• AI's workplace impact: Matthew Call, associate professor at Texas A&M University’s Mays School of Management, examines how AI widens performance gaps between superstars and average employees in organizations, published in WSJ's C-Suite Strategies, to highlight risks to team cohesion and suggest mitigations.
• Superstars master AI first: High performers quickly learn advanced AI features using domain expertise to ask nuanced questions and refine outputs for better results, while average workers use basic functions.
• Expertise enhances AI value: Research shows experts better accept correct AI recommendations and reject incorrect ones, leading to more accurate work compared to peers.
• Systematic habits boost AI: Organized stars provide structured inputs that AI tools respond to effectively, yielding superior outcomes over random approaches by average employees.
• Managerial treatment favors stars: Top performers receive autonomy to experiment with AI early, gaining passes for errors, unlike average workers who await guidance.
• Credit bias disadvantages averages: Stars get outsized recognition for AI-assisted work due to status, while average employees' contributions are undervalued or attributed to AI.
• Encourage AI experimentation: Companies should offer sandbox time, cross-training, and literacy programs on prompt engineering to help all employees leverage AI.
• Level playing field strategies: Share AI knowledge via templates and prompts, redesign evaluations for AI transparency, and train managers to reduce bias in assessing augmented work.

---

While this may be good news for superstars, it’s problematic for companies, because AI-amplified performance gaps will intensify the workplace tensions and resentment that stars can sometimes create, undermining team cohesion—and ultimately hurting the collaborative work that drives business success. Organizations that fail to address this may find their best talent harder to keep and their remaining employees harder to motivate.

How expertise amplifies AI advantages

Think about your own organization. When a new tool arrives that promises to make everyone more productive—advanced Excel features, sophisticated customer-relationship-management systems or powerful analytics platforms—who actually masters it first? It’s usually the superstars who dive deep, discover hidden capabilities and find creative applications no one else thought of, while the average employees tend to stick to basic functions.

AI follows the same pattern as every other workplace tool: The superstars are the first to embrace it. What’s more, research shows that stars also leverage their “domain expertise”—that is, their in-depth knowledge of a subject or business—to extract fundamentally more value (and catch more mistakes) from AI systems than average performers.

SHARE YOUR THOUGHTS

What impact is AI having on your workplace? Join the conversation below.

Imagine a star consultant working on bringing a new product or service to market. Instead of asking AI to “analyze this market” and receiving generic insights, the star uses years of experience to ask more nuances and targeted questions about competitive dynamics, regulations and barriers. Stars’ deep expertise will lead them to better refine the commands or questions they give AI, rather than accepting the first output. This results in more useful and accurate results.

In addition, research finds that employees with more expertise than their peers are significantly better at accepting AI recommendations when they are correct and, more important, rejecting them when they are wrong.

Stars have another advantage: They work more systematically in general, meaning they are more organized and thoughtful in how they approach tasks compared with the average worker. Research finds that these are exactly the kinds of people who get dramatically better results from AI tools than those who just dive in randomly. AI tools respond best to clear, structured inputs—exactly what stars naturally provide through their organized work habits.

Extra credit

The way managers treat superstars only exacerbates their advantages.

The reputation and status of top performers grant them autonomy and discretion in their work, according to my research. That means stars are more likely to dive in and start experimenting with AI immediately. While average employees wait for official guidance or follow company-approved templates for fear of making mistakes, stars will test boundaries, discover creative applications and build personalized workflows long before their organizations catch up. If an AI experiment goes sideways, they are more likely to get a pass—or at least the benefit of the doubt.

Emmanuel Polanco

Then there is the question of getting credit.

Decades of research show high-status individuals gain outsize credit for doing work similar to that of low-status employees. That suggests that when AI assistance is invisible—which it often is—observers are likely to fill in the gaps based on what they already believe about the employee. Stars will get the benefit of the doubt: Their AI-enhanced work becomes proof of their superior judgment and strategic thinking. Average performers face the opposite assumption: If the work is exceptional, AI must have done it.

This creates a vicious double bind for average employees. They’re already less equipped to leverage AI strategically, but even when they do manage to produce outstanding AI-assisted work, they’re unlikely to get the career-advancing recognition that comes with it. Sometimes just the suspicion of AI involvement is enough to diminish how others view their contributions.

How to level the field

So, what can companies do to prevent AI from turning stars into an untouchable class? I suggest three things:

• Encourage everyone to experiment with AI. While stars are quietly building personal AI workflows, most employees are waiting for official guidance that may never come. Smart leaders should create “AI sandbox” time where all employees can test tools without fear of making mistakes, and establish cross-training programs that pair average performers with early adopters.

More important, they should invest in AI-literacy training that goes beyond basic tool usage to include prompt engineering, output evaluation and strategic-task delegation. The goal isn’t to eliminate stars’ expertise advantage, but to teach the learnable skills that can level the playing field.

• Spread the knowledge. Since AI responds best to clear, detailed inputs, leaders need to train average performers to adopt work habits that will enable them to get the most from AI. This means providing templates for organizing information and creating shared collections of effective AI prompts, strategies and use cases. Rather than letting stars hoard their discoveries, make sharing knowledge a standard practice. When one employee discovers an effective AI workflow, capture it and spread it across teams so it becomes something everyone can use rather than one person’s secret advantage.

