Saturday, 17 October 2026
Saturday, 17 October 2026
- In brief
- 1Britain's competition watchdog moves to give Microsoft a special status that would let it set rules on how Windows, Office and Teams are sold
- 2China wants licences to export its biggest AI
- 3Taiwan to make AI chip sales to China a crime
- 4Anthropic drafts its own AI safety law
- 5Open AI models now carry most of one gateway's work
- 6New York council questions SpaceX over Grok
- 7Poland summons platforms over school threats
- 8Waymo's robotaxis head to Toronto
- 9London drivers push back on robotaxis
- 10Nvidia buys stakes in five UK start-ups
- 11iPhone Duo pre-orders sell out on day one
- 12China's CXMT makes nearly $20 billion
- 13Taiwan raises $4 billion for a Texas supplier park
- 14A Chinese start-up shows a key chip-machine part
- 15Thai police cut a cable feeding a scam compound
- 16CERN and Mistral plan an AI model for physics
- 17Banks and shops block outside AI agents
- 18Replit says it dropped Anthropic
- 19AI reviewers delay the next PostgreSQL
- 20487 researchers want hacking AIs kept offline
- 21MI5 sets up a centre to judge AI threats
- 22A conservative group sets out its AI plan
- 23Anthropic's staff expect AI to do their jobs
- 24A tracker counts AI-linked job cuts
- 25Irish tech workers join a union
- 26Microsoft's president hits back on green cards
- 27A UK boss gives AI only to his best staff
- 28OpenAI gives $5 million after a maths row
- 29Microsoft brings Defender to Linux desktops
- 30Munich and Berlin move to free software
- 31Capita set to miss a pension deadline
- 32Gmail answers spoken questions for paying users
- 33AI chats nobody else can read
- 34MapQuest shows where cameras watch
- 35An AI designs a homemade gun
- 36Police use an AI-made child to catch a suspect
- 37Fake experts trouble a research firm
- 38A tool checks lawyers' AI citations
- 39Elorian's first model looks twice
- TruthThe truth, across the day
17 Oct · Tech · In brief
Tech · Saturday, 17 October 2026
Tech, in brief
39 stories · 55 sources
Swipe up for the day in brief ↑Scroll down for the day in brief ↓
What is moving · 1 of 4
Governments guarding AI know-how
China proposed that exports of its largest AI models and AI chip designs need a licence.
What is moving · 2 of 4
Buyers switching AI suppliers
Almost one in three teams on Vercel's AI service changed their main AI supplier in September.
What is moving · 3 of 4
Officials questioning AI firms
New York City's council questioned SpaceX about sexual images its chatbot Grok made of real women.
What is moving · 4 of 4
Workers worried about AI jobs
London's ride-hail drivers plan a protest at Wayve, a robotaxi company, this weekend.
Part of the day's signals
2 pressures seen across the day run through Tech too.
39 stories
- 1Britain's competition watchdog moves to give Microsoft a special status that would let it set rules on how Windows, Office and Teams are sold
- 2China wants licences to export its biggest AI
- 3Taiwan to make AI chip sales to China a crime
- 4Anthropic drafts its own AI safety law
- 5Open AI models now carry most of one gateway's work
12 sources
Tech, in brief
- 5China proposes export controls on its advanced AI models and chip designs
- 8Taiwan to Curb AI Chip Exports to China as Part of US Trade Deal
- 12China startup demoes prototype EUV light source - 15W of output power and 90% domestic components, still years away from viability
- 15Open-weight models win hearts, minds, and wallets
- 18Replit is no longer working with Anthropic, CEO says
- 19OpenRouter challenger closes seed round
- 21Anthropic quietly pitches its own AI safety bill to lawmakers, Politico reports
- 26SpaceX AI policy leader grilled by NYC Council on Grok safety, sexualized images
- 29'We should all be worried': calls grow to limit London rollout of robotaxis
- 42Poland tells tech platforms to curb violent content after school attacks
- 44How many jobs is your state losing to AI? This dashboard tracks the cuts - and the cost
- 48In Europe's tech capital, Irish labor unions are booming amid AI-driven layoffs
Britain's competition watchdog moves to give Microsoft a special status that would let it set rules on how Windows, Office and Teams are sold
Swipe up for the story ↑Scroll down for the story ↓
Britain's Competition and Markets Authority (CMA), the UK's competition watchdog, proposed on Friday 16 October to give Microsoft 'strategic market status' in business software.
PC systems, office programs, security, servers and databases
Who is involved
The Competition and Markets Authority (CMA)
Britain's competition watchdog; it proposed the special status for Microsoft
Who is involved
Sarah Cardell
the CMA's chief executive; she says businesses should be free to mix Microsoft's products with rivals' [1]
Why it matters
Watch whether the CMA's first rules, if it goes ahead in February, change what Microsoft charges to run Windows Server and SQL Server on rival clouds, which cloud companies have long complained about.
What's really going on
Britain's competition regulator says Microsoft's discount for buying Windows, Office, Teams and its security tools together makes a rival program cost a business twice, and it wants powers by February 2027 to change how Microsoft sells them.
Why it works on people
A discount feels like a gift to the buyer, so it is hard to see it as the thing that keeps a rival out.
Who gains
Google and other rival cloud companies
If the CMA limits how Microsoft licenses Windows Server and SQL Server on other clouds, running those programs on Google's or Amazon's servers could get cheaper.
Who pays
Microsoft
Special status could limit the bundles and licence terms that, the CMA says, make rivals cost a business twice.
Nobody knows yet
What rules the CMA would actually set for Microsoft.
The CMA has listed possible steps, on bundling, cloud licences and Copilot, but rules come only after a final decision due by 13 February 2027.
Good news
Amazon Web Services and Microsoft now let customers move their data to another cloud company without a fee, and both have linked their data centres to Google's.
The short version
That's the story in short.
The whole story has 14 more parts.
or swipe up for the next storyor scroll on for the next story ↓
The whole story, in 14 parts ↓
Part 1 of 14 · Who else is involved
Microsoft
by far the biggest seller of business software in Britain for decades; it rejects the CMA's view [1][2]
Part 1 of 14 · Who else is involved
Amazon Web Services (AWS)
Amazon's cloud computing arm; told with Microsoft to make switching clouds easier [3]
Part 2 of 14 · How it unfolded
Oct 2025
Google and Apple are designated over their phone platforms
Mar 2026
the CMA accepts promises from AWS and Microsoft on cloud switching
Part 2 of 14 · How it unfolded
Mar 2026
the CMA accepts promises from AWS and Microsoft on cloud switching
May 2026
the CMA opens its investigation into Microsoft's software
Part 3 of 14
That status would let the CMA step in and set rules that Microsoft must follow.
Part 3 of 14
In June it used the same power to set rules for Google's search engine.
Part 4 of 14
The proposal covers Windows, the Office programs such as Word, Excel, Outlook and PowerPoint, Teams, and Microsoft's security, server and database software.
Part 5 of 14
About 20 million people a month in Britain use Microsoft's business software, Reuters reports.
Part 6 of 14
The CMA says Microsoft holds a strong position in all five areas it studied, and that customers find it hard to switch away.
Part 7 of 14
Microsoft sells whole bundles of programs at a discount, the CMA found.
Part 7 of 14
So a business that picks a rival's program for one job would pay twice, because that job is already covered in its Microsoft bundle.
Part 8 of 14
The CMA says it would look first at how Microsoft licenses its software to rival cloud companies, and at whether businesses can freely pick AI helpers other than Copilot, Microsoft's own.
Part 9 of 14
Its list of possible rules includes stopping Microsoft's technical design from blocking other providers, limiting bundling that harms competition, and stopping Microsoft from using its software power in cloud computing.
Part 10 of 14
Microsoft said it does not share the CMA's view and that all its products, including Copilot, are open and work alongside rivals.
Part 11 of 14
The Open Cloud Coalition, a group of cloud companies that includes Google, called it a 'once-in-a-generation opportunity' to end Microsoft's licensing terms.
Part 12 of 14
If the CMA goes ahead, Microsoft would be the third company under Britain's new digital rules, after Google and Apple.
Part 13 of 14
The same day, the CMA said Amazon Web Services and Microsoft have made it easier to move between clouds, but must do more.
Part 13 of 14
Both have linked their data centres to rivals', starting with Google's, and dropped fees for customers who move their data to another provider.
Part 14 of 14
Anyone can comment until 13 November, and the CMA will make its final decision by 13 February 2027.
Part 14 of 14
The US Federal Trade Commission and the European Commission are weighing similar steps, The Register reports.
4 sources
Britain's competition watchdog moves to give Microsoft a special status that would let it set rules on how Windows, Office and Teams are sold
The full story
Britain's competition watchdog moves to give Microsoft a special status that would let it set rules on how Windows, Office and Teams are sold
Britain's Competition and Markets Authority (CMA), the UK's competition watchdog, proposed on Friday 16 October to give Microsoft 'strategic market status' in business software.
That status would let the CMA step in and set rules that Microsoft must follow.
The proposal covers Windows, the Office programs such as Word, Excel, Outlook and PowerPoint, Teams, and Microsoft's security, server and database software.
About 20 million people a month in Britain use Microsoft's business software, Reuters reports.
The CMA says Microsoft holds a strong position in all five areas it studied, and that customers find it hard to switch away.
Microsoft sells whole bundles of programs at a discount, the CMA found.
The CMA says it would look first at how Microsoft licenses its software to rival cloud companies, and at whether businesses can freely pick AI helpers other than Copilot, Microsoft's own.
Its list of possible rules includes stopping Microsoft's technical design from blocking other providers, limiting bundling that harms competition, and stopping Microsoft from using its software power in cloud computing.
Microsoft said it does not share the CMA's view and that all its products, including Copilot, are open and work alongside rivals.
The Open Cloud Coalition, a group of cloud companies that includes Google, called it a 'once-in-a-generation opportunity' to end Microsoft's licensing terms.
If the CMA goes ahead, Microsoft would be the third company under Britain's new digital rules, after Google and Apple.
The same day, the CMA said Amazon Web Services and Microsoft have made it easier to move between clouds, but must do more.
Anyone can comment until 13 November, and the CMA will make its final decision by 13 February 2027.
Why it matters. Watch whether the CMA's first rules, if it goes ahead in February, change what Microsoft charges to run Windows Server and SQL Server on rival clouds, which cloud companies have long complained about.
China wants licences to export its biggest AI
Swipe up for the story ↑Scroll down for the story ↓
China's commerce and science ministries proposed on Friday that 23 kinds of AI technology need a licence to leave the country.
They include the inside workings of large models, AI agents, chip designs and big computer clusters.
The model rules would cover models with at least 2 trillion parameters, or ones ranked in the world's top ten, and Moonshot's Kimi K3 and Alibaba's Qwen3.8-Max are above that size.
Open-source code is exempt, and the public can comment until 15 November.
Why it matters
Firms that build on Chinese models now face a new question over whether those licences can be pulled, a Council on Foreign Relations fellow said.
Why it matters
The draft comes a month after the US and China agreed to cut tariffs and start talks on AI.
Who pays
US firms building on Chinese open models
China's draft lets it require licences for the biggest models, which could limit the deals these firms rely on, a Council on Foreign Relations fellow said.
Nobody knows yet
Which Chinese AI models the export licences will catch.
The draft names no model. Bloomberg and AP say Kimi K3 and Qwen3.8-Max are above the size line, and the final rules come after comments close on 15 November.
Part of a signal
China keeping its know-how at home
Seen in 3 of the 12 beats
Also seen in Biotