Does playing chess consume as much energy as running a marathon? No. So why do some think it does? CC-licensed photo by Ben Schumin 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.

---

There’s another post coming this week at the Social Warming Substack on Friday at 0845 UK time. Free signup.

---

A selection of 10 links for you. Forked. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

---

Sam Altman slams Meta’s AI talent-poaching spree: ‘missionaries will beat mercenaries’ • WIRED

Zoë Schiffer:

»

OpenAI’s chief research officer, Mark Chen, told staff that it felt like “someone has broken into our home and stolen something.”

Altman struck a different tone about the departures in his note on Monday.

“Meta has gotten a few great people for sure, but on the whole, it is hard to overstate how much they didn’t get their top people and had to go quite far down their list; they have been trying to recruit people for a super long time, and I’ve lost track of how many people from here they’ve tried to get to be their Chief Scientist,” he wrote. “I am proud of how mission-oriented our industry is as a whole; of course there will always be some mercenaries.”

He added that “Missionaries will beat mercenaries” and noted that OpenAI is assessing compensation for the entire research organization. “I believe there is much, much more upside to OpenAl stock than Meta stock,” he wrote. “But I think it’s important that huge upside comes after huge success; what Meta is doing will, in my opinion, lead to very deep cultural problems. We will have more to share about this soon but it’s very important to me we do it fairly and not just for people who Meta happened to target.”

Altman then made his pitch for people to remain at OpenAI. “I have never been more confident in our research roadmap,” he wrote. “We are making an unprecedented bet on compute, but I love that we are doing it and I’m confident we will make good use of it. Most importantly of all, I think we have the most special team and culture in the world. We have work to do to improve our culture for sure; we have been through insane hypergrowth. But we have the core right in a way that I don’t think anyone else quite does, and I’m confident we can fix the problems.”

“And maybe more importantly than that, we actually care about building AGI in a good way,” he added. “Other companies care more about this as an instrumental goal to some other mission. But this is our top thing, and always will be. Long after Meta has moved on to their next flavor of the week, or defending their social moat, we will be here, day after day, year after year, figuring out how to do what we do better than anyone else. A lot of other efforts will rise and fall too.”

«

Lots of noise about the huge golden hellos that Zuckerberg is offering, but if they’re going to add huge value to a company, then it makes sense, doesn’t it. Rather like hiring a huge sports star for a team because you know they’ll score. Not sure about the missionaries/mercenaries suggestion, though.
unique link to this extract

---

Wimbledon 2025: how has the introduction of electronic line-calling gone? • BBC Sport

Jess Anderson:

»

The absence of line judges at Wimbledon on day one of the Grand Slam has certainly been noticeable.

For the first time electronic line calling has been introduced at the All England Club with the well-dressed line judges replaced by AI.

In the absence of the 300 line judges that have been used for the past 148 years, up to 18 cameras, developed by HawkEye, are situated around each court to track the progress of the ball and determine whether it is in or out.

The technology is already in operation at the US and Australian Opens but its introduction at SW19 has been a topic of discussion. The emptiness of the courts is noticeable on Centre Court and Court One, where the vast space behind the baseline is now occupied only by the ball kids.

Britain’s Cameron Norrie said it “looks cool” with the line umpires in place and contributes to the “tradition” of the tournament. “Obviously there’s a lot of jobs and people that love tennis, which will definitely be missed from them,” he said. But while many players agree line judges are part of the spectacle, few can argue with the accuracy of the calls. “As a player it’s pretty black or white with the calls,” added Norrie. “In, out… there’s no mistake, nothing happening. Definitely you’ve got to feel for those linesmen and those people. That’s a bit tough for them, but it’s pretty black or white with the calling.”

The theatre of players challenging the calls has also been a notable absence with fans unable to get involved with the drama of a close call being replayed on the big screen. American 12th seed Frances Tiafoe said he would have liked to see Wimbledon keep line judges. “I actually like [it] with them [line judges] on the court, because I think for fanfare it’s better,” he said. “If I were to hit a serve on a big point, you go up with the challenge, is it in, is it out? The crowd is, like, ‘ohhh’. There’s none of that. Now if I hit a good serve and they call it out, you may still think it’s in, but it doesn’t matter. I think that kind of kills it.”