• Redesign employee-evaluation systems to account for AI-augmented work. The bias that gives stars disproportionate credit for AI-assisted work will only worsen if left unchecked. To fix that, companies should establish clearer guidelines about AI disclosure. They should develop evaluation criteria that fairly assess AI-assisted work regardless of the performer’s existing status. And they should train managers to recognize when bias might be skewing their assessment of an employee’s performance. Consider implementing “AI transparency” practices where teams share how they use AI tools, making the assistance visible rather than hidden.

Without these systemic changes, AI risks creating a two-tier workforce where a small group captures most opportunities and everyone else falls further behind.

Matthew Call is an associate professor in the department of management at Texas A&M University’s Mays School of Management. He can be reached at reports@wsj.com.

C-Suite Strategies

Read the full report

Why AI Will Widen the Gap Between Superstars and Everybody Else

Meet My New Boss. He’s Such a Good Dog.

The Underrated Power of ‘Glue Employees’ Who Hold Everything Together

Support During Menopause Is Becoming the Latest Workplace Perk

What Is the One Trait That Makes for a Great Manager? You Might Be Surprised

How Much Pay Will Workers Sacrifice for Remote Work?

Want to Keep Your Best Employees? Workplace Culture Matters

Next Time You Are Invited to a Meeting, Maybe Don’t Say ‘Maybe’

Talk to enough people of family-starting age, and you’re likely to hear the complaint that the right house is too expensive—or doesn’t exist. A decent house would mean either a 90-minute commute or a move to a city with no jobs.

Along with childcare and education expenses, housing costs were one of the most prevalent factors named among Americans under 40 interested in starting families, according to an Institute for Family Studies survey conducted last year. Twenty-five percent of respondents cited housing costs, while just 8 percent reported lack of supportive family policy as an issue and 10 percent identified debt burdens.

But you don’t need surveys to prove that housing matters. Kids take up space. Americans today have expectations of space and privacy that require extra bedrooms for family formation to occur. Moreover, families desire other housing- and neighborhood-related goods: safety from crime, good schools, attractive parks, and sidewalks or safe streets for children’s bicycles.

We explored these preferences in the IFS survey. Using a series of forced-choice-style questions that required survey respondents to select among various tradeoffs, we found that single-family houses with three or more bedrooms are, without any close competition, the preferred arrangement for American family formation. This result holds for every religious, geographic, political, and demographic subgroup we could analyze. And it’s not just true in the abstract—vacancy rates of starter homes, along with the implied budgetary constraints we imposed on our respondents selecting these homes, suggest that this desire is an almost-universal norm of American culture today.

On that basis, we suggested numerous policies to generate more of this kind of housing. With America’s birth rate at record lows, the least policymakers can do is try to remove needless barriers to the supply of family-friendly homes. We then did a second study, asking, “How can we make real estate development slightly less catastrophically bad for families?”

To understand where this report is coming from, you first need to understand how American housing is changing. The figure below shows the share of new housing accounted for by apartments.

Apartment construction is running at a record high, even as we know that apartments are not where Americans want to raise their families. Even worse, the apartments are getting smaller.

The steady decline in bedroom counts and square footage in apartments over the past 30 years means that the boom in apartments is adding even less family-friendly housing than you might think. The buildings going up don’t contain an even mix of unit sizes: more than half of the units are one bedroom or less, and barely 5 percent are the family-friendly, three-bedroom units that Americans want.

Many reasons explain this shift—some regulatory, some just a feature of the private market. Delayed family formation and a large Millennial generation coming of age in post-Great Recession credit crunch created a massive surge in demand for rental apartments. Developers surged production of this housing.

But many developers failed to realize, first, that this shift was not permanent, and second, that future cohorts would be smaller. Gen Z is smaller than the Millennial generation, and millennials are now marrying and having babies. As a result, vacancy rates for small apartments are rising and now stand at double the vacancy rates for bigger units, even as family-friendly apartments with three or more bedrooms increasingly get bought up as vacation properties and Airbnbs.

We wanted to understand these dynamics better. If we built more family-friendly apartments, would Americans actually be interested in them? Would they pay for them? And would they have more babies?

The answers we found were yes, yes, and yes. Strikingly, bedroom counts matter, but square footage doesn’t. We found that showing respondents scenarios with two bedrooms rather than one increased demand just as much as showing them a $2,000 difference in rent. (Obviously, two-bedroom units don’t go for $2,000 more—because people who can afford that amount move into townhouses or the suburbs, or because the people who prefer those bedrooms can’t afford them.) Square footage variations drove comparatively far smaller differences in preferred scenarios in our survey.

Market prices represent the intersection of willingness and ability to pay for a product. Our survey measures only willingness, and willingness is measured within the survey’s constraints—in this case, that all options presented were apartments.

Nonetheless, this tells us that there is meaningful pent-up demand for more bedrooms. Americans, especially those who have or want children, value bedrooms even more than square footage. The open floor plan is anathema to American family life.

Next, my co-author Bobby Fijan and I produced realistic renderings of apartment floor plans. We showed respondents these floor plans in randomized head-to-head matchups and asked them to select one. While selections were relatively close (partly because the floor plans were similar, with many respondents reporting that they didn’t see a difference between them), the “Extra bedroom” layout was preferred at every apartment size we tested.

When we asked respondents why they made their choice, the answer was clear: people who chose the extra bedroom unit really did choose it because of the number of bedrooms.

