Start Up No.2607: how AI changes work, China’s car are ready to go, Trump blocks EV chargers (again), Gemini grows, and more


The UK Ministry of Justice intends to delete a huge database of court records, for no obviously good reason. CC-licensed photo by Hc_07 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 10 links for you. Poor judgement. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.


AI doesn’t reduce work—it intensifies it • Simon Willison

Simon Willison:

»

AI Doesn’t Reduce Work—It Intensifies It (via) Aruna Ranganathan and Xingqi Maggie Ye from Berkeley Haas School of Business report initial findings in the HBR from their April to December 2025 study of 200 employees at a “U.S.-based technology company”.

This captures an effect I’ve been observing in my own work with LLMs: the productivity boost these things can provide is exhausting.

»

AI introduced a new rhythm in which workers managed several active threads at once: manually writing code while AI generated an alternative version, running multiple agents in parallel, or reviving long-deferred tasks because AI could “handle them” in the background. They did this, in part, because they felt they had a “partner” that could help them move through their workload.

While this sense of having a “partner” enabled a feeling of momentum, the reality was a continual switching of attention, frequent checking of AI outputs, and a growing number of open tasks. This created cognitive load and a sense of always juggling, even as the work felt productive.

«

I’m frequently finding myself with work on two or three projects running parallel. I can get so much done, but after just an hour or two my mental energy for the day feels almost entirely depleted.

I’ve had conversations with people recently who are losing sleep because they’re finding building yet another feature with “just one more prompt” irresistible.

«

I think this is especially true for programmers, where projects can be endless Forth Bridge repaintings (with new bits added on all the time). But there’s certainly something subtly addictive about the way that once the answer to your prompt has unfurled, the chatbot says “Is there anything else I can help with?”
unique link to this extract


I drove three Chinese cars; here’s why they would clean up in the US • The Verge

John Voelcker:

»

It only took a brief drive in a Zeekr 7X to convince me: Chinese cars are now competitive and could be sold in the US tomorrow. The compact battery-electric crossover, a bestseller in Europe, is aimed directly at the Tesla Model Y with its five seats, two rows, impressive road grip, energetic performance, and smooth ride. Its price in China is around $32,000 — about $7,000 cheaper than Elon Musk’s crossover.

If you’ve followed automotive news recently, you might come away thinking Chinese cars are destined for the US, with Geely among the best positioned to break through first. That’s certainly a possibility — especially after President Donald Trump seemingly threw open the doors to Chinese automakers in a recent speech. The quid pro quo: They must be built in the US. Geely recently confirmed it was “actively evaluating” a possible entry into the world’s second-largest auto market and would have a decision within three years.

Geely may be best placed to build cars here: Volvo Cars, which it controls, has had an auto plant up and running in South Carolina since 2018, where it builds Volvo and Polestar vehicles. (Volvo stated, “We do not have any plans to produce cars on behalf of Geely there.”) If Geely does attempt to enter the US market under its own brands, it likely won’t happen before 2029.

Still, climbing out of the Zeekr 7X, I wondered — even if it were built in the US — whether it could legally be sold here. Will restrictions on automotive software that originates in China keep it out indefinitely? Or will Geely find a way to convince the US government that its technology poses no threat to US national security?

Most US drivers won’t have heard of Geely, if they think about Chinese cars at all. Yet last year, parent company Geely Holding sold more than 4.1 million cars, making it China’s second-largest carmaker after BYD. It offers multiple brands that include Volvo, Polestar, Lotus, Malaysia’s Proton, Geely, Lynk & Co, Zeekr, and others.

«

Had not absorbed that Volvo is now a Chinese-owned brand. But also: of course Trump’s admin will block this.
unique link to this extract


Trump’s team stalls EV charging money again • The Verge

Andrew Hawkins:

»

Last month, a federal judge ordered the US Department of Transportation to unfreeze $5bn from the federal program dedicated to building more EV chargers. But today, Transportation Secretary announced a new requirement that all federally funded EV chargers be “100 percent” built in America. Since most EV chargers are sourced from China, this will essentially refreeze the funds and indefinitely delay the installation of more chargers.

«

It’s so petty, and vindictive, and retrograde.
unique link to this extract


Apple and Google pledge not to discriminate against third-party apps in UK deal • The Guardian

Dan Milmo:

»

Apple and Google have committed to avoid discriminating against apps that compete with their own products under an agreement with the UK’s competition watchdog, as they avoided legally binding measures for their mobile platforms.

The US tech companies have vowed to be more transparent about vetting third-party apps before letting them on their app stores and not discriminate against third-party apps in app search rankings.

They have also agreed not to use data from third-party apps unfairly, such as using information about app updates to tweak their own offerings. Apple has also committed to giving app developers an easier means of requesting use of its features such as the digital wallet, and live translation for AirPod users.

The commitments have been secured as part of a new regulatory regime overseen by the Competition and Markets Authority, (CMA), which has the power to impose changes on how Apple and Google operate their mobile platforms after deciding last year that they had “substantial, entrenched” market power. However, the CMA has opted to allow voluntary commitments rather than impose formal changes.

Tom Smith, a competition lawyer at Geradin Partners and a former CMA director who represents app developers challenging Apple and Google’s dominance of the mobile platform market, said the changes were “lightweight” and had no legal bite.

The CMA did not name specific Google and Apple apps that compete with third-party rivals but both firms offer their own music services in the form of YouTube Music and Apple Music. Google and Apple’s app stores are vital shopfronts for app developers because the Android and iOS platforms are used by the overwhelming majority of UK mobile phone users.

However, the commitments do not cover a significant bugbear for app developers – the fees of up to 30% that Apple and Google charge developers to sell products via their app stores. The CMA is still considering whether to tackle how Apple and Google charge developers.

«

Bear in mind that in October Apple in the UK lost a £1.5bn collective action case which found it had overcharged for apps and subscriptions, so you’d hope the CMA might look at the reasoning there for some guidance.
unique link to this extract


OpenAI’s lead is contracting as AI competition intensifies • Big Technology

Alex Kantrowitz:

»

OpenAI’s rivals are cutting into ChatGPT’s lead. The top chatbot’s market share fell from 69.1% to 45.3% between January 2025 and January 2026 among daily U.S. users of its mobile app. Gemini, in the same time period, rose from 14.7% to 25.1% and Grok rose from 1.6% to 15.2%.

The data, obtained by Big Technology from mobile insights firm Apptopia, indicates the chatbot race has tightened meaningfully over the past year with Google’s surge showing up in the numbers. Overall, the chatbot market increased 152% since last January, according to Apptopia, with ChatGPT exhibiting healthy download growth.

On desktop and mobile web, a similar pattern appears, according to analytics firm Similarweb. Visits to ChatGPT went from 3.8 billion to 5.7 billion between January 2025 and January 2026, a 50% increase, while visits to Gemini went from 267.7 million to 2 billion, a 647% increase. ChatGPT is still far and away the leader in visits, but it has company in the race now.

“ChatGPT showed really strong growth for most of 2025,” said David Carr, insights news and research editor at Similarweb. “That said, we did see a ChatGPT traffic dip in November / December, which coincided with a growth spurt for Gemini. Preliminary data for January show ChatGPT traffic recovering but not back to its peak of 6 billion + visits in October. And Gemini continues to grow strongly, up another 17% month over month based on a preliminary estimate for January.”

«

Gemini is really starting to move in there. Maybe at some point in the future OpenAI will make an antitrust complaint about Google using its search dominance to push its AI products. But the search antitrust trial seemed to allow them to do that.
unique link to this extract


“Hate brings views”: confessions of a London fake news TikToker • London Centric

Katherine Denkinson and Jim Waterson:

»

The man on the recording is baffled. He can’t understand how London Centric traced his anonymous hate-filled London TikTok account back to his employer by geolocating the wheelie bins in his videos.

“I thought no one’s gonna notice that,” he says. “Why would someone?”

Last summer, the man says, he found himself sitting in his car, analysing trends on TikTok. His day job was conducting viewings for an estate agency but he was trying to come up with an idea for a viral video account that could be run as a money-making side-hustle.

“I was thinking of unique videos I can do for people,” he says on the tape.

That’s when he had a brainwave: “Hate brings views.”

At that time protests outside asylum hotels were spreading across the country. The man says he noticed “far-right people” were among the most engaged on TikTok. They were easy to rile up: “They hate such videos of illegal migrants. I was like, why not?”

The result was the account Reform_UK_2025, which co-opted the logo and name of Nigel Farage’s political movement without permission from the party. It posted video tours of Londoners’ homes accompanied by an AI-generated voice claiming properties in Knightsbridge and Chelsea had been handed over to illegal immigrants for free. It smeared residents, who were visible in some of the videos, as rapists and said that others proclaimed their hatred of the UK while collecting the keys.

It was an instant hit, attracting millions of views. It was also, the man confesses, all lies.

«

London Centric previously wrote about TikTokers secretly filming fake anti-immigrant videos inside Londoners’ homes. It’s a fabulous publication, but geolocating wheelie bins might be the absolute best yet.

But yes, this is what Social Warming was pointing to. Though at the time the idea of making tons of money out of generating hate – rather than just finding people who share your hate – wasn’t as developed as now.
unique link to this extract


Spain will ban social media for under-16s • CNN Business

Hanna Ziady:

»

Spain will ban social media for under-16s and require platforms to employ strict age verification tools, joining Australia, France and Denmark in moves to curb the influence of digital platforms on children.

“Our children are exposed to a space they were never meant to navigate alone. A space of addiction, abuse, pornography, manipulation and violence,” Spanish Prime Minister Pedro Sánchez said Tuesday. “We will no longer accept that, we will protect them from the digital Wild West.”

Speaking at the World Governments Summit in Dubai, Sánchez said his government would also introduce new laws to hold social media executives criminally liable for failing to remove illegal or hateful content.

The new rules will also sanction people and platforms that amplify illegal content, including via algorithms. “We will turn algorithmic manipulation and amplification of illegal content into a new criminal offense,” he said. “Spreading hate must come at a cost.”

The process of passing legislation will begin next week. Other proposed measures include developing a “hate and polarization footprint,” Sánchez said, a system to track and quantify how digital platforms fuel division and amplify hate.

In December, Australia became the first country in the world to implement a social media ban for under-16s, barring access to 10 platforms, including Facebook, TikTok, Instagram, Snapchat and X. Britain is considering a similar move, while France and Denmark have recently announced plans to stop under-15s accessing social media.

«

This is a very interesting development. Give it long enough, and if the bans are effective enough (even if only at schools) then we should, within five years or so, have enough data for a longitudinal study.
unique link to this extract


ICE activity is pushing readers to nonprofit news sites that cover immigrant communities • Nieman Journalism Lab

Joshua Benton:

»

In our recent rankings of web traffic at the top local newspapers and public media outlets, there’s been a consistent trend: Wherever ICE unleashes its controversial deportation tactics, the audience’s attention follows. Operation Midway Blitz sent readers to Chicago’s WBEZ, Operation Charlotte’s Web did the same for Charlotte’s WFAE, and the chaos in Minneapolis led Minnesota Public Radio to the top of the traffic rankings.

For another class of digital publishers, nonprofit outlets, the effect can be even more intense, since some focus specifically on immigrant communities being targeted. Our latest nonprofit rankings — for October, November, and December 2025 — show that clearly.

— In October, the biggest percentage increase in traffic any site saw was at NepYork, a site that covers the Nepali community in New York City, where visits skyrocketed from around 20,000 to 289,000. (ICE deportations of Nepali citizens have been on the rise.) Other big gainers included The Maine Monitor (up 153%), California’s Fresnoland (up 152%), Fort Worth Report (up 51%), and The TRiiBE, a site about Black Chicago (up 107%).

«

(Thanks Gregory B for the link.)
unique link to this extract


The big split driving the tricky politics of AI data centres • POLITICO

Brendan Bordelon and Gabby Miller:

»

The tech industry is facing fierce local backlash to data centre projects around the country. But a new poll suggests national opinion is still up for grabs.

Cities from Madison, Wisconsin, to Chandler, Arizona, are rejecting new data centres — the hulking, server-packed complexes that make up the backbone of the booming artificial intelligence industry — citing everything from rising electricity costs to depleted water tables and air pollution.

Nationally, however, the tech giants behind the rapid rollout of data centres have a window to shape public opinion despite opposition they’re seeing on the local level, according to new results from The POLITICO Poll. The survey, conducted by London-based independent polling company Public First, found that most voters are blasé — even mildly positive — about the possibility of having a data centre in their area, associating them with new jobs and other economic benefits.

But the industry’s standing is also precarious, and the poll suggests a partisan split is emerging: People increasingly see the tech companies as aligned with Republicans, the survey shows. And Democratic Govs. Abigail Spanberger of Virginia and Mikie Sherrill of New Jersey won races in November in part by campaigning to force data centre operators to help upgrade the electric grid and keep utility rates down.

As the AI-driven projects spread, the map the industry will have to defend politically and financially will keep expanding, far beyond traditional clusters of data centre growth in places like Virginia and Texas.

“I think it’s going to be a big issue in the midterm elections,” said Brad Carson, a former Oklahoma Democratic representative and head of Public First, a super PAC pushing for AI regulations that has no relation to POLITICO’s polling partner. He said that while most people can tune out data centres as an abstraction, “we know there are discrete pockets of people — often quite conservative in their politics — who care a lot about it.”

Carson said people who are against data centres “are likely to vote on that issue, right, because, ‘I don’t want a data centre in my neighborhood, I’m opposed to it.’ The guys who are for [data centres] are like me — they’re a million miles away from the nearest data centre.”

«

(Thanks Joe S for the link.)
unique link to this extract


Ministry of Justice orders deletion of UK’s largest court reporting archive • The Times

George Greenwood:

»

The Ministry of Justice is ordering the deletion of a large archive of court records, raising open justice concerns.

Courtsdesk, a data analysis company that supports media and campaigners in monitoring court records, has been ordered by the government to delete its archive, which provides a crucial tool for journalists covering the justice system.

The project was approved by the lord chancellor in 2021 to explore how a “national digital news feed of listings and registers can improve coverage of the courts by the news media” by opening up magistrate court records.

According to Courtsdesk, the platform has since been used by more than 1,500 journalists from 39 media organisations and the data provided has highlighted serious failures in the courts system.

It said journalists were given no advance notice of 1.6m criminal hearings, the number of court cases listed was accurate on just 4.2% of sitting days and half a million weekend cases were heard with no notification to the press.

Two-thirds of all courts routinely heard cases that the media was not told about in advance. Seventeen courts that sent outcome records had not once published an advance listing in the entire period, the company’s research found.

In November, HM Courts and Tribunal Service (HMCTS) issued the company a cessation notice, citing what it called “unauthorised sharing” of court data, on the basis of a test feature, claiming this was a “data protection issue.”

«

Governments are always in favour of open data, right up to the point where it starts to embarrass them.
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

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.