«

There aren’t any really good explanations of how it works, but this one is decent. And, of course, Wikipedia.
unique link to this extract

---

No, chess grandmasters do not burn 6000 calories per day • A Note On The Production of Facts

Adam Strandberg:

»

I work on metabolism and have some interest in neurons, so I have on several occasions run into the claim that chess grandmasters burn 6000 calories per day during tournaments. I found this implausible and decided to investigate where it came from. While I am not the first person on the internet to express skepticism of such a large number, nobody seems to have worked out the precise source of the claim. I assumed when I dug into it that I would find a specific methodological error. But while methods enter the story, the real problem is that the number was completely made up.

As far as I can tell, the “patient zero” that caused this claim to become so widespread is this 2019 ESPN article:

»

Robert Sapolsky, who studies stress in primates at Stanford University, says a chess player can burn up to 6,000 calories a day while playing in a tournament, three times what an average person consumes in a day. Based on breathing rates (which triple during competition), blood pressure (which elevates) and muscle contractions before, during and after major tournaments, Sapolsky suggests that grandmasters’ stress responses to chess are on par with what elite athletes experience.

«

This story was then picked up by many outlets…

«

The game of academic telephone that was played to produce this crazy number (which is comparable with running more than two marathons) is quite entertaining, if it hadn’t produced this zombie statistic. What it suggests is: nobody checks anything for sense.
unique link to this extract

---

Scientists identify culprit behind biggest-ever US honey bee die-off • Science

Joanna Thompson:

»

U.S. beekeepers had a disastrous winter. Between June 2024 and January 2025, a full 62% of commercial honey bee colonies in the United States died, according to an extensive survey. It was the largest die-off on record, coming on the heels of a 55% die-off the previous winter.

As soon as scientists at the U.S. Department of Agriculture (USDA) caught wind of the record-breaking die-offs, they sprang into action—but their efforts were slowed by a series of federal funding cuts and layoffs by President Donald Trump’s administration. Now, six months later, USDA scientists have finally identified a culprit.

According to a preprint posted to the bioRxiv server this month, nearly all the dead colonies tested positive for bee viruses spread by parasitic mites. Alarmingly, every single one of the mites the researchers screened was resistant to amitraz, the only viable mite-specific pesticide—or miticide—of its kind left in humans’ arsenal.

Tracking the rise of miticide resistance is critical, experts say. Honey bees pollinate more than 90 commercial crops in the United States, generate between $20bn and $30bn in agricultural revenue, and play a key role in keeping the US food supply stable.

“There is a lot at stake,” says Danielle Downey, the executive director of Project Apis m., the nonprofit that conducted the bee die-off survey. USDA did not provide comment on its research to Science after multiple inquiries spanning nearly three weeks, with one spokesperson citing a need “to move [the request] through agency clearance.”

Miticide-resistant varroa mites have been a growing issue for beekeepers for years, so much so that breeders have sought to develop mite-resistant bee varieties. Since the 1980s, the parasites have evolved global resistance to at least four major classes of miticide. Unfortunately, effective new compounds are notoriously difficult to develop, and amitraz represented one of the best remaining treatments. But the preprint suggests amitraz could soon fall by the wayside.

«

It’s always bad news with the bees, unfortunately.
unique link to this extract

---

Trump’s tariff threat pushes Canada to scrap digital services tax • Ars Technica

Ashley Belanger:

»

In a sudden reversal, Canada has caved and will remove its digital services tax after trade talks with the US suddenly fell apart this weekend.

Blocked just hours before taking effect, the controversial digital services tax (DST) would have charged big US tech companies like Apple, Google, and Meta a 3% tax on all digital services revenue earned from Canadian users. Frustrating US tech giants, Canada also sought to collect retroactive taxes dating back to 2022.

Over the weekend, President Donald Trump claimed the tax was a “direct and blatant attack” on US tech companies and terminated the trade talks, while threatening to impose a new tariff rate on Canadian goods by July 4.

On Sunday, Canadian Prime Minister Mark Carney seemingly bowed to Trump’s pressure campaign, abruptly doing an “about turn” after previously refusing to pause the DST despite Trump’s opposition, NBC News reported.