We used several other questions to test preferences. We had respondents score the importance of various amenities (bedroom counts clobbered every other amenity). We had them rank floor plans in order of preference (more bedrooms led to better rankings). We asked them about housings costs and ideal housing types.

We found, across every method we checked, compelling evidence that Americans prefer apartments with more bedrooms. For the 750-square-foot floor plans, we even found that childless-by-choice respondents had stronger preferences for the extra-room floor plan than family-minded respondents! And yet over 90 percent of 750-square-foot apartments built in America today will be one bedroom—or none.

When my co-author and I spoke to developers and property management companies about this, we were surprised by their responses. They almost uniformly said they were starting to notice these trends in their own data. The returns on buildings with lots of studio units weren’t as good as anticipated, and family-friendly units were better investments than expected.

Given that most investment in multifamily housing is cautious, slow-moving, and tends to replicate floor plan mixes from prior projects, this response is unsurprising. There isn’t much experimentation in the industry, and with low interest rates, builders didn’t need to deliver highly competitive returns, since the cost of capital was low. This is almost a paradigmatic case of malinvestment.

Unfortunately, it’s malinvestment that will linger in American neighborhoods for decades to come. But we can learn the lesson and improve.

The Low-Income Housing Tax Credit could be rewritten to encourage builders to produce a certain number of low-income available bedrooms instead of units, since children are disproportionately likely to be poor. The Build Now Act being considered by Congress could be revised to reward localities for adding more bedrooms, instead of adding postal addresses. The current proposal, which Republicans have supported thus far, subsidizes localities that convert family houses into multiunit apartments, even if the number of residents in the building declines because three children move out and one renting adult moves in.

Localities that set parking requirements per bedroom can repeal those rules or set them per unit instead. Rules allowing age-restricted units to count for affordable-housing mandates can be repealed. Senior citizens are richer than median Americans—we don’t need to subsidize their housing. Housing authorities that build apartments can target bedroom counts over unit counts.

Private builders should adjust their outlook as well. Family-friendly units will sell. At the same square footage, more bedrooms are in demand, and buildings with those units may be cheaper to hold in the long run.

Some in the real-estate industry have already noticed this and have been pivoting quietly. In our report, we’re shouting it from the rooftops: it’s time for family-friendly apartments.

Top Photo by H. Armstrong Roberts/ClassicStock/Getty Images

The party is in full swing in the Theresienwiese. Thousands of people are packed into scores of tents on the edge of Munich, singing boisterously, dancing on sodden benches and swigging down vast quantities of beer.

Welcome to Oktoberfest, the world’s biggest beer festival, which began in 1810 to celebrate a royal wedding and this year drew over six million revellers from across the world. Like a giant Christmas party, it is filled with good-natured revellers and resounds with goofy German humour. Locals mix with tourists and celebrities. The Bayern Munich football team had been in for a few steins, with Harry Kane dutifully decked out in lederhosen — while Arnold Schwarzenegger was spotted a few days earlier chaotically conducting a band. Armies of waiting staff rush around delivering plates piled high with pork and vast glasses of frothy booze. “It’s better than I thought it would be,” says Osvaldo from Mexico City, as he polishes off his fifth stein. “We don’t have good beer because it is mixed with other shit,” he says. “But here it’s different — you don’t need lemon and all that other stuff.”

Beer, of course, is more than just a drink in Germany. It has special significance; it is part of their nation’s cultural heritage. And it also has economic clout: Germany is one of the world’s biggest beer producers with around 1,500 breweries, many of them based in Bavaria. But beyond the revelry, all is not well for this cornerstone of German culture. For just like the country’s pivotal car industry, the cherished brewing sector is struggling. Sales of alcoholic drinks fell more than 6% in the first half of this year compared with last year. And the average German now consumes only 88 litres of beer a year, down from 126 at the start of the century — a decline fuelled in large part by abstemious younger generations.

Even here at Oktoberfest, the problems are apparent. “I don’t really like beer — it’s so heavy,” confesses one young woman outside a raucous tent. Her friend agrees. Young Germans, she tells me, are significantly changing their approach to well-being. “They want to live healthier.”

Not that everyone is into wellness. In one tent, I come across a lively group snorting what I presume was cocaine from the top of a bald man’s shiny head. One gives me a big grin as he sniffs deeply, then wipes his nose. Two tables down, another ebullient gang of pals pass around a mirror with chunky white lines laid out on it. Minutes later, I pass yet another high-spirited gaggle tapping small piles of powder onto their clenched fists. Yet as Benedict, a young data analyst, tells me when I describe these enthusiastic scenes of debauchery, Oktoberfest “is the biggest drug festival in the world if you count alcohol as a drug”. And it seems many do. In a YouGov survey, almost half the 18-24 age group said they were teetotal — the highest in Europe. Consumption of alcohol among younger teenagers is at an all-time low, storing up future problems for the beer sector.

The German drinks industry is far from alone in confronting such a situation. Times are tough in Britain, too, with scores of breweries and hundreds of pubs closing annually; wine sales in France have crashed, and in the US increasing numbers of the young in the land that once had prohibition are more likely to be teetotal.

