In future table tennis players might find an AI-powered robot across the table, as one just defeated some top-ranked amateurs. CC-licensed photo by Gaël Marziou on Flickr.
You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.
A selection of 9 links for you. Table stakes. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.
Meta to lay off 10% of workforce in AI push • The New York Times
Mike Isaac and Eli Tan:
»
Meta plans to cut 10% of its work force, or roughly 8,000 employees, and close another 6,000 open roles, according to an internal memo on Thursday, as the company spends heavily on developing artificial intelligence.
Meta, which owns Facebook, Instagram and WhatsApp, employed more than 78,000 people at the end of 2025. Mark Zuckerberg, Meta’s chief executive, has said he expects much of the work done in the technology industry to eventually be overtaken by AI-powered systems, including coding assistants that help engineers write software.
“We’re doing this as part of our continued effort to run the company more efficiently and to allow us to offset the other investments we’re making,” Janelle Gale, Meta’s chief people officer, said in the memo to employees. “This is not an easy trade-off and it will mean letting go of people who have made meaningful contributions to Meta during their time here.”
A spokesman for Meta confirmed the cuts and declined further comment.
Across the technology industry, companies have been laying off employees as they experiment with AI. In February, Block, the financial technology company that owns Square, Cash App and Tidal, said it was cutting 40% of its work force as it embraced new AI tools. Microsoft on Thursday said it was offering buyouts [redundancy] to 7% of its work force as it invests in AI.
Mr. Zuckerberg is reorganizing his company around AI products in a fierce race to lead in the technology against rivals like OpenAI, Google and Anthropic. He has made no secret of his AI ambitions and has described developing AI-powered social media products that are a kind of “personal superintelligence” that he hopes people will incorporate into their daily lives.
“At Meta, we have the resources to build the massive infrastructure required and the ability to deliver new technology to billions of people,” Mr. Zuckerberg said in a video posted to his Facebook page in July.
«
So the idea is that technology workers will be replaced by the thing they’ve created. It’s an odd version of Frankenstein. The tech firms that are investing heavily in AI seem to expand and contract like an old-fashioned bellows, trying to get the fire going and then putting people aside as soon as the flames seem to be alight.
unique link to this extract
Microsoft offers voluntary retirement to long-serving employees • The Verge
Tom Warren:
»
Microsoft is changing up its annual rewards and performance programs today, and offering long-serving employees in the US the ability to voluntarily retire. It’s the first time in Microsoft’s more than 50-year history that the company has offered such a voluntary retirement program.
“Many of these employees have spent years, and in some cases, decades, shaping Microsoft into what it is today,” says Microsoft’s HR chief Amy Coleman in a memo seen by The Verge. “For those who may be considering their next chapter, we’re offering a one‑time Voluntary Retirement Program.” Microsoft says it applies to only a “small percentage of our US employees.”
US employees whose combined years of service added to their age totals 70 or more will be eligible for voluntary retirement, and Coleman says this will include “generous company support.” It’s not clear if this is a precursor to more layoffs at Microsoft, but it certainly looks like a method to avoid a bigger round of layoffs ahead of Microsoft’s new financial year in July.
«
That “service plus age” thing is intriguing. You’re aged 50, you’ve been there for 20 years. You’re 45, you’ve been there 25 years – so you joined aged 20. That seems like the ceiling there. (John Gruber points out there’s someone at Apple who totals 114, which I believe is Chris Espinosa: joined aged 14 in 1976, now aged 64 plus those 50 years of service.
Anyway, Microsoft seems to think that old is bad now.
unique link to this extract
You’re about to feel the AI money squeeze • The Verge
Hayden Field:
»
Between 2024 and 2029, [senior research analyst Will Sommer] said, [research company] Gartner estimates that capital investment in AI data centers will reach about $6.3 trillion — a “massive amount of money.”
To avoid a write-down of these assets, major AI model providers would ideally generate a return on invested capital (ROIC) of about 25%, Sommer said. (That’s about what Amazon, Microsoft, and Google tend to earn on their overall capital investments.) On the other hand, if the returns fall below 12%, institutional capital loses interest — there’s better money elsewhere, Sommer said. Below 7%, you’re in writedown territory, which is “an unmitigated disaster for all of the investors in this technology,” Sommer said.
To reach that bare minimum of 7%, Gartner forecasts that large AI companies would need to earn cumulatively close to $7 trillion in AI-driven revenue through 2029, which is close to $2 trillion per year by the end of the period. In order to achieve “historic returns,” the providers would need to earn nearly $8.2 trillion in the same period.
