhD student, she began a small

Cornell reading group with a fellow graduate student to study topics like housing instability, health-care access, and inequality. She then embarked on a new project to see whether her computational skills could support efforts to alleviate poverty. Eventually, she found the Poverty Tracker study, a detailed data set on the financial shocks—unexpected expenses like medical bills or parking tickets—experienced by more than 2,000 New York families. Over many conversations with the study’s authors, social workers, and nonprofits serving marginalized communities, she learned about their needs and told them how she could help. Abebe then developed a model that showed how the frequency and type of shocks affected a family’s economic status. Five years later, the project is still ongoing. She’s now collaborating with nonprofits to improve her model and working with policymakers through the California Policy Lab to use it as a tool for preventing homelessness. Her reading group has also since grown into a 2,000-person community and is holding its inaugural conference later this year. Abebe sees it as a way to incentivize more researchers to flip the norms of AI. While traditional computer science conferences emphasize advancing computational techniques for the sake of doing so, the new one will publish work that first seeks to deeply understand a social issue. The work is no less technical, but it builds the foundation for more socially meaningful AI to emerge. “These changes that we’re fighting for—it’s not just for marginalized groups,” she says. “It’s actually for everyone.” Q Karen Hao is a senior editor for AI at MIT Technology Review.

CHANGE PRESENTS OPPORTUNITIES Make decisions that make a difference with the Financial Times Read more at ft.com/newagenda

54

A year into the covid-19 pandemic, Apple commemorated the growing array of devices featuring its custom M1 chip with great fanfare, including a “Mission Implausible” ad on TV featuring a young man running across the rooftops of its “spaceship” campus in Cupertino and infiltrating the facility to “steal” the breakthrough microprocessor from a MacBook and place it inside an iPad Pro. Apple’s custom-designed chip is the latest triumph for Moore’s Law, the observation turned self-fulfilling prophecy that chipmakers can double the number of transistors on a chip every few years. The M1 packs 16 billion transistors on a microprocessor the size of a large postage stamp. It’s a marvel of today’s semiconductor manufacturing prowess. But even as Apple celebrated the M1, the world was facing an economically devastating shortage of microchips, particularly the relatively cheap ones that make many of today’s technologies possible. Automakers have been shutting down assembly lines and laying off workers because they can’t get enough $1 chips. Manufacturers have resorted to building vehicles without the chips necessary for navigation systems, digital rear-view mirrors, display touch screens, and fuel management systems. Overall, the global automotive industry could lose more than $110 billion to the shortage in 2021. Production has also slowed for smartphones, laptops, video-game consoles, TVs, and even smart appliances, all because of the lack of cheap microchips. Their use is so essential and so widespread that some observers think the chip crisis could threaten the global economic recovery from the pandemic. The global shortage is shining a harsh spotlight on the semiconductor industry’s ability to deliver cheaper and more powerful microchips. The longstanding promise of chips with ever more capabilities inspired engineers, programmers, and product designers to create generations of new products and services. Moore’s Law has been more than just a road map for the semiconductor industry—it has governed technological change over the last half-century. Now that promise of more computing power everywhere is crumpling, but not because chipmakers have

THE GREAT CHIP DDIV A shortage of microchips is threatening to slow digital innovation inspired by the promise of ever faster, cheaper computing power.

By Jeremy Hsu

COURTESY OF ASML

55

This tool, called an extreme ultraviolet lithography machine, allows manufacturers to make leading-edge chips with features a few nanometers in size.

56

Change

finally run up against the physical limits of technology to make ever smaller transistors. Instead, the growing costs of sustaining Moore’s Law have encouraged consolidation among chipmakers and created more choke points in the immensely complex business of chip production. Even as microchips have become essential in so many products, their development and manufacturing have come to be dominated by a small number of producers with limited capacity—and appetite—for churning out the commodity chips that are a staple for today’s technologies. And because making chips requires hundreds of manufacturing steps and months of production time, the semiconductor industry cannot quickly pivot to satisfy the pandemic-fueled surge in demand. After decades of fretting about how we will carve out features as small as a few nanometers on silicon wafers, the spirit of Moore’s Law—the expectation that cheap, powerful chips will be readily available—is now being threatened by something far more mundane: inflexible supply chains.

now beyond the reach of many companies. In addition to Apple, only the largest tech companies that require the highest computing performance, such as Qualcomm, AMD, and Nvidia, are willing to pay hundreds of millions of dollars to design a chip for leading-edge nodes, says Sri Samavedam, senior vice president of CMOS technologies at Imec, an international research institute based in Leuven, Belgium. Many more companies are producing laptops, TVs, and cars that use chips made with older technologies, and a spike in demand for these is at the heart of the current chip shortage. Simply put, a majority of chip customers can’t afford—or don’t want to pay for—the latest chips; a typical car today uses dozens of microchips, while an electric vehicle uses many more. It quickly adds up. Instead, makers of things like cars have stuck with chips made using older technologies What’s more, many of today’s most popular electronics simply don’t require leading-edge chips. “It doesn’t make sense to put, for example, an A14 [iPhone and iPad] chip in every single computer that we have in the world,” says A lonely frontier Hassan Khan, a former doctoral researcher Twenty years ago, the world had 25 manuat Carnegie Mellon University who studfacturers making leading-edge chips. Today, ied the public policy implications of the ONLY THE LARGEST TECH only Taiwan Semiconductor Manufacturing end of Moore’s Law and currently works COMPANIES ARE at Apple. “You don’t need it in your smart Company (TSMC) in Taiwan, Intel in the WILLING United States, and Samsung in South Korea thermometer at home, and you don’t need TO PAY HUNDREDS OF MILLIONS OF DOLLARS have the facilities, or fabs, that produce 15 of them in your car, because it’s very TO DESIGN the most advanced chips. And Intel, long power hungry and it’s very expensive.” A CHIP FOR LEADING-EDGE a technology leader, is struggling to keep The problem is that even as more users NODES. up, having repeatedly missed deadlines for rely on older and cheaper chip technolproducing its latest generations. ogies, the giants of the semiconductor One reason for the consolidation is industry have focused on building new that building a facility to make the most leading-edge fabs. TSMC, Samsung, and advanced chips costs between $5 billion and $20 billion. These Intel have all recently announced billions of dollars in investfabs make chips with features as small as a few nanometers; in ments for the latest manufacturing facilities. Yes, they’re expenindustry jargon they’re called 5-nanometer and 7-nanometer sive, but that’s where the profits are—and for the last 50 years, nodes. Much of the cost of new fabs goes toward buying the it has been where the future is. TSMC, the world’s largest contract manufacturer for chips, latest equipment, such as a tool called an extreme ultraviolet lithography (EUV) machine that costs more than $100 million. earned almost 60% of its 2020 revenue from making leading-edge Made solely by ASML in the Netherlands, EUV machines are used chips with features 16 nanometers and smaller, including Apple’s to etch detailed circuit patterns with nanometer-size features. M1 chip made with the 5-nanometer manufacturing process. Making the problem worse is that “nobody is building semiChipmakers have been working on EUV technology for more than two decades. After billions of dollars of investment, EUV conductor manufacturing equipment to support older technolomachines were first used in commercial chip production in 2018. gies,” says Dale Ford, chief analyst at the Electronic Components “That tool is 20 years late, 10x over budget, because it’s amazing,” Industry Association, a trade association based in Alpharetta, says David Kanter, executive director of an open engineering Georgia. “And so we’re kind of stuck between a rock and a hard consortium focused on machine learning. “It’s almost magical spot here.” that it even works. It’s totally like science fiction.” Such gargantuan effort made it possible to create the bil- Low-end chips lions of tiny transistors in Apple’s M1 chip, which was made by All this matters to users of technology not only because of the TSMC; it’s among the first generation of leading-edge chips to supply disruption it’s causing today, but also because it threatrely fully on EUV. ens the development of many potential innovations. In addition Paying for the best chips makes sense for Apple because these to being harder to come by, cheaper commodity chips are also chips go into the latest MacBook and iPhone models, which sell becoming relatively more expensive, since each chip generaby the millions at luxury-brand prices. “The only company that is tion has required more costly equipment and facilities than the actually using EUV in high volume is Apple, and they sell $1,000 generations before. smartphones for which they have insane margin,” Kanter says. Some consumer products will simply demand more powerful Not only are the fabs for manufacturing such chips expen- chips. The buildout of faster 5G mobile networks and the rise sive, but the cost of designing the immensely complex circuits is of computing applications reliant on 5G speeds could compel