But in Bavaria, the heartland of the traditional brewing industry, the crisis is hitting hard.  Schönbrunn, close to the Czechia border, has been the home of the Lang-Bräu brewery for 172 years, surviving some of the most tumultuous events in European history. But after struggling for the past two decades, it has closed. When I visit, the brewer Richard Hopf, 38, explains what it was like to close a family-run firm after six generations. “Some people were crying — most of the village is very sad,” he says, trying to capture the significance of small breweries in these close communities. “In the past there was the church, but for the last 30 years it was the brewery,” he tells me “People would come here, buy your beer, drink one, talk. It was a big place for getting together. Everyone knows each other here.”

Hopf shows me around the maze of rooms where for decades his family fermented, cooled and stored their beer, pointing out proudly the machines which could fill 6,000 bottles an hour, but which now stand silent. This plant used to produce 1.5 million litres annually, exporting about 20% of their output to China and Korea at one point — although most sales have been within a 30-mile radius.

As we tour the ghostly plant, he tells me about his father winning a burst of global media attention after making “Erotik bier” — with a highly-suggestive label — as a joke. But those carefree days are long gone. As the cost of energy and staff rose in recent years — and with cut-throat supermarkets dominating sales — beer prices remained largely static. “The end result,” he says wistfully, “is the market just goes down and down and down.” And in such a climate, the family firm could not get the financing needed to modernise — so these days their storage rooms are stacked high with empty bottles.

While the closure has been painful, Hopf and his brother were left with little choice. The machinery and buildings were constructed in the decades after the Second World War and were too dated for efficiency. “You must think 20 years in the future,” he explains. “You have to forecast the beer you might sell then, but we knew things were only going to get worse because the baby boom generation is dying and younger ones drink less beer.”

It is a similar story in Bamberg. A hundred kilometres away, set in rolling hills, it reflects centuries of history and is famed for its number of breweries and its diversity of beers — including the local speciality, Rauchbier, a smoked draft served straight from oak barrels. Inside Schlenkerla, a bustling time-warp tavern, one manager explains the appeal beer’s appeal, telling me their brands sell around the world. “I was recently in Lviv, in Ukraine,” she says, “and I saw a whole shelf of our beer there.”

But having survived the war largely unscathed, its breweries know they face a struggle. And it soon becomes clear that this 28-year-old saleswoman is emblematic of the sector’s problems. “Bavarians are more likely to drink beer than other parts of Germany but I prefer wine,” she tells me. “I don’t drink so much personally. I am very boring. And all my friends are into sport or health. They have maybe a beer at the weekend with their dinner. But when you work here it seems like only older people drink.” When I ask her what she meant by “old”, she replies: 40.

The manager of an Irish pub over the road echoes the sentiment. “I know it is a national symbol but it’s not for me,” he confesses. In Wirtshaus Graser, the town’s newest microbrewery, the barmaid says business is good — but even she sounds a warning, admitting that Gen Z Germans seem far more abstemious than their predecessors. “I am from the generation that partied and drank a lot during their teenage times.” Her younger sister, she says, “is not as adventurous as we were”.

Falling sales, rising costs and generational change have created an existential crisis for German beer makers — and sales of the alcohol-free products that now comprise almost one-tenth of their output will not fully plug the gap. But exports are falling even faster than the intensifying domestic decline, leading the head of one major brewery to warn two months ago about a “landslide” that would sweep away both big and small firms. “The brewing world is crumbling,” says Stefan Blaschak, managing director of Oettinger, which is closing one of its four plants. It’s a huge dent in the self-image and confidence of the nation that powers the European economy — something which cannot be masked by all the beer-soaked revelry, drinking chants and sea of lederhosen at Oktoberfest.

Concealed by the jollity of Bavaria’s biergartens and the frenzied Oktoberfest partying, another national champion is flailing in an economy that is strugging to adapt to a world of rapid political, societal and technological change. The number of German jobless hit 3.02 million in August, the highest figure for a decade, as the manufacturing giants that drove the post-war economic boom shed staff. Industrial output slid 4.3% on the previous month as car production crashed to its lowest level since the start of this century. All those dirndl-clad waitresses clutching overflowing steins of golden Helles might be the nation’s fetching global image, but it is pretty sobering to see that Germans, of all people, are losing their thirst for beer.

---

• AI Actress Introduction: Tilly Norwood, an AI-generated actress created by London-based Particle6 Productions, debuted at the Zurich Film Festival recently in a satirical video, sparking debates on technology's impact on Hollywood.
• Media Backlash: Over 107 negative stories appeared in print media shortly after, criticizing the AI gimmick despite its success in generating global publicity.
• Emily Blunt's Reaction: Actress Emily Blunt urged studios and agencies to avoid AI actresses like Tilly, expressing discomfort during a Variety interview.
• Opportunities in AI: AI could create new interactive opportunities benefiting existing stars, while failure to adapt might advantage upstart studios abroad.
• Existing Tech Acceptance: Hollywood has not objected to technologies that digitize and distribute performances globally, disrupting markets for other entertainers.
• Protectionism Critique: Trump's proposed 100% tariffs on foreign films echo SAG's resistance, potentially harming U.S. industry's competitiveness like past protectionist failures.
• Historical Reforms: Bipartisan efforts by Carter and Reagan, including deregulation and tax cuts, spurred economic growth, venture capital, and innovations leading to AI like Tilly.
• Path Forward: Embracing innovation counters debt crisis warnings, unlike reactionary opposition to Tilly, which stifles progress seen in past surpluses and recoveries.

---

BPC > Only use to renew if text is incomplete or updated: | archive.fo

BPC > Full article text fetched from (no need to report issue for external site): | archive.today | archive.vn

Tilly Norwood, an AI-generated 'actress,' smiles in an AI-generated image.

But what is the case _against_Tilly Norwood?

Why should screen actors be exempted from adapting to new technologies when adapting has been the fate of industrial man for three centuries?

Not a she but an it, Tilly Norwood was created by the London-based Particle6 Productions. Introduced at the Zurich Film Festival a couple of weekends ago, she appeared fleetingly in a two-minute satirical video. One observer said the 20-something computer-generated actress might be cast as a “sweet kindergarten teacher” or a “non-mean sorority sister.”

Cue the algorithmic press outrage. Across 107 print stories in the Factiva database in the next few hours, all were negative, all in the same way. One writer, on deadline in New Zealand, was at least awake enough to notice the obvious calculation by Tilly’s creators. Their gimmick paid off beyond all expectations in global publicity.

At the opposite extreme? In the frog, an optic nerve responds to the movement of a fly. Out pops the tongue—or, in this case, out pops a New York Times column uncritically linking the pathos of the Screen Actors Guild to the AI doom coming for all of humanity.

Thanks to an accident of podcasting, the actress Emily Blunt would become the global voice of Hollywood reaction. On being told of Tilly’s existence by an interviewer for Variety, “Please stop,” she insisted, while demanding that mainline studios and talent agencies refrain from trafficking with AI actresses.

Except, what if there’s a real opportunity here? Shouldn’t somebody try to find out? Might not failure on this score hand the future to upstart studios and film industries in other countries?

Too, I don’t remember movie types objecting to the technology that already allows their images and performances to be turned into data, massaged in a million ways, then distributed cheaply to global customers in a manner that necessarily upsets the market for myriad other potential entertainers.

The hysterical posturing is even more implausible when you realize existing stars will be AI’s biggest beneficiaries when it spawns lucrative new interactive opportunities.

Like many Trump-era kerfuffles, the Tilly Norwood panic shows signs of disappearing as fast as it came. Maybe it shouldn’t.

It lands as the Trump-whispering actor Jon Voight has renewed his push for tax breaks to counter the Los Angeles-based movie industry’s precipitous decline. It comes as Mr. Trump revives his own ill-advised proposal from earlier in the year. His idea of 100% tariffs on TV shows and films shot or edited elsewhere would test a proposition proved too many times. Protectionism is among the surest ways to set a globally competitive domestic industry on a path to ruination.

Ms. Blunt might be shocked how much she and Mr. Trump are birds of a feather when it comes to believing that blocking competition is the way to help a struggling Hollywood.

Our political lore tends to draw false dichotomies between partisan eras. It was not either party but the Democratic-Republican duo of Johnson and Nixon that created the problems of the 1970s economy.

It wasn’t one party or the other but the Carter-Reagan duo that rescued the economy, though honorable mention is due Ford advisers who first hatched deregulation.

In 1976 Jimmy Carter gave no hint that he would spearhead major regulatory reforms across the transportation and energy industries or appoint inflation hard-liner Paul Volcker to the Federal Reserve. Mr. Carter signed a capital-gains tax cut that would give rise to the modern venture-capital industry and spur the technology that led to Tilly.

In a televised debate on Sept. 23, 1976, Mr. Carter said the words, “The best way to control inflation is to cut down government waste . . . and to let the American free-enterprise system work freely and effectively.”

Nobody, including him, guessed he’d given voice to the animating spirit of American reform that would transcend the parties over the next 20 years. A Hollywood movie will never be made about it. The growth recovery that began thanks to the Carter-Reagan reforms produced four consecutive budget surpluses in the second Clinton term. Bill Clinton’s Republican Fed chairman, Alan Greenspan, would fret about the national debt disappearing faster than the financial plumbing could adjust to its absence.

This story, which might appeal to Clint Eastwood, shows America’s problems have a solution. The country today is widely understood to be headed toward some kind of debt or spending meltdown. Economists have warned for years that such a crisis is inevitable. But an alternative exists. It won’t be found, though, in the reactionary loud-mouthing against innovation and opportunity that greeted the AI actress Tilly Norwood.

• Lisa Su's OpenAI Deal: AMD CEO Lisa Su, 55, born in Taiwan and immigrated to US at age 3, secured a megadeal with OpenAI this week for 6 gigawatts of AI chips, generating tens of billions in revenue, positioning AMD against Nvidia in AI technology.
• Early Life and Education: Su repaired her brother's remote-controlled car as a child, attended Bronx High School of Science, and earned electrical engineering degrees from MIT, developing interest in semiconductors.
• Career Progression at AMD: Joined AMD in 2012 as senior vice president, became COO, then CEO in 2014, turning around the company from near bankruptcy with market value rising from $3 billion to $348 billion.
• Engineering Focus: As an engineer's engineer fascinated by device physics, Su emphasized designing powerful semiconductors, avoiding costly manufacturing like Intel, and targeting high-performance computing for AI.
• Market Share Gains: Under Su, AMD captured 41% market share in data-center CPUs, launched MI300 GPU series in 2023 for AI, and powers gaming consoles like PlayStation and Xbox.
• Response to Criticism: After a critical SemiAnalysis article on AMD's software bugs, Su engaged directly with the author for 90 minutes, acknowledged issues, and publicly thanked for feedback as a gift.
• Company Background: AMD, founded in 1969 by Fairchild Semiconductor veterans, historically trailed Intel but under Su shifted focus from mainstream to high-performance parts in data centers and AI.
• Rivalry with Nvidia: Nvidia holds over 75% AI chip market with CUDA software advantage; its CEO Jensen Huang praised Su's turnaround but questioned the OpenAI deal's stock giveaway structure.

---

BPC > Only use to renew if text is incomplete or updated: | archive.is

BPC > Full article text fetched from (no need to report issue for external site): | archive.today | archive.is

As a child, Lisa Su opened up her brother’s remote-controlled car when it broke, and, as she tells it, identified a loose wire connection and taught herself how to repair the toy.

In 2021, she became the first woman to be awarded the Robert N. Noyce medal, considered the Nobel Prize of microelectronics. This week, she muscled in on the battle over the newest and most advanced technology in the world that has been dominated by boldfaced names like Sam Altman and Jensen Huang.

Su pulled off a megadeal with OpenAI, unveiled Monday, that positions her company, Advanced Micro Devices, to take on archrival Nvidia. Now the market is watching for her to strike additional deals and sign up more big artificial-intelligence firms as customers for AMD’s chips. The deal caps Su’s decadelong turnaround effort. 

As a CEO in the old Silicon Valley mold—an engineer who intimately knows the technology she’s selling—Su is deftly navigating challenges that have tripped up other tech-industry leaders. While rivals like Intel got bogged down in costly manufacturing investments, Su focused squarely on what she believes AMD does best: designing ever more powerful semiconductors. She has transformed AMD into the most serious competitor to Huang and Nvidia, the 800-pound gorilla of the AI chips industry.

Since taking the helm in 2014, she’s pulled AMD back from the verge of bankruptcy and built the company’s market value from less than $3 billion to an astounding $348 billion.

Under the terms of this week’s agreement, Altman’s OpenAI, the maker of ChatGPT, is expected to buy 6 gigawatts worth of advanced AI chips, handing AMD tens of billions of dollars in revenue over the next few years. The company’s share price rose more than 33% this past week on the news. 

From left, Sam Altman of OpenAI, AMD’s Lisa Su and Michael Intrator of CoreWeave arrive to testify at a Senate committee hearing on AI in May.

“If you ask, what do we do when we wake up in the morning, it’s about, how do we push the envelope in high-performance computing?” Su recently told the venture capitalist Reid Hoffman in a podcast interview. “And it turns out that today, that high-performance computing is all about AI.” 

Born in Taiwan to a mathematician father and accountant mother, Su, 55 years old, immigrated to the U.S. in the early 1970s, when she was 3 years old. Her family settled in Queens, N.Y., and she attended the prestigious Bronx High School of Science. Su is a distant cousin of Huang, the co-founder and CEO of Nvidia, but didn’t meet him until well into her career. 

She later obtained bachelors, masters and doctoral degrees in electrical engineering from the Massachusetts Institute of Technology, where she grew interested in computer chips while working in a semiconductor lab as an undergraduate. 

Early in her career, she worked at IBM, helping design some of the earliest copper-wire interconnections for semiconductors and learning how to bring new technology to market as a product manager. She joined AMD in 2012 as a senior vice president and quickly rose to chief operating officer.

Su has described herself as an “engineer’s engineer” who is fascinated by device physics, a discipline that studies how machines can interact with tiny particles like electrons and protons. Her understanding of the nuts and bolts of microelectronics has primed her to lead AMD at a time when building AI supercomputers requires finding increasingly inventive ways to cram billions of transistors that are a fraction of the width of a human hair onto postage-stamp sized slices of silicon.

“At the end of the day, the beauty of engineering is, it actually matters what you do,” Su told Hoffman in her interview.

Although AMD is based in Santa Clara, Calif., Su lives with her husband in Austin, Texas, and commutes to Silicon Valley several times a month. 

Su, in 2015, launched a turnaround that built AMD from less than $3 billion in market value to $348 billion.

Just before last Christmas, SemiAnalysis, an influential tech research shop, published a long, technically complex article comparing AMD’s newest AI chips unfavorably to Nvidia’s. While AMD’s hardware is sophisticated enough to compete with its bigger rival, the article said, its software platform was full of bugs and impossible to use in large-scale AI training. 

The day after it published, Su called up the author, SemiAnalysis founder Dylan Patel, and spent an hour and a half going over his arguments. She acknowledged problems in AMD’s products and listened to Patel’s detailed recommendations for how to fix them. Afterward, she thanked Patel for his help in a social-media post. 

“Feedback is a gift, even when it’s critical,” she wrote. 

AMD was founded in 1969 by Jerry Sanders and other veterans of Fairchild Semiconductor, a pioneering developer of what were known as integrated circuits, the forerunners of modern chips. For the next two decades, AMD was consistently playing catch-up to Intel, another spawn of Fairchild that created the microprocessors that were used in most personal computers starting in the 1980s. 

When Su took over as CEO, AMD had “lost the plot” when it came to data-center chips, and its market share was near zero percent, said Forrest Norrod, AMD’s head of data-center products and her first hire as CEO. Su developed a systematic approach and today AMD has a market share of about 41% in data-center CPUs, the central computer brains that power servers used for cloud computing and AI functions, he said. 

“AMD before Lisa was trying to chase the bulk of the market, the mainstream parts rather than high-performance parts,” Norrod said. “She understood that our customers care the most about the high-performance parts, and if you’re focused on the bulk of the market, then your competitors can outmaneuver you.”

AMD founder Jerry Sanders holds up a silicon wafer in 1997.

“When you’re talking to her, she’ll test you on what you’re thinking both on the business side and the technology side,” he said.

She quickly set teams to work designing new graphics cards for videogames—AMD’s chips power PlayStation and Xbox consoles, among other gaming systems—and more powerful data-center and PC chips, steadily eating away at Intel’s market share. Intel meanwhile fell behind.

AMD’s biggest foray into artificial intelligence began in late 2023 with its launch of the MI300 series of graphics processing units, or GPUs, the chips required for the advanced computing workloads behind AI.

Nvidia controls more than 75% of the AI data-center chip market, most analysts estimate, and has a major head start over rivals, in part because of its proprietary CUDA software system, which provides thousands of developers with a streamlined programming language that they can use to create applications using only Nvidia’s chips. 

Huang, the Nvidia CEO, has praised Su for her AMD turnaround and called her “terrific.” This past week, however, he questioned the wisdom of her new deal with OpenAI, under which OpenAI stands to receive 160 million shares of AMD stock in phases if certain milestones are met. 

“Considering they were so excited about their next-generation product, I’m surprised that they would give away 10% of the company before they even built it,” Huang said of AMD in an interview on CNBC. “Anyhow, it’s clever, I guess.”

Write to Robbie Whelan at robbie.whelan@wsj.com

• Putin's Succession Overview: RAND commentary addresses Vladimir Putin's potential death as Russia's longest-serving leader since Stalin, turning 73 on October 7 in Moscow, urging Western preparation due to risks of instability amid ongoing Ukraine war and nuclear concerns.
• Russian Political System: Describes authoritarian dictatorship with personality cult features, centered on ageing loyalists controlling key sectors like security, energy, and banking through personal loyalties rather than formal laws.
• Potential Successors: Lists figures from Kremlin, government, security, military, energy, banking, and Putin's inner circle, such as Nikolai Patrushev, Mikhail Mishustin, Igor Sechin, selected via Kremlinology for proximity to Putin.
• Loyalty Dynamics: Emphasizes loyalty over competence, with circle of trust shrinking; examples include Sergei Shoigu's soft landing after Ukraine invasion mishandling, tying elites' fates to Putin alone.
• Transition Risks: Predicts surprises in succession, drawing on historical precedents like Yeltsin's 1999 choice and Soviet collapse, potentially leading to chaos, violence, or elite power grabs as in "The Death of Stalin."
• Western Preparation Steps: Recommends scenario planning, wargaming by NATO and EU to test responses to instability, nuclear control, regional unrest, and Russia-China shifts.
• Intelligence and Expertise: Calls for enhanced mapping of elite networks, assets, and rivalries using diplomats, academics, and defectors, informed by Russian political culture traits like autocracy and kleptocracy.
• Strategic Measures: Advocates reinforcing deterrence in Ukraine and Europe, improving information messaging, and viewing succession as opportunity for de-escalation while avoiding past misjudgments on Russia's democratic transition.

---

Benjamin Franklin wrote that nothing is certain but death and taxes. In Russia, one might quip that only death and the overwhelming power of the tsar are constant. But what happens when the tsar dies? Vladimir Putin is the longest-serving Russian ruler since Joseph Stalin. He turns 73 on Oct. 7, old for a Russian, but even if he is fanatical about his health, he cannot continue indefinitely. It is essential to consider who might succeed him.

This is a challenging task—under Putin's rule, the Russian political system has, once again, become an authoritarian dictatorship bearing some of the features of a personality cult. Under these conditions Putin has not, as far as we can tell, been publicly grooming a successor, presumably because his personal authority would begin to ebb to them, and because they would become a target for others who fear losing their influence. Under Putin, the Russian system of governance centres on a cadre of mainly ageing men (and, increasingly, their sons and daughters), who serve as custodians of the strategic verticals of the political economy. These include the presidential administration, security services and armed forces, as well as the energy, industrial, technocratic and banking sectors, among others.

Conceptually, this system can be understood as a “limited access order”, in which those with the potential to challenge for power are bought off with rent-generating assets. These men owe their positions personally to Putin, and they have been allowed to enrich themselves through their verticals in return for their loyalty. In this system, personal understandings and customs are more important than formal rules and laws.

A longlist of potential successors can be drawn from these men using old fashioned Kremlinology—collating observations about their proximity to and apparent favour with Putin—to divine their respective prospects. As imperfect as it may be, this approach is useful for understanding the ups and downs of the Putin court.

Assessments can be made of various figures from, for example, the Kremlin (such as Sergei Kiriyenko or Nikolai Patrushev), the government (Mikhail Mishustin or Marat Khusnullin), the security services (Sergei Naryshkin), the military industrial complex (Andrei Belousov or Sergei Chemezov), the energy complex (Igor Sechin), banking (Andrei Kostin) and Putin's old friends (Yurii Kovalchuk and Arkadii Rotenberg). Such men have their own power bases and are surely among the plausible runners and riders.