OpenAI has already made $600bn in spending commitments through 2030, the company said in February, which Sommer says is already a “massive step down” from the $1.4 trillion it had planned before. Based on OpenAI’s revenue forecasts and potential compound annual growth, Sommer said that even in the best-case scenario, he predicts that the lab would only hit a fraction of the overall spend required to hit that 7% ROIC.
How do model providers like OpenAI make this money? By selling access to what are known as tokens. A token is essentially a unit of data input that an AI model can understand and process — it could be text, images, audio, or something else. One token is generally worth about four characters in the English language — the word “bathroom,” for instance, would likely be processed as two tokens. One paragraph in English is generally about 100 tokens, and a 1,500-word essay may be about 2,050 tokens, per an OpenAI estimate.
To hit investors’ revenue expectations, providers would need to process a “mind-bending” number of tokens, Sommer said.
By most measures, companies’ numbers are already pretty big. Google announced it was processing 1.3 quadrillion tokens in October, for instance. If you add all the providers’ estimates up, Sommer said, you get 100 to 200 quadrillion tokens a year. But to achieve the the $2 trillion in annual spend Gartner calculated, providers would need to be generating, by conservative estimates, a cumulative 10 sextillion tokens per year. (To make that slightly less abstract, a quadrillion has 15 zeros, and a sextillion has 21.)
«
There just doesn’t seem to be any way to meet those targets. The writedown is coming; the only question is when. Plus Nilay Patel points out that AI is not popular. Hard to generate trillions in revenue from something that is not popular.
unique link to this extract
Intellectuals are f*cking idiots • Mark Manson
Mark Manson:
»
Malcolm Caldwell was the consummate intellectual. He had spent his entire life studying Southeast Asian history and economic development. He had written hundreds of articles and over a dozen books on the subject. He was a professor and researcher at one of the most prestigious universities in the world and was celebrated and supported for his views.
Much of his work dealt with English colonialism in Asia and its dire political consequences. As a result, Caldwell evolved into a staunch Marxist, far to the left of the leftiest leftist who ever lefted.
Just to give you an idea how far left we’re talking, Caldwell visited North Korea in the 1960s and came away saying good things about it. When the Vietnam War started, Caldwell tried to host a fundraiser in London… for the Vietcong.
So when communist revolutionaries took control of Cambodia, Caldwell showed enthusiastic support. The new communist leader of Cambodia was a man by the name of Pol Pot and he had radical new ideas of how to achieve a communist utopia — ideas that had existed in Marxist thought but had yet to actually be attempted in any communist country. Caldwell had been waiting for decades for a communist revolutionary who fully implemented his Marxist dreams. Caldwell came to believe Pol Pot was his man.
…when news of the genocide and atrocities began to leak out of Cambodia, Caldwell refused to believe it. He defended Pol Pot’s regime and wrote off the atrocities as simply more western capitalist propaganda. His unwavering support eventually earned him an exclusive invitation to visit Cambodia by Pol Pot’s government. Caldwell accepted. And in December of 1978, he boarded that fateful flight to Asia.
Once there, Caldwell toured the country. He met the leadership and learned about their policies firsthand. But the climax of his trip was the last evening — a private audience with Pol Pot himself. Reportedly, Caldwell was “euphoric” with excitement and anticipation. Once in private, Caldwell and Pol Pot had a long intellectual conversation. In his enthusiasm, Caldwell began sharing some of his ideas for the Cambodian regime. He began to offer feedback and dare I say, potentially even a little criticism. Pol Pot, not used to being lectured to by a professor, promptly had Caldwell killed that night.
Malcolm Caldwell is what I like to refer to as an intelligent idiot. A man with an encyclopedic breadth of knowledge and understanding, a world-class mind with powerful thoughts, and yet absolutely no idea how to apply any of it.
The world seems to be full of intelligent idiots. The examples are endless.
«
He lists a few of them. And they’re interesting.
‘Staggering’ number of people believe unproven claims about vaccines, raw milk and more • Nature
Helen Pearson:
»
More than two-thirds of the public believe at least one false or unproven health claim, such as the idea that taking paracetamol during pregnancy causes autism, a new survey finds. The results hint that a large, and potentially growing, number of people are questioning scientific evidence.