But it wasn’t just Trump pushing Carney to reconsider the tax. A nonprofit representing CEOs and leaders of some of Canada’s biggest businesses, the Business Council of Canada, had warned that Carney defending the tax risked “undermining Canada’s economic relationship with its most important trading partner,” Al Jazeera reported.

If Trump were to impose new tariffs on Canada, it could have “large ripple effects across both economies,” the Council warned, potentially disrupting markets for automobiles, minerals, energy, and aluminum.

«

Every effort to impose a “digital tax” seems to run into the ground in North America. The picture in Europe seems confused – the UK and other European OECD countries have laws for them but the road to implementation is long.
unique link to this extract

---

Using AI to humiliate women: the men behind deepfake pornography • Der Spiegel

Marvin Milatz and Max Hoppenstedt:

»

Nudify apps are not hidden in obscure forums or on pornography platforms, rather they are freely available on the internet. The only limitation: Many of these services only work with women’s bodies. The AI programs they use have apparently never been trained to produce naked pictures of men. Images of women in underwear are usually free, with faked photos of subjects in typical pornographic poses available for a price of just a few euros.

Clothoff is one of the leading apps on the market. In just the first six months of 2024, the website received 27 million visitors, with an average of 200,000 pictures being produced by the program each day, according to the company. Thousands of women have likely become victims of the app. The creators of Clothoff are among the most unscrupulous nudify operators and offer photo montages with schoolgirl outfits and pregnant women in sex poses. The app has recently begun marketing the ability to create fake videos with a picture. According to company information, the function has already been used over a million times.

In August 2024, public officials in San Francisco, California went public with a lawsuit against Clothoff and several other nudify apps. They demanded that the services cease operation due to the distribution of child and youth pornography. The investigators from the heart of Silicon Valley were likely also motivated by the fact that cases had become public at several schools in the state in which AI-generated nude images of girls had been circulated. Thus far, however, officials have experienced only moderate success in identifying the people behind the apps or getting them to suspend their services. The operators of Clothoff, in any case, seem unimpressed by the lawsuit.

Just how lucrative Clothoff’s business has become can now be seen by the statements of a whistleblower who has access to internal company information. Working for the app, he says, initially felt like being part of an exciting startup. “But over time, people became just cynical and obsessed with money.” For his own safety and out of concern that the company may seek to retaliate, his identity cannot be revealed. He, too, initially joined the project in the hopes that it would turn out to be profitable, but he now feels partly responsible for what Clothoff has become. Which is why he is going public with his information.

«

unique link to this extract

---

How plug-in solar panels could help Britons save 30% on energy bills • The Independent

Howard Mustoe:

»

Britons could soon buy cheap solar panels which can be put on balconies, sheds and terraces to cut their energy bills.

The plan, using so-called plug-in solar modules which typically cut bills by 30% in countries where they are already used, has been announced by the government as part of a broader plan to step up solar power access.

Using balconies for solar panels is already common in Germany, where 1.5 million homes use the technology. Locals have named it balkonkraftwerk, or balcony power plant.

Using solar panels this way is a lot cheaper than installing them on a roof, where scaffolding and hiring specialist workers means that even a modest eight-panel array will cost about £5,000.

It will also unlock solar power for many of Britain’s 5.4 million households which rent. Presently, only homeowners can opt to fit solar panels unless they agree a deal with their landlord. Even then, if they moved, they would lose their panels. The proposed system is portable.

This so-called plug-in power plant also means no expensive fitting is required. Instead, the panels are attached to an inverter, which steps up the voltage to the 240V used by your home’s mains supply. It is then attached by a regular plug. To avoid electric shock, the inverters detect when they are unplugged and isolate the plug and its exposed electrical pins.

The government is investigating whether a similar safe system can be deployed in the UK.

Unlike a full rooftop solar system, no power can be sold back to the grid with plug-in panels. Instead, the aim is to cut electricity bills during the day from appliances like fridges, freezers and computers used by home workers.

«

At last something useful. As solar prices come down, small and convenient systems like this become more useful to deploy just anywhere.
unique link to this extract

---

Scaling human judgment in Community Notes with LLMs • Arxiv

Haiwen Li and others from MIT and X.com:

»

The core tasks of a Community Notes contributor—researching a claim, synthesizing diverse sources, and drafting a neutral, well-evidenced summary—are capabilities at which “Deep Research” LLMs show promise. Indeed, work from 2024 has demonstrated that a fully automated pipeline for generating Community Notes with LLMs can, under certain circumstances, produce notes that are of similar quality to human-written notes—at a fraction of the time and effort. LLM-written CNs have the potential to be faster to produce, less effort to generate, and of high quality, hence are an attractive direction to pursue.

While LLM-written CNs are compelling, critical questions remain: can LLMs consistently produce accurate notes that are well-received across perspectives? If yes, can we use them to accelerate the addition of informative context in a way that is valued and trusted across viewpoints, while avoiding the fate of becoming just a new version of top-down fact-checking that can lack broad trust? We believe both are possible.

Rather than replacing humans, LLMs can complement and enhance their work. Trust in CN stems not from who drafts the notes, but from who evaluates them. It is the collective judgment of a diverse and engaged community that grounds people’s trust in CN.

This paper argues for a new paradigm for Community Notes in the LLM era: an open ecosystem where notes from both human writers and LLMs are submitted into a single pool, and the decision of which notes are helpful enough to show remains in the hands of the people. The system’s legitimacy is still upheld by its foundational principle: a community of diverse human raters that collectively serve as the ultimate evaluator and arbiter of what is helpful.

«

Getting an LLM to help write the advisory notes sounds good, though the real problem with Community Notes remains how long it takes for them to be “approved”. There’s no indication of how many approvals are needed for a Note to appear, but it seems to be a long time. And meanwhile the false claims circulate unchecked.
unique link to this extract

---

Google kills the fact-checking snippet in search results • Nieman Journalism Lab

Clara Jiménez Cruz:

»

Hidden in a developer blog earlier this month, Google announced that it will stop using the fact-checking snippets in search.

For a decade, the fact-checking snippet, under the name of ClaimReview, has been a way to protect users worldwide by showcasing fact-checked information when a Google user searched explicitly for an already debunked claim. It exposed citizens to reliable information first by enhancing search results for fact-check articles.

ClaimReview enabled Google to expose fact-checked content to over 120 million European Union citizens in the first half of 2024 (see data by country).

I am the founder and CEO of fact-checking foundation Maldita.es and chair of the European Fact-Checking Standards Network. Google did not inform fact-checkers that the 10-year collaboration was coming to an end, let alone consult with us on the decision to stop using the fact-checks that we provided for free.

The company says its data shows the fact-checking snippet is “not commonly used in Search” and no longer provides “significant additional value for users,” but did not share the data or analysis that led it to make this decision.

This year’s Reuters Institute Digital News Report asked 92,000 online news consumers in 46 markets which sources they look for in search results when they’re checking information they suspect may be fake. The Reuters Institute’s independent data does not match Google’s analysis: 25% of news consumers globally said they’re looking for a fact-check when they conduct a search, including 38% in the U.S.

«

In the blogpost, Google talks about “simplification”, but it’s hard not to think that the fact checking could contradict the AI chunks at the top of the search results, which would never do. (Thanks Gregory B for the link.)
unique link to this extract

---

Three-dimensional time: a mathematical framework for fundamental physics • Reports in Advances of Physical Sciences: Vol 09

Gunther Kletetschka:

»

This paper introduces a theoretical framework based on three-dimensional time, where the three temporal dimensions emerge from fundamental symmetry requirements. The necessity for exactly three temporal dimensions arises from observed quantum-classical-cosmological transitions that manifest at three distinct scales: Planck-scale quantum phenomena, interaction-scale processes, and cosmological evolution.

These temporal scales directly generate three particle generations through eigenvalue equations of the temporal metric, naturally explaining both the number of generations and their mass hierarchy. The framework introduces a metric structure with three temporal and three spatial dimensions, preserving causality and unitarity while extending standard quantum mechanics and field theory.

«

Just thought I’d bring this to your attention, in case you needed an explanation for being late: “the time shape was wrong for the train, unfortunately.”
unique link to this extract

---

• 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