Read smarter, not harder A subscription to MIT Technology Review includes • Unlimited web access • Exclusive, subscriberonly stories • Digital version of each issue

• Subscriber-only app • The Algorithm newsletter • Access to 120+ years of publication archives

Subscribe today at technologyreview.com/subscribe

58

Change

investment in specialized chips designed for networking equip- 13%. Taiwan’s TSMC alone has nearly 55% of the foundry market ment that talks to dozens or hundreds of Internet-connected that handles consumer chip manufacturing orders. devices. Automotive features such as advanced driver-assistance Looming over everything is the US-China rivalry. China’s systems and in-vehicle “infotainment” systems may also benefit national champion firm SMIC has been building fabs that are from leading-edge chips, as evidenced by electric-vehicle maker still five or six years behind the cutting edge in chip technoloTesla’s reported partnerships with both TSMC and Samsung on gies. But it’s possible that Chinese foundries could help meet the chip development for future self-driving cars. global demand for chips built on older nodes in the coming years. But buying the latest leading-edge chips or investing in spe- “Given the state subsidies they receive, it’s possible Chinese cialized chip designs may not be practical for many companies foundries will be the lowest-cost manufacturers as they stand when developing products for an “intelligence everywhere” future. up fabs at the 22-nanometer and 14-nanometer nodes,” Khan Makers of consumer devices such as a Wi-Fi-enabled sous vide says. “Chinese fabs may not be competitive at the frontier, but machine are unlikely to spend the money to develop specialized they could supply a growing portion of demand.” chips on their own for the sake of adding even fancier features, The global semiconductor industry will need to almost double Kanter says. Instead, they will likely fall back on whatever chips overall capacity by 2030 to keep pace with demand, according made using older technologies can provide. to the Semiconductor Industry Association (SIA), a WashingtonAnd lower-cost items such as clothing, he says, have “razor- based industry group, which has advocated for strengthening thin margins” that leave little wiggle room for more expensive the global supply chain rather than attempting to build fully chips that would add a dollar—let alone $10 or $20—to each “self-sufficient” domestic manufacturing capability. item’s price tag. That means the climbing But in a nod to the importance of price of computing power may prevent advanced chips for national security and the development of clothing that could, critical infrastructure, the SIA suggests for example, detect and respond to voice that the US provide “market-driven incenTHE MAJORITY OF TODAY’S CHIP commands or changes in the weather. tives” for companies to build two or three CUSTOMERS The world can probably live without new leading-edge fabs domestically. That MAKE DO WITH THE fancier sous vide machines, but the lack could help ensure that the nation’s core CHEAPER COMMODITY CHIPS THAT of ever cheaper and more powerful chips telecommunications networks and data REPRESENT would come with a real cost: the end of an centers—along with the US military—have A TRADE-OFF BETWEEN COST AND era of inventions fueled by Moore’s Law a domestic supply of chips. PERFORMANCE. The White House photo ops with the and its decades-old promise that increaspresident called to mind the role that the ingly affordable computation power will be available for the next innovation. government has played since the dawn The majority of today’s chip customof the semiconductor industry that gave ers make do with the cheaper commodity chips that represent Silicon Valley its name. “Making that front and center is not somea trade-off between cost and performance. And it’s the supply thing that the president has talked about in that way since Ronald of such commodity chips that appears far from adequate as the Reagan,” says Margaret O’Mara, a historian at the University global demand for computing power grows. of Washington in Seattle. “Biden sitting there waving a wafer “It is still the case that semiconductor usage in vehicles is around—I don’t think I’ve seen that in a presidential hand ever.” going up, semiconductor usage in your toaster oven and for The US government became “the Valley’s first, and perhaps all kinds of things is going up,” says Willy Shih, a professor of its greatest, venture capitalist,” O’Mara wrote in her 2019 book management practice at Harvard Business School. “So then the The Code: Silicon Valley and the Remaking of America. Large question is, where is the shortage going to hit next?” government orders for chips to supply NASA’s Apollo program and the military’s Minuteman intercontinental ballistic missiles A global concern encouraged chipmakers to begin mass production and helped In early 2021, President Joe Biden signed an executive order lower the cost of the first silicon chips from $1,000 each in 1960 mandating supply chain reviews for chips and threw his support to just $25 by 1965. The price drop made computing power affordable to many behind a bipartisan push in Congress to approve at least $50 billion for semiconductor manufacturing and research. Biden also beyond just deep-pocketed government agencies. It kick-started held two White House summits with leaders from the semicon- the golden age of Moore’s Law, in which customers reaped the ductor and auto industries, including an April 12 meeting during benefits of cheaper chips that also delivered better performance which he prominently displayed a silicon wafer. every few years. And you might not know its promise was in peril The actions won’t solve the imbalance between chip demand if all you had to go on was Apple’s latest ad. While I was interviewing O’Mara for this story, a delivery and supply anytime soon. But at the very least, experts say, today’s crisis represents an opportunity for the US government to try person showed up at her door as if on cue. to finally fix the supply chain and reverse the overall slowdown “Speaking of chips, I’m just getting handed my brand-new in semiconductor innovation—and perhaps shore up the US’s computer,” she said with a laugh. “Yes, I’ve got my new MacBook capacity to make the badly needed chips. with my M1 chip.” An estimated 75% of all chip manufacturing capacity was based Jeremy Hsu is a technology and science journalist based in East Asia as of 2019, with the US share sitting at approximately in New York City.

Don’t believe everything you hear. The award-winning podcast In Machines We Trust thoughtfully examines the far-reaching impact of artificial intelligence on our daily lives. Download it wherever you listen.