The survey, of more than 16,000 people across 16 countries, asked whether they believed claims that are not supported by research, including that the “risk of childhood vaccinations outweighs benefits”, “fluoride in water is harmful” and “raw milk is healthier than pasteurized”.
For each statement, between 25% and 32% of respondents said they believed it, and another sizeable percentage (17–39%) said they didn’t know whether it was true. In total, 70% of respondents believed at least one of the claims. The findings, which have not been peer reviewed and were published on Wednesday by the Edelman Trust Institute in New York City, were described as ‘staggering’ in an accompanying article by the think tank’s chief executive, Richard Edelman.
The result “blows the lid off of this idea” that such beliefs are held by only a fringe population of individuals who are uninformed or ideologically driven, says David Bersoff, head of research at the Edelman Trust Institute. “This is not like a small problematic group.”
“There has definitely been a growing number of people who question widely accepted scientific evidence,” agrees Heidi Larson, who studies confidence in vaccines at the London School of Hygiene & Tropical Medicine. “It’s important to pay attention to.”
Other recent studies have highlighted how commonly people question scientific consensus or evidence-based medical practices, at least in certain contentious areas, such as vaccines. One global 2023 study found that during the COVID-19 pandemic, people’s confidence that vaccines are important for children fell in 52 of 55 countries.
«
Experts suggest there’s actually too much information: people can’t decide what is true. That suggests they’ve never learnt how to parse factual information. Anecdotally, I (and others) observe that literacy, especially among Americans, is plummeting. The combination is potentially lethal.
unique link to this extract
AI robot outplays humans in table tennis milestone • Financial Times
Michael Peel:
»
An AI-powered robot has beaten expert table tennis players in a landmark machine-over-human triumph in a major competitive sport.
The mechanical maestro, known as Ace, uses a network of cameras and AI to achieve the rapid planning and reaction times needed to compete.
The invention made by Japanese tech group Sony highlights how researchers are using AI to improve robots’ ability to adapt to physical tasks they have struggled with, particularly those involving people.
…“Table tennis is a game of enormous complexity that requires split-second decisions as well as speed and power,” said Peter Dürr, director of Sony AI in Zurich, who led the Ace project. “This research breakthrough highlights the potential of physical AI agents to perform real-time interactive tasks, and represents a significant step toward creating robots with broader applications in fast, precise and real-time human interactions.”
Ace beat three out of five elite table tennis players who had more than ten years of training, and scored 48 points versus 70 in two defeats to professionals, according to a paper published in Nature on Wednesday.
The robot had improved further, Sony said: since the paper’s submission, it had played four further matches against humans, beating two elite players and winning one out of two matches against professionals.
The robot handled spin and unexpected changes in trajectory caused by the ball clipping the net on its way over, the researchers said. It outscored the elite players in aces — points won directly from serve — by 16 to eight.
«
Reminiscent of DeepMind’s AlphaGo beating the European champion Go professional Fan Hui in October 2015. Then it beat the human world champion. But now, humans use AI-based Go computers to inform their play, and still play other humans. Maybe this will become the sparring champion everyone wants to play against?
unique link to this extract
Diet Coke shortage in India: why your favourite zero-calorie Cola has vanished from Blinkit, Zepto and store shelves • Open Magazine
Shailja Tiwari:
»
Somewhere in India right now, a person is refreshing [shopping apps] Blinkit or Zepto for the seventeenth time today. The result is the same: out of stock. It has been that way since mid-April. The silver can — the diet-cola lifeline of India’s urban millennial and Gen Z population — has vanished from shelves in Bangalore, Mumbai, and Ahmedabad, sending its loyal fanbase into collective mourning.
Quick-commerce apps showed it unavailable. Supermarkets said it was gone. Convenience stores shrugged. One Reddit user from Bangalore reported walking into a bar — “underage, by their own admission” — and the bar was out too. Their final, haunted question: “Like is the Diet Coke ka ship stuck in Strait of Hormuz or what?”
With shelf space empty and hot takes incoming, the internet convened its own inquiry. Several theories emerged with great confidence and varying degrees of accuracy.
Theory one: Instagram creators hoarding empty cans for DIY craft projects disrupted the aluminium recycling cycle. Creative. Chaotic. Completely unverifiable. Acquitted.
Theory two: a ship stuck in the Strait of Hormuz. Beautiful poetry. Also acquitted.