These men are “made”, in so far as they have served Putin for years and appear destined to remain with him to the end. As Putin ages, it is quite natural for him to stick with the loyalists around him. Loyalty trumps competence in Putin's Russia, as demonstrated when Sergei Shoigu, who botched 2022's full-scale invasion of Ukraine (but who has bonded with Putin on holidays in Siberia), was belatedly removed as defence minister but landed softly, becoming Head of the Security Council. Putin has brought loyalists closer and closer to him—in autocracies and kleptocracies the circle of trust shrinks over time.

However, observational analysis of the men in that circle is only useful to a point. In fact, the loyalty of these men ought to raise questions about their succession prospects. Russia's limited access system works by tying the fate of the elite to the leader alone: when he dies, they will almost certainly see themselves as vulnerable and, in fact, little is known about how they would behave in a transition scenario. Due to their associations with the previous regime, they will surely have to protect themselves immediately.

History suggests that the transition may throw up surprises. Few Kremlin watchers foresaw that Putin would be picked by Yeltsin in 1999 or, for that matter, that Medvedev would step in for Putin in 2008. Indeed, the collapse of the Soviet Union was characterised by implosion: regional power grabs, a rush to asserts control over strategic sectors of the economy, and Gorbachev leaving Moscow while events unfolded out of his control. In the comedy The Death of Stalin, potential successors to the Vozhd compete in a chaotic grab for power. This parody may be as instructive as anything Kremlinology can offer. Putin might control everything now, but vertical logic means that no-one can be sure of anything in the moment he dies.

Western countries should be preparing now to handle a transition that, due to Russia's governance structure and political culture, could be surprising, drawn out, and violent.

It follows that Western countries should be preparing now to handle a transition that, due to Russia's governance structure and political culture, could be surprising, drawn out, and violent. Unlike The Death of Stalin, the prospect is, of course, no laughing matter, especially considering the need to ensure that the world's largest stock of nuclear weapons remains in safe hands. Indeed, the transition will be strategically crucial, given the ongoing war in Ukraine and China's interest in the outcome. Changes in the Kremlin will also have significant impact in other regions, e.g., affecting the stability of the regime in Belarus, or the fate of pro-Russian elements in the Caucasus, Central Asia, Africa, and elsewhere.

How can NATO allies prepare for a Russian succession? As it stands, there is very little information available publicly about this, given that attention has been on the war in Ukraine. This is concerning, as the departure of Putin will not only be a Russian drama but a strategic shock to Europe, NATO and the wider world. It requires concrete planning.

The first step is to embed scenario planning and wargaming into Western institutional thinking on the issue. NATO and the EU should run regular joint table-top exercises to test assumptions and rehearse responses—diplomatic, informational, military and economic. Scenarios should range from sudden instability in Moscow, to elite rivalries spilling into violence, to a more managed succession. Planners must consider not only nuclear command and control, but also the possibility of regional unrest, opportunistic moves, or shifts in Russia's alignment with China. They must also test how potential responses might be interpreted by various Russian cadres, and the population at large.

Second, and in part to inform such wargames, Western governments need to improve intelligence on Russia's elites. Succession will be determined not in public, but within opaque patronage networks. Mapping these networks, tracking their assets, and understanding their rivalries will be essential for anticipating potential contenders. Western states should invest in a deeper pool of expertise, drawing on diplomats, academics, and Russian defectors to provide insight into the dynamics of the inner circle. Certain traits in Russian history, political culture and political economy provide a framework for thinking about the potential interactions, factions and contests that could emerge in various scenarios: autocracy, kleptocracy, orthodoxy, militarism, imperialism, and other Russian pathologies should inform this work.

Western states should invest in a deeper pool of expertise, drawing on diplomats, academics, and Russian defectors to provide insight into the dynamics of the inner circle.

Third, the West must reinforce its front line. A leadership transition could tempt Moscow to externalise its instability, through escalation in Ukraine, hybrid operations in Europe, or threats to NATO members. Maintaining deterrence on the eastern flank, accelerating support for Ukraine's military modernisation, and closing loopholes in sanctions enforcement will all be necessary to prevent Russia exploiting a moment of uncertainty. In the 1990s, Western leaders came to the belief that Russia would transition from totalitarianism to liberal democracy and a market economy underpinned by the rule of law; a misreading that should not be repeated.

Fourth, it will be vital to prepare for an intensification of the information battle. A succession struggle will generate confusion, rumour and competing narratives aimed both at Russian society and international audiences. Crafting credible and coherent messaging in advance (about accountability for the war, about possible pathways to de-escalation, and about the unity of the Western alliance) will help shape the environment in those crucial first hours and days.

Finally, and cautiously, policymakers must see that succession is not only a risk but also an opportunity. The danger of instability, renewed militarism or competition among the elite is real. At the same time, the transition could be smooth or even present an apparent reformer. It may leave Russia weakened, inward-looking, or perhaps open to some change. A range of eventualities can be stress-tested using techniques such as back-casting, scenario analysis, and assumptions-based planning. These methods can provide the evidence base for a unified approach that enables Western partners to seize emerging opportunities while avoiding the mistakes of the 1990s and 2000s.

Putin's absence will not by itself make Russia less dangerous. Without foresight the West risks being caught unprepared. Strategic stability in Europe's future depends on treating Putin's mortality as a concrete policy challenge for which preparations must begin now.

• Share on LinkedIn
• Share on X
• Share on Facebook
• Email