OPPOSITE: GETTY IMAGES (BRANCH,COUPLES); PEABODY ESSEX MUSEUM VIA WIKIMEDIA; PUBLIC DOMAIN VIA WIKIMEDIA (ENGRAVING, TICKET, STAMP); ALAN RICHARDS PHOTOGRAPHER; FROM THE SHELBY WHITE AND LEON LEVY ARCHIVES CENTER, INSTITUTE FOR ADVANCED STUDY, PRINCETON, NJ, USA (BLACKBOARD);

61

One day in late March, People’s Daily, the Chinese Communist Party’s official newspaper, shared a pair of photos on Chinese social media. The first, in black and white, was of the signing of the Boxer Protocol, a 1901 treaty between the Qing empire, which ruled China at the time, and 11 foreign nations. Troops from eight of these countries, including the US, had occupied Beijing following sieges on their embassies by a peasant militia known as the Boxers. Among a litany of concessions, the Qing government agreed to pay the eight occupying powers an indemnity of 450 million taels of silver (about $10 billion in today’s dollars), almost twice its annual revenue. The Boxer Protocol is etched into the Chinese consciousness as a searing reminder of the country at its weakest. The second image, in vivid color, was from the previous day, at an acrimonious summit held in Alaska between top Chinese and American officials. It was the first high-level meeting between the two governments during the Biden administration. The officials criticized one another’s governments for human rights abuses and belligerence on the international stage. At the end of the opening session, Yang Jiechi, director of foreign affairs for the Chinese Communist Party, scolded his American counterparts: “Haven’t we, the Chinese people, suffered from foreign bullies long enough? Haven’t we been penned in by foreign nations and stopped from progress long enough?” The People’s Daily post quoted Yang as saying further: “You, the United States, are not qualified to claim that you are speaking to China from a position of strength.” This struck a nerve; the post has been liked almost 2 million times, and Yang’s quote has found its way to T-shirts, stickers, and cell-phone covers sold in China. To many in the country, the harsh words carry the sweet taste of revenge. China is finally strong enough to stand up to the most powerful nation on earth and demand to be treated as its equal. From the last Chinese empire to the current People’s Republic, generations of politicians and intellectuals have sought ways to build a strong China. Some imported tools and ideas from the West. Others left China for a better education, but the homeland still beckoned. They pondered the relationships between East and West, tradition and modernity, national allegiance and cosmopolitan ideals. Their accomplishments and regrets have shaped the path of China’s development and mapped the contours of Chinese identity. I’m a product of their complex legacy. I grew up in Hefei, a medium-sized city in central-eastern China. The Hefei of my childhood was a humble place, known for ancient battlegrounds, sesame snacks, and a few good universities. I spent the first 19 years of my life

THOSE WHO FALL BEHIND GET BEATEN UP Can science build a strong China?

By Yangyang Cheng

62

Change

there and left in 2009 to pursue my PhD in physics in the US, where I now live and work. Watching my birth country’s ascent conjures up mixed feelings. I’m glad that the majority of Chinese people enjoy a higher standard of living. I’m also alarmed by the hardened edge to China’s new superpower status. Economic growth and technological advancements have not ushered in more political freedoms or a more tolerant society. The Chinese government has become more authoritarian and its people more nationalistic. The world feels more fractured today. Hefei is now a budding metropolis with new research centers, manufacturing plants, and technology startups. For two of the city’s proudest sons, born a century apart, a strong homeland armed with science and technology was the aspiration of a lifetime. One of these men was the late Qing’s most revered statesman. The other is one of the first two Nobel laureates from China. The Boxer Protocol marked the end of one career and laid the foundation for the other. I grew up with their names and have been returning to their stories. They teach me about the forces that propelled China’s rise, the way lives can be squeezed by the pressures of geopolitics, and the risks of using science for state power.

IN 1823,

Li Hongzhang was born to a wealthy household in Hefei, then a small provincial capital surrounded by farmland. Like his father and brother before him, Li excelled in the imperial exams, China’s centuries-old system for selecting officials. Over six feet tall and with a piercing gaze, he commanded space and attention. He distinguished himself in suppressing peasant rebellions and rose quickly in the imperial court to become the Qing empire’s highest-ranking governor, its commerce minister, and its de facto foreign minister. After China lost to British and French forces in the Opium Wars, Li and his allies launched a wide range of reforms. They called it the Movement for Western Affairs, also known as “self-strengthening.” The strategy was best summed by the scholar Wei Yuan in an 1844 book, Illustrated Treatise on the Maritime Kingdoms: “Learn advanced technologies from the barbarians to keep barbarian invaders at bay.” To the Chinese literati, the world was divided between hua, the homeland of civilized glory, and yi, the places where barbarians dwelled. British gunboats on the southern shore had shaken but not shattered this centuries-old belief. Proponents of self-strengthening claimed that Chinese tradition was the base onto which Western technology could be grafted for practical use. As the historian Philip Kuhn has argued, such logic also implied that technology was culturally neutral and could be detached from political systems. A classically trained scholar and battle-tested general, Li championed both civilian and military enterprises. He

petitioned the emperor to construct the first Chinese railroad and founded the country’s first privately owned steamship company. He allocated generous government funding for the Beiyang Fleet, China’s first modern navy. In 1865, Li oversaw the establishment of the Jiangnan Arsenal, the largest weapons factory in East Asia at the time. In addition to producing advanced machinery for war, the arsenal also included a school and a translation bureau, which translated scores of Western textbooks on science, engineering, and mathematics, establishing the vocabulary in which these subjects would be discussed in China. Li also supervised China’s first overseas education program, which sent a cohort of Chinese boys aged 10 to 16 to San Francisco in the summer of 1872. After a promising start, the mission was derailed by anti-Chinese racism in the US and conservative obstruction at home. Some students, upon returning to China, were held and questioned by the authorities about their loyalty. After nine bumpy years, the program was shut down in 1881 on the eve of the Chinese Exclusion Act. Meanwhile, neighboring Japan had adopted not only the West’s technology but also its governing methods, transforming a feudal society into a modern industrial state with a formidable military. For centuries, the Chinese elite had looked down upon Japan, dismissing it as small and inferior. When the two countries went to war in 1894, ostensibly over the status of Korea, the real prize was status as the preeminent Asian power. Japan won decisively. It was six years after this devastating loss that Li signed the Boxer Protocol on behalf of the Qing government. He died two months later. By the start of the 20th century, the last Chinese empire had lost its legitimacy. Armed rebellions were erupting across the country. The Qing regime was overthrown in 1911, and the Republic of China was born. Progressive intellectuals saw Chinese tradition as “rotten and decayed,” a cultural albatross holding their country back. They believed that national salvation demanded embracing Western ideas. The few dissenting voices were sidelined. China’s path to westernization received some early assistance from the US. Hoping to improve relations between the two countries, the US government decided to return almost half the American portion of the indemnity China had agreed to pay in the Boxer Protocol. With the US side dictating the terms, part of the remittance went toward a program known as the Boxer Indemnity Scholarships, which provided one of the few pathways for Chinese students to study in the US. The bulk of the returned payment was used to establish a Westernstyle preparatory school, which became Tsinghua University, China’s premier technological institution. Li Hongzhang could not have imagined that after his death, the most shameful chapter of his career would, at the whimsical hand of geopolitics, contribute to his lifelong dream of bringing Western science and education to China. Tsinghua took its motto from the ancient

PEOPLE’S DAILY VIA WEIBO

63 after the Communist victory, and pioneered the field of semiconductor physics in the country. Deng Jiaxian, Yang’s best friend since adolescence, boarded a ship back nine days after receiving his PhD from Purdue. He became a leader in China’s fledgling nuclear weapons program. Some overseas Chinese scientists, dreading Communist rule, followed the Nationalist government to Taiwan, Yang’s former mentor Wu Ta-You among them. But Yang opted to stay in the US after getting his doctorate, moving in 1949 to the Institute for Advanced Study in Princeton, New Jersey. There he would spend the better part of the next two decades. He would not see any of his old friends for many years. In 1957, Yang and Tsung-Dao Lee, a fellow Chinese graduate of the University of Chicago, won the Nobel Prize for proposing that when some elementary particles decay, they do so in a way that distinguishes left from right. They were the first Chinese laureates. Speaking at the Nobel banquet, Yang noted that the prize had first been awarded in 1901, the same year as the Boxer Protocol. “As I stand here today and tell you about these, I am heavy with an awareness of the fact that I am in more than one sense a product of both the Chinese and Western cultures, in harmony and in conflict,” he said. Yang became a US citizen in 1964 and moved to Stony Brook University on Long Island in 1966 as the founding director of its Institute for Theoretical Physics, which was later named after him. As People’s Daily text of I Ching, the Book of Changes: “The the relationship between the US and China juxtaposed a began to thaw, Yang visited his homeland in work of self-strengthening is ceaseless. The photograph from the virtuous carry the world with generosity.” 1971—his first trip in a quarter of a century. A signing of the Boxer Protocol with one from lot had changed. His father’s health was failthe Alaska summit. ing. The Cultural Revolution was raging, and The post has been IN 1945, liked almost 2 million both Western science and Chinese tradition times. had been deemed heresy. Many of Yang’s fora young man named Chen Ning Yang gradmer colleagues, including Huang and Deng, uated from Tsinghua and arrived at the University of were persecuted and forced to perform hard labor. The Chicago for his PhD on a Boxer Indemnity Scholarship. Nobel laureate, on the other hand, was received like a foreign dignitary. He met with officials at the highest Inspired by the autobiography of Benjamin Franklin, which he had read as a child, the aspiring physicist from levels of the Chinese government and advocated for Hefei gave himself the English name Frank. the importance of basic research. In the years that followed, Yang visited China reguAfter World War II ended, Nationalists and Communists continued to battle in China. Yang and larly. At first, his trips drew attention from the FBI, which his small cohort of overseas Chinese students faced saw exchanges with Chinese scientists as suspect. But by the late 1970s, hostilities had waned. Mao Zedong a pressing dilemma: Should they stay in the West— despite its racism and anticommunist paranoia—and was dead. The Cultural Revolution was over. Beijing enjoy social stability, material comfort, and career adopted reforms and opening-up policies. Chinese opportunities? Or should they return to their impovstudents could go abroad for study. Yang helped raise erished homeland after graduation and help it rebuild? funding for Chinese scholars to come to the US and for In a long letter to Yang in 1947, his college classmate international experts to travel to conferences in China, Huang Kun wrote, “It’s difficult to imagine how intellecwhere he also helped establish new research centers. When Deng Jiaxian died in 1986, Yang wrote an emotuals like us can affect the fate of a nation. Independent tional eulogy for his friend, who had devoted his life to minds like us, once we go back, will certainly get crushed like grains in a mill … but I still sincerely believe that China’s nuclear defense. It concluded with a song from whether China has us makes a difference.” 1906, one of his father’s favorites: “[T]he sons of China, Huang was studying in England at the University they hold the sky aloft with a single hand … The crimson never fades from their blood spilled in the sand.” of Bristol. He returned to China in 1951, two years

Change

Yang retired from Stony Brook in 1999 and moved back to China a few years later to teach freshman physics at Tsinghua. In 2015, he renounced his US citizenship and became a citizen of the People’s Republic of China. In an essay remembering his father, Yang recounted his earlier decision to emigrate. He wrote, “I know that until his final days, in a corner of his heart, my father never forgave me for abandoning my homeland.”

IN 2007,

when he was 85 years old, Yang stopped by our hometown on an autumn day and gave a talk at my university. My roommates and I waited outside the venue hours in advance, earning precious seats in the packed auditorium. He took the stage to thunderous applause and delivered a presentation in English about his Nobelwinning work. I was a little perplexed by his choice of language. One of my roommates muttered, wondering whether Yang was too good to speak in his mother tongue. We listened attentively nevertheless, grateful to be in the same room as the great scientist. A college junior and physics major, I was preparing to apply to graduate school in the US. I’d been raised with the notion that the best of China would leave China. Two years after hearing Yang in person, I too enrolled at the University of Chicago. I received my PhD in 2015 and stayed in the US for postdoctoral research. Months before I bid farewell to my homeland, the central government launched its flagship overseas recruitment program, the Thousand Talents Plan, encouraging scientists and tech entrepreneurs to move to China with the promise of generous personal compensation and robust research funding. In the decade since, scores of similar programs have sprung up. Some, like Thousand Talents, are supported by the central government. Others are financed by local municipalities. Beijing’s aggressive pursuit of foreign-trained talent is an indicator of the country’s new wealth and technological ambition. Though most of these programs are not exclusive to people of Chinese origin, the promotional materials routinely appeal to sentiments of national belonging, calling on the Chinese diaspora to come home. Bold red Chinese characters headlined the web page for the Thousand Talents Plan: “The motherland needs you. The motherland welcomes you. The motherland places her hope in you.” These days, though, the website isn’t accessible. Since 2020, mentions of the Thousand Talents Plan have largely disappeared from the Chinese internet. Though the program continues, its name is censored on search engines and forbidden in official documents in China. Since the final years of the Obama administration, the Chinese government’s overseas recruitment has come under intensifying scrutiny from US law enforcement. In 2018, the Justice Department started a China Initiative intended to combat economic espionage, with a focus

on academic exchange between the two countries. The US government has also placed various restrictions on Chinese students, shortening their visas and denying access to facilities in disciplines deemed “sensitive.” There are real problems of illicit behavior in Chinese talent programs. Earlier this year, a chemist associated with Thousand Talents was convicted in Tennessee of stealing trade secrets for BPA-free beverage can liners. A hospital researcher in Ohio pled guilty to stealing designs for exosome isolation used in medical diagnosis. Some US-based scientists failed to disclose additional income from China in federal grant proposals or on tax returns. All these are cases of individual greed or negligence. Yet the FBI considers them part of a “China threat” that demands a “whole-of-society” response. The Biden administration is reportedly considering changes to the China Initiative, which many science associations and civil rights groups have criticized as “racial profiling.” But no official announcements have been made. New cases have opened under Biden; restrictions on Chinese students remain in effect. Seen from China, the sanctions, prosecutions, and export controls imposed by the US look like continuations of foreign “bullying.” What has changed in the past 120 years is China’s status. It is now not a crumbling empire but a rising superpower. Policymakers in

W. & D. DOWNEY, PUBLIC DOMAIN, VIA WIKIMEDIA

64

ENERGY.GOV, PUBLIC DOMAIN, VIA WIKIMEDIA

65 It was a very difficult decision, and I’m still reckoning with the sense of loss associated with the change. But with every passing day, news from my birth country and my adopted home reminds me of why I made the choice. Advancements in science and technology have created unprecedented wealth—as well as inequality and capacity to cause harm. In the fevered race for power and supremacy, concerns about ethics and sustainability are drowned out by jingoistic cheers. My mother has been trying to persuade me to move back to China. She tells me how Hefei has shed its rusty, blue-collar image to become a modern city. “It has a new subway system! Do you know how fast it is?” she says over the phone. The sincerity in her voice breaks my heart. I want to say that I do not care for fast trains or new buildings—I really don’t—but I also know that my mother does not care for these things either. Her pride in her country’s development is genuine. If there’s anything she loves more than her homeland, though, it’s her child. My mother wants me to come back not because of some lofty ideals of patriotism, though she believes in them; nor for my career advancement, though the Chinese government has been investing heavily in the fundamental sciences. My mother wants me to come back because she is afraid. My mother is afraid that the borders between the US and China will be closed again as they were during the pandemic, shut down by forces just as Above: countries use similar techno-nationalistic laninvisible as a virus and even more deadly. She Yang (seated, left) guage to describe science as a tool of national fears for my safety in a foreign land that is in with fellow Nobel greatness and scientists as strategic assets in many ways increasingly hostile to my race and Prize winners (clockwise from left) Val geopolitics. Both governments are pursuing nationality. What my mother does not know, Fitch, James Cronin, military use of technologies like quantum or refuses to accept, is that the homeland is Samuel C.C. Ting, and Isidor Isaac Rabi not safe for me either. A state can command computing and artificial intelligence. “We do not seek conflict, but we welcome the world’s second-largest economy and a Opposite: stiff competition,” National Security Advisor strong military, and still be too fragile to allow Li (center) pictured with Lord Salisbury Jake Sullivan said at the Alaska summit. Yang dissent. Sometimes, life as a Chinese person (left) and Lord Curzon Jiechi responded by arguing that past confronmeans following one’s conscience with no (right) during a trip to England in 1896 tations between the two countries had only refuge in sight. At Li Hongzhang’s family temple on the damaged the US, while China pulled through. outskirts of Hefei, there is an old yulan tree. Tall and Much of the Chinese public relishes the prospect of competing against the US. Take a popular saying fragrant, yulan was a favorite of royalty. Legend has of Mao’s: “Those who fall behind will get beaten up!” it that this tree was a gift from the Japanese prime The expression originated from a speech by Joseph minister on Li’s 70th birthday. Li planted it himself. Stalin, who stressed the importance of industrialization In less than a year, the two countries would be at war. The tree has outlived both men and the empires they for the Soviet Union. For the Chinese public, largely served. It blossoms every year and occasionally bears unaware of its origins, it evokes the recent past, when fruit. It is a witness, and also a teacher. One day, when a weak China was plundered by foreigners. When I was little, my mother often repeated the expression at I’m able to go back to China and to Hefei, I hope to home, distilling a century of national humiliation into visit Li’s old residence. a personal motivation for excellence. It was only later, I hope to be there in the spring, when the yulan in adulthood, that I began to question the underlying blooms. Its flowers will be the purest white. Its petals logic: Is a competition between nations meaningful? will be thick and smooth. Its branches will lift into the By what metric, and to what end? sky. When the sun hits at just the right spot, its shadow After 11 years of designing particle detectors and will carry the shape of home. searching for dark matter, I left physics at the end of Yangyang Cheng is a particle physicist and a 2020 for a position working on science policy and ethics. postdoctoral fellow at Yale Law School.

FAST TIME Climate change is warping geological time, compressing the time scales of natural processes. In photographs taken around the world, Ian van Coller has documented these shifts, reflected in rocks, sediment, and the shrinking of glaciers. Van Coller collaborates with scientists who annotate his images, pointing out key geological features. He also uses historical photos to show changes, juxtaposing the black-and-white images taken by earlier expeditions with today’s landscapes; peaks once covered in snow are now bare rock.

Opposite: A mud core from Fairy Lake in Montana, superimposed against the surrounding mountains, reveals thousands of years of vegetative history. Geographer James Benes annotated the photo.

67

Quelccaya Glacier in Peru, seen here in 2017, is receding. The foreground rocks show signs of glacial erosion and were likely still covered 10 years ago. Each layer in the ice represents a year’s worth of snow. Annotated by geographer Carsten Braun.

68

70

A photograph taken in 2020 shows just how little is left of the glacier at Mount Stanley in Uganda. The photo from a 1906 expedition shows the glacier below Elena Peak; today what’s left is dirty ice, a sign the glacier will soon be gone. (Carsten Braun)

71

72

At nearby Mount Baker, also in Uganda, the story is similar. Dotted lines are an attempt to estimate the ice seen in earlier expeditions on Semper Peak, which is now bare rock. There is no sign of what the 1906 photo labels Moore Glacier. (Carsten Braun)

73

RE

“I have a stalker—an online harasser who has moved into real-world stalking—which made me more fearful. ” —p. 87

VIEW

Books, policy, and culture in perspective

JAFFE

Power play Workers across the technology industry are organizing in new and old ways to demand more say in how some of the world’s wealthiest companies operate.

he workers at the Amazon fulfillment center in Bessemer, Alabama, wanted a union. The center opened in March of last year, just as stay-at-home orders for covid-19 went into effect. While much of the world economy tanked, some sectors thrived, including tech—Amazon founder Jeff Bezos would add some $75 billion to his own net worth in 2020. Back at Bessemer, though, workers were being pressed to work harder and longer, and they felt dehumanized. They wanted dignity, not just higher wages.

T

ANDREA DAQUINO

SARAH

Review

The workers’ push to join the Retail, Wholesale, and Department Store Union (RWDSU) was always going to be an uphill battle. Amazon used its ever-growing wealth to fight the union’s campaign. Bosses used social distancing protocols meant to stop covid-19 transmission as a pretext to stymie communication between workers. Employerfavoring federal and state labor laws allowed management to run a scorched-earth anti-union effort: Amazon hired anti-union consultants, flooded employees with text messages and signs urging them to “vote no,” and held “captive audience” meetings where workers were required to sit through anti-union lectures. The results were painful for supporters: 738 votes for the union, 1,798 against. But even as the votes were being counted, workers around the country were agitating. On April 7, self-organized workers under the mantle of Amazonians United Chicagoland struck against the company’s “megacycle,” a grueling 10-hour overnight warehouse shift. Workers organized in California’s Inland Empire. Nationwide, hundreds of employees from at least 50 Amazon facilities refused to work during the pandemic. A group in Staten Island was moving to start its own grassroots union. In the Twin Cities, the Awood Center, a workers’ advocacy group for East African immigrants, convinced the company to sit down with workers and come to an agreement about accommodations for religious observances. The Bessemer fight, and Amazon organizing as a whole, reflect a new groundswell of interest in organizing among tech workers. But that current has also raised a question: What is a “tech” worker, anyway? The term could reasonably be applied to anyone from programmers to data

center staff to warehouse pickers to assembly-line autoworkers in a Tesla factory. The reality is that organizing in “tech” is sort of like organizing in “industry” in the 1930s. Back then, the Congress of Industrial Organizations shifted the focus of the labor movement from organizing the skilled trades to bringing together the “unskilled” workers in massive new factories. The new labor movement was epitomized by Detroit’s auto plants but emerged across a wide variety of industries characterized by new technology and scientific management tactics. This era of industry required a new kind of union, and labor struggled for decades before hitting upon methods that worked—and, importantly, getting the backing of the Depression-era federal government. Organizing tech workers will require a similar effort, a similar reorganization of labor tactics, and, quite possibly, a similarly supportive federal government. As the result in Bessemer shows, today’s workers are up against the world’s richest companies—companies with the world’s most sophisticated surveillance and information systems, not to mention millions to spend on anti-union consultants. For both sides in this struggle, though, the bottom line is not money but power. SHARED POWER

To understand tech’s new labor movement, says Emma Kinema, an organizer with the Communications Workers of America’s tech-sector organizing project CODE-CWA, one must understand that tech is everywhere, that most workers are in some sense working with technology, and that nevertheless there is a consistency to the thing we call the “tech industry,” even if it is massive and diverse.

75

It’s also important to remember that the culture of Silicon Valley was anti-union from the start; one reason California became the tech hub of choice was that the Boston area, where many early leaders in the industry got their start, had a long-established union presence. Logic Magazine’s Ben Tarnoff notes that the common perks and amenities of the tech workplace—free food, toys and games, and casual dress—began as explicitly antiunion measures. This culture, which sociologist Andrew Ross has called “no-collar,” was designed to engender not just loyalty but a love for and identification with the company. Labor historian Nelson Lichtenstein says Big Tech has a tendency to lean on its transformational image to paper over any labor complaints and minimize them as quibbles that are impeding the evolution of the world. That’s nothing new, he argues—Henry Ford responded in much the same way when workers in his factories spoke up, and Ford learned the tactic from the railroad magnates who preceded him. Despite the bosses’ pushback, workers in these trailblazing sectors did eventually unionize—though it took years and quite a few failed efforts. Steelworkers held massive strikes in 1919 but failed. “In the ’20s and ’30s everyone thought steel was impossible to organize,” says Kinema. “Those workers were too well paid; their industry was too new; they had these new, modern ways of management.” But by 1937, US Steel recognized the union. “Can you imagine if Google workers went on strike?” she asks. The site reliability engineers alone, who maintain the technical infrastructure of Google, “could shut down half the internet.” Chewy Shaw is one of those engineers, as well as the executive

76

Change

P O W E R P L AY

vice chair of the Alphabet Workers Union, which is part of CWA. The union went public in January of 2021 with just over 200 members and now has more than 800, including Google programming staff like Shaw and executive chair Parul Koul, as well as researchers, data-center workers, temps, and vendors. The union is not big enough to level the internet anytime soon, nor is a strike imminent (Alphabet, Google’s parent company, employs 135,000 people). But it has challenged the company to do better by its workers, leveraging its complement of “no-collar” staff like Shaw and Koul to win real changes for employees with less power and security. As one example, the union supported Shannon Wait, a data-center technician employed through a subcontractor in South Carolina, through a wrongful suspension this March for speaking to coworkers about her working conditions. The National Labor Relations Board reversed the suspension and ordered the company to post notices informing workers that they have the right to organize. Much of this work builds on previous tech-worker activism that gained steam during Donald Trump’s presidency, when liberal-leaning no-collar employees learned that the bosses they had thought shared their values were in fact happy to work with the administration. When Google workers realized they were building infrastructure for Project Maven, an artificial-intelligence project for the US military, they noted that programmers working on the software might not even know that their code could be used for drone attacks. Thousands of Google employees signed a letter protesting the company’s involvement in Project Maven in early 2018, and Google let its Maven contract expire the

THE GOOGLE WALKOUT UNDERSCORED THE FACT THAT MANY WORKERS, EVEN THOSE WITH THE HIGHEST SALARIES OR THE MOST JOB SECURITY, DIDN’T FEEL VALUED BY THE COMPANY.

next year. That wasn’t workers’ only concern, though—a massive international work stoppage followed in November 2018, centering on sexual harassment and discrimination at the company. The Google walkout underscored the fact that many workers, even those with the highest salaries or the most job security, didn’t feel valued by the company. And while many observers scoffed—venture capitalist Mike Solana wrote on Twitter that workers like Shaw and Koul “aren’t oppressed coal miners”—this feeling led to the formation of the Alphabet Workers Union. “Asking for respect on the job is not specific to coal miners, and that’s really why we all do this,” Koul says. These workers want to leverage the power they have within the company as part of a broader workingclass movement. That meant refusing to work on Maven; and now it means demanding, in solidarity with the Black Lives Matter movement, that the company not sell tech to police. This organizing takes cues from unions like the Chicago Teachers Union, which made racial justice and economic inequality across the city central to its demands—and won those fights through two widely publicized strikes in 2012 and 2019. Longtime union strategist Stephen Lerner says that through such “bargaining for the common good” at Google, the workers challenge the company’s impact on society, not just their own treatment. “I don’t think the tech organizing would have any of the kind of resonance it has now if people opened up with ‘Well, we need a better 401(k) plan,’” Lerner says. THE RISE OF THE TECH WORKER

Even in the early 1990s, when Lerner went to war with Apple as an organizer of the Justice for Janitors

campaign and won union rights for subcontracted cleaning workers across the tech sector, the question of “Who is a tech worker?” loomed large. Through those successful campaigns, Lerner helped extend the definition of a tech worker to virtually anyone who makes a tech company run. Cori Crider, an attorney with Foxglove, a firm that aims to challenge the power of Big Tech, has been working with subcontracted content moderators—real humans who sift through posts with violence and racism and graphic sex every day, trying to determine what violates a constantly shifting set of rules. Those workers are often bound by nondisclosure agreements that keep them from speaking publicly about their working conditions. That allows companies like Facebook to deny they exist—an assertion the company stuck with last year even after reports emerged that moderators working for the outsourcing firm Accenture were being pushed back into the office during the pandemic. Tech workers outside the normal definition of “employees” are still finding ways to organize and protect themselves. Coworker.org, a campaign platform for labor organizing, is using donations from well-off tech workers to build a “solidarity fund” distributed to workers on the other side of the tech supply chain. Gig workers on Amazon’s Mechanical Turk platform are using the site Turkopticon to come together and fight for better terms. At the other end of the techworker spectrum are those building electric cars at Tesla’s plant in Fremont, California. Before Elon Musk’s company bought the Fremont facility, it was known as New United Motors Manufacturing, Inc., or NUMMI, a collaboration between General Motors and Toyota where Japanese “lean production”

Review

was brought to America. NUMMI didn’t survive GM’s bankruptcy in 2008, and Tesla snatched it up. Cooperating with the United Auto Workers was one of NUMMI’s big innovations, but Tesla’s gone another way. Recently, an administrative judge at the NLRB ruled that several of the company’s actions in response to worker organizing were illegal—including a couple of Musk’s tweets as well as harassment of workers passing out union pamphlets, banning of pro-union T-shirts and buttons, and the interrogation of organizers and firing of one. The NLRB’s penalties amount to little more than a finger-wag—Musk must read a statement telling workers that they have the right to unionize, and rehire the fired worker. He’s appealed the decision anyway. The workers at the plant, even the union supporters, are enthusiastic about producing electric vehicles, but they note that the technical sophistication of the plant does not prevent a lot of backbreaking manual labor—or injuries. Jose Moran, one of the leaders of the union drive and a former NUMMI worker, wrote a blog post about the things he wanted to improve, including the grueling pace of the work and some badly designed machinery. Autoworkers have struggled with machinery since the days of Henry Ford. But Tesla workers’ stories echo the complaints of autoworkers in the 1960s who battled “speed-up”—the way management would use new technology to ratchet up the pace of work—in places like Lordstown, Ohio, and Detroit. A wave of rebellions within the unions and wildcat strikes challenged the idea that automation was making their jobs easier. As machines sped up the manufacturing process, workers had to hustle faster to keep up. The autoworkers at Tesla, far from representing a labor

aristocracy among autoworkers, say they make less than unionized workers at GM and Ford. As Moran wrote, “I often feel like I am working for a company of the future under working conditions of the past.” THE LONG GAME

In the Amazon warehouses, too, everything old is new again. “The auto industry tried to do lots of automation back in the ’80s, ’70s, whatever, and they basically plateaued out where they couldn’t do it anymore. And Tesla basically tried to do the same thing,” says Tyler Hamilton, an Amazon warehouse worker from Minneapolis. “It’s the same thing with Amazon. There’s only so much you can do with automation.” Mohamed Mire, a coworker of Hamilton’s, explains that most of Amazon’s vaunted technology goes to tracking the workers rather than making the work efficient. Scanners that the workers use to scan packages also keep track of their so-called “time off task,” and they get written up if their productivity rate falls. Robots that Hamilton likens to “giant Roombas” carry merchandise around the warehouse but malfunction often—lately his job has included setting the robots right when they stop working. Data from Amazon shows that injury rates are higher at facilities with robots than without them. Hamilton and Mire work with the Awood Center, which—since it is a worker center rather than a union— doesn’t go through NLRB elections but instead organizes through direct action. Awood members have won some concessions from Amazon, particularly around prayer time (many are practicing Muslims) and accommodations for fasting during Ramadan. They’ve also gotten people hired back who’d been fired. Despite the results in Alabama, workers like Hamilton and Mire have

77

no intention of slowing their organizing. But Amazon’s heavy-handed tactics—including the hiring of actual Pinkertons, security agents from a company that has been helping employers break unions since the 19th century—are also unlikely to stop. The NLRB is deciding whether to consolidate complaints against the company across its various regions— there have been at least 37 in 20 US cities since the beginning of the pandemic. RWDSU filed 23 complaints of unfair labor practices just in Bessemer, including the charge that Amazon illegally threatened workers with layoffs or the facility’s closure. There is clearly still a long way to go before tech workers win at the bargaining table, but history offers them plenty of models to look to. Lichtenstein, the labor historian, points to the International Longshore and Warehouse Union, a powerful West Coast waterfront union that inked an agreement with shippers in 1958 to get dockworkers a slice of the gains from automation. When selective use of automation led to more injuries, the union actually pushed for more tech to improve safety. They drove wages for what had been insecure, contingent work to over $150,000 a year. In the current struggle, the Biden administration has signaled support both for sweeping labor law reform—which would make many of Amazon’s tactics at Bessemer illegal—and potentially for regulating Big Tech. And Hamilton notes, “It took something like 50 years to unionize US Steel. Amazon’s warehouses were just built a handful of years ago. If it’s not this year or next year, it’ll be five years from now.” Q Sarah Jaffe is a fellow at Type Media Center and author of Work Won’t Love You Back.

78

Change

THE

SUROWIECKI

The big breakup It’s become trendy to talk of forcing the biggest US tech companies to loosen their monopolistic hold on the digital economy. But it won’t be easy.

or Apple, Amazon, Facebook, and Alphabet, covid-19 was an economic blessing. Even as the F pandemic sent the global economy into a deep recession and cratered most companies’ profits, these companies—often referred to as the “Big Four” of technology—not only survived but thrived. Collectively, they now have annual revenue of well over a trillion dollars, and the value of their stocks has soared: together they’re worth $2.5 trillion more than they were 15 months ago. Yet at the same time, they have come under unprecedented attack from politicians and government

ANDREA DAQUINO

JAMES

Review

regulators in the US and Europe. While congressional hearings on charges that Facebook has been censoring conservatives or not doing enough to restrain disinformation and hate speech may have gotten most of the headlines and public attention, the companies are facing far more substantive threats, in the form of new lawsuits, proposed bills, and regulations. This past fall, the Federal Trade Commission and 48 state attorneys general filed suit against Facebook, charging it with illegally maintaining a monopoly over the social-networking space “through a years-long course of anticompetitive conduct.” Soon after, the US Department of Justice and 11 state attorneys general filed suit against Google, charging it with illegally maintaining a monopoly over the search and search advertising markets. Apple is currently locked in a civil trial with game developer Epic Games, which is challenging Apple’s control of its App Store on antitrust grounds. Last summer, the US House Judiciary Committee concluded a 19-month investigation into alleged anticompetitive activity by the tech titans. The resulting 450-page report described the companies as “the kinds of monopolies we last saw in the era of oil barons and railroad tycoons” and recommended that the government take action against them. It’s easy, of course, to dismiss anything that comes out of Washington or Brussels as political posturing, but in this case that would be a mistake. President Joe Biden has named some of Big Tech’s sharpest and most vocal critics— including Columbia University professor Tim Wu, author of the book The Curse of Bigness, and Lina Khan, who served as special

counsel to the Judiciary Committee during its investigation—to important roles in his administration. Europe is putting in place tougher regulations to try to limit Big Tech’s power. And antitrust action, at least with regard to the tech industry, has become that rarest of things: a bipartisan issue in Congress. What’s arguably more important is that we’re in the middle of a radical shift in the intellectual discussion—one that has made it much easier to go after Big Tech. In many ways, we seem to be going back to the antitrust vision that determined US policy toward big companies for much of the 20th century, a vision that’s much more skeptical of the virtues of size and much more willing to be aggressive in keeping companies from exercising monopoly power. America’s key antitrust laws were written around the turn of the 20th century. The Sherman Antitrust Act of 1890 and the Clayton Act of 1914 remain on the books today. They were written in broad, far-reaching (and ill-defined) language, targeting monopolists who engaged in what they called “restraint of trade.” And they were driven in large part by the desire to curb the giant trusts that had, via a series of mergers and acquisitions, come to dominate America’s industrial economy. The quintessential example was Standard Oil, which had built an empire that gave it essentially complete control over the oil business in the US. But antitrust law wasn’t just used to block mergers. It was also used to stop a host of practices that were deemed anticompetitive, including some that nowadays seem routine, like aggressive discounting or tying the purchase of one good to the purchase of another. This all changed with the Reagan administration in the 1980s. Instead

79

of worrying about big companies’ impact on competitors or suppliers, regulators and courts started to focus almost entirely on what was called “consumer welfare.” If a merger, or a company’s practices, could be shown to lead to higher prices, then it made sense to step in. If it didn’t, antitrust regulators generally took a hands-off approach. That’s why Facebook’s acquisitions of Instagram and WhatsApp, Amazon’s acquisition of Zappos, and Google’s acquisitions of DoubleClick, YouTube, Waze, and ITA all sailed through the regulatory approval process without a hitch. No longer, though. Over the past four or five years, scholars, politicians, and public advocates have begun to push a new idea of what antitrust policy should be, arguing that we need to move away from that narrow focus on consumer welfare—which in practice has usually meant a focus on prices—toward consideration of a much wider range of possible harms from companies’ exercise of market power: damage to suppliers, workers, competitors, customer choice, and even the political system as a whole. They’ve done so, not surprisingly, with the Big Four squarely in mind. But what exactly would reining in Big Tech’s power look like? Short answer: It depends very much on which company you’re going after. THE TARGETS

While antitrust advocates often rhetorically lump Apple, Amazon, Google, and Facebook together, creating a memorable image of four giant “gatekeepers” collectively controlling access to the digital economy, in reality the four companies have very different businesses that raise very different antitrust questions and will lend themselves to very different antitrust solutions.

80

Change

THE BIG BREAKUP

Take, for a start, Apple. It is the most valuable company in the world, as of this writing worth more than $2 trillion. It’s also the most profitable company in the world. And yet, when it comes to discussions of antitrust and Big Tech, Apple often seems like an afterthought. In Wu’s book, Apple barely makes an appearance, and in Senator Amy Klobuchar’s new book, Antitrust, which is a ringing call for remaking and enforcing anti-monopolization policy, the discussions of Apple seem more cursory than central to her thesis. That may be in large part because Apple has become a behemoth mostly on its own—while it has made plenty of acquisitions, its recent growth is mainly due to the simple fact that it has introduced three of the most successful and lucrative technology products in history, and that it has continued to convince customers to keep upgrading to the next generation of products. Even in this new world, it is not illegal to become hugely successful by building the proverbial better mousetrap. To be sure, Apple has antitrust issues, which center on its requirement that all developers who are making apps for the iPhone and iPad sell their goods through the App Store, with Apple collecting a 30% fee. So it’s possible Apple will end up having to let developers sell directly to consumers, or even allow independent app stores. Even so, it could still collect a licensing fee from any app that wanted to be on the iPhone. And most users would, in all likelihood, continue to use the App Store regardless, if only out of habit and convenience. So in the grand scheme of things, Apple wouldn’t seem to have that much to worry about from increasing antitrust pressures.

The Curse of Bigness Antitrust in the New Gilded Age

Antitrust Taking On Monopoly Power from the Gilded Age to the Digital Age

Amazon’s situation is more complicated. It, too, has the fact of organic growth going for it; while it has made its share of acquisitions, it has grown mostly on its own, driven by its relentless appetite for selling more, its huge investment in infrastructure, and its willingness to spend huge amounts of money in order to win and keep customers. Its biggest antitrust problem stems, paradoxically, from something it created itself: Amazon Marketplace. Marketplace was the result of a decision that, at the time, seemed crazy to many: allowing outside sellers to compete with Amazon products and sell on its platform, with Amazon taking a cut of the proceeds. It turned out to be a genius move: Marketplace now accounts for a huge chunk of Amazon’s sales and an even bigger chunk of its retail profits. But Marketplace has also become the place where Amazon’s exercise of power is most visible and most obviously problematic. As Brad Stone details in his new book Amazon Unbound, many Marketplace sellers accuse the company of juking search results to reward those who use its fulfillment services rather than filling orders on their own; rewarding sellers who advertise on the site; boosting Amazon’s own house-brand products in the rankings; and, most famously, using Marketplace data to identify particularly successful products and then mimicking them to undercut Marketplace sellers. Whether Amazon is a retail monopolist is an open question—its total sales remain well below Walmart’s, and even in online commerce its market share is below 50%. But it does unquestionably control Marketplace, and the sellers who use it don’t have many other places to go. That’s why politicians like

Senator Elizabeth Warren have argued that Amazon should be required to spin off Marketplace, while others have suggested that tough regulations be imposed on how it manages the site. Even so, it’s not surprising that when the government was deciding which companies to file antitrust lawsuits against, it went after Google and Facebook first. Those companies are the easiest to fit into a traditional definition of a monopoly—more than 90% of all internet searches are performed through Google, and it and Facebook together control around 80% of the digital ad market. Google’s acquisitions of DoubleClick and ITA played key roles in fueling its evolution. It faces a lawsuit in Europe for tinkering with search results to put its own shopping-comparison engine higher up in the rankings and the sites for rival services lower down. Perhaps most important, Google effectively holds the economic fate of websites across the world in its hands—a change to its search engine or to YouTube’s algorithms can cost people thousands of customers or viewers. None of this might have mattered much in the days when regulators worried mainly about a monopoly’s impact on consumer prices, since just about everything Google does is free to consumers. But under the new antitrust model, the company’s sheer reach makes it a good target. Not, though, as good a target as Facebook. If you had to bet, in fact, on which company is most likely to suffer real consequences from the revolution in antitrust policy, you would be smart to bet on Facebook. It gets 61% of all social media visits in the US. It’s been famously ruthless in snuffing out competitors, either by duplicating their features—as it did with

Review

Snapchat and Twitter—or by simply acquiring them. Its acquisitions of WhatsApp and Instagram look like precisely the kind of anticompetitive acquisitions that regulations were designed to stop. And its lack of transparency about the way it uses customer data has made it notorious. BUT WHAT WOULD A BREAKUP REALLY LOOK LIKE?

The Big Four are unquestionably in the government’s crosshairs. Yet their stocks are more valuable than ever, which suggests that investors, at least, are betting that the antitrust hullaballoo won’t add up to much. Why? One reason is that in going after Big Tech, trustbusters are going after some of the most popular companies in America. Surveys routinely find that Amazon is the most trusted company in the US, with Google and Apple not far behind in the “most admired” rankings. Facebook is the exception; but even if people don’t like it, they find it useful. Antitrust advocates want to take other kinds of harms into account, but they’re not saying that consumer interests should be ignored. And the benefits people get from these companies are easy to show, while the harms they’re inflicting on users can be hard, if not impossible, to define, often resting on somewhat abstract ideas of restricted consumer choice and the costs of lost future innovation. These costs are arguably real, but it’s not obvious they’re enough to build popular support for remedies like breaking the companies up. And while in theory we’re talking about the law, in practice all decisions about what cases to bring and whom to bring them against—and which regulations and remedies politicians

THE BENEFITS PEOPLE GET FROM THESE COMPANIES ARE EASY TO SHOW, WHILE THE HARMS THEY’RE INFLICTING ON USERS CAN BE HARD, IF NOT IMPOSSIBLE, TO DEFINE.

and regulators push for—are shaped by politics, which means in turn that they’re shaped by popular opinion. It’s unlikely any president is going to want to be seen as the person who broke up Google, particularly if it means worse search engines and maps. What this suggests is that even if public rhetoric suggests a campaign to cut Big Tech down to size, we’ll likely end up instead with a series of company-specific remedies. Amazon may have to comply with stricter regulations on Marketplace, including curbs to its power to manipulate its search results or perhaps even its ability to compete with Marketplace sellers. Apple’s monopoly on the App Store may end. Google may face stricter regulations on what it can do with data, and how its search engine’s ranking works. These would not be trivial changes, which is why the companies can be expected to fight them. And yet in most cases, it’s hard to see that they would be transformative. In fact, in recent years these companies have already had to change various questionable practices in response to court cases or inquiries from regulators. It hasn’t kept them from missing a beat. Facebook, which is the least popular of the Big Four, might be different. It may be at risk of the kind of breakup that happened to Standard Oil and AT&T, with Instagram and WhatsApp spun out as independent companies. That would be logistically difficult, since Facebook has worked assiduously to integrate the three services. But it’s not impossible. And it’s a logical, easy-to-understand remedy that might inject some competition into social media. Even so, it’s not clear this would fundamentally dent Facebook’s hold on users, given the

81

treasure trove of data it controls and the power of network effects. Indeed, if the new antitrust movement really wants to change the digital economy, challenging the Big Four’s various sketchy practices is not going to be enough. These companies’ greatest competitive advantage isn’t the legally dubious stuff they’re doing—it’s their perfectly legal access to enormous amounts of detailed and granular user data. That data helps them understand their users better than anyone else and make continuous improvements to their products and services—which in turn helps them keep their current users and add new ones, which gives them access to more data, and so on. It is the key to their growth. Truly challenging the power of the Big Four would mean rethinking how data is gathered and used by companies, and who gets access to it. It might mean requiring that data be shared, that algorithms be transparent, and that consumers have far more control over what they share and what they don’t. For that to happen, the new trustbusters will have to make the case that even if we like what our digital overlords are doing with our data, it’s still wrong for a small number of companies to control so much of it. In a way, they need to make the case that, as in the past, at some point bigness in and of itself is a curse. Big Tech has made that a hard sell in America, simply because the companies have created so much value for consumers. We’re going to find out if that’s enough to keep them safe in this new world. Q James Surowiecki is the author of The Wisdom of Crowds and formerly wrote the Financial Page for the New Yorker.

Change

THE

GERNOT

WAGNER

Cheap solar PV and expensive climate change The rapidly dropping price of solar power has transformed how we think about clean energy.

n late 2007, less than 10 years into the company’s I existence, Google came out swinging on the clean energy front. To a fanfare of plaudits up and down Silicon Valley and well beyond, it declared “RE