The real culprit is considerably more boring than either: aluminium. According to metal market analysts, aluminium supply came under serious pressure heading into April 2026, with prices rising roughly 14–20% in recent weeks due to geopolitical disruption, production outages, and reduced imports. Domestic mills are stretched thin. Lead times are longer. The cans are simply hard to source.
…In fairness, this is a localised supply hiccup, not a national catastrophe. But context matters: Diet Coke sales in India actually doubled last year, as low-sugar beverages surged to represent 30% of Coca-Cola India’s total volumes in 2025 — up from just 5% in 2020. The fanbase is real, large, and very online.
«
If aluminium prices are going up, there’s nothing in the short term that’s going to bring them down. In fact we have some bad news coming up for the fanbase…
unique link to this extract
Aluminium faces ‘black swan’ supply shock, Mercuria says • Reuters
Pratima Desai:
»
The global aluminium market is experiencing a “black swan” event as disruptions due to the Middle East war trigger a supply shock that will lead to major shortages this year, according to the top metals analyst at commodity trader Mercuria.
The region accounts for about 7 million metric tons of annual aluminium smelting capacity, or roughly 9% of the estimated global supply this year. Aluminium is a key material for the transport, construction and packaging industries.
“The scale of the supply shock we’re seeing in the aluminium market is probably the largest single supply shock a base metals market has suffered in the post-2000 era,” Nick Snowdon, head of metals and mining research at Mercuria, said on the sidelines of the Financial Times Commodities Global Summit in Lausanne, Switzerland.
“We are already in a ‘black swan’ event. No one could have foreseen something on this scale,” he told Reuters.
Concerns about supplies due to disruptions stemming from the U.S.-Israeli war with Iran fuelled a rally on the London Metal Exchange, pushing aluminium prices to a four-year high at $3,672 a ton on April 16.
Mercuria estimates the market will face, at a minimum, a deficit of roughly 2 million tons between now and the end of the year. Snowdon said this estimate may prove conservative, as it assumes a near-term improvement in alumina flows via the Strait of Hormuz will enable some smelters to restart production this quarter.
“That shortfall compares with about 1.5 million tons of visible inventory and just over 3 million tons of total global stock, including non-visible units, leaving the market with limited buffers,” Snowdon said.
A larger deficit is possible if the conflict is extended and flows of alumina – a feedstock for aluminium production – to the Gulf are limited, he added.
Middle East aluminium cannot easily be replaced. In China, the world’s top producer, there is an annual output limit of 45 million tons, while the US and Europe have little idled capacity that could return.
«
The only good news from the Iran war • The New York Times
David Wallace-Wells:
»
The war has plunged the energy world into a kind of time warp: The last boats to exit the Strait of Hormuz before the war are just now approaching their refineries, and seven weeks is not a terrifically long time in which to recalibrate national energy systems.
But between February and March, the world didn’t just tighten its fossil fuel use; it spent money on new green stuff. China’s solar, battery and EV exports grew 39%. The country’s solar exports alone more than doubled during that time. In South Korea, March saw new EV registrations double compared with the same time last year. In New Zealand, new registrations nearly doubled over two weeks.
For the past five years, competition-minded trade economists have lamented Chinese “overcapacity” — that one country was producing more clean energy tech than the world really wanted, flooding the market with cheap products that prevented other countries from even competing. As the Carnegie Endowment’s Noah Gordon put it over the weekend, the response to the Iran war makes it look like this analysis had it more or less backward: the problem wasn’t Chinese overcapacity, it was constrained demand everywhere else in the world. As the economist Isabella Weber put it, “it turns out Chinese ‘overcapacity’ is the world’s largest buffer stock against a crisis like the one we face.”
Around the world, new clean energy policy — which had mostly stalled out or retreated over the past few years — is now following suit. In recent weeks, new clean energy investments and initiatives have been declared by France, Egypt, India, Indonesia, Turkey — and many others. The president of the E.U. Commission called for an expedited rollout of grid improvement and electrification plans. Jigar Shah, a former Biden administration official and a liberal energy guru, says he expects monthly global clean energy spending to double by the end of this year.
«
| • Why do social networks drive us a little mad? • Why does angry content seem to dominate what we see? • How much of a role do algorithms play in affecting what we see and do online? • What can we do about it? • Did Facebook have any inkling of what was coming in Myanmar in 2016? Read Social Warming, my latest book, and find answers – and more. |
Errata, corrigenda and ai no corrida: none notified
The Overspill: when there's more that I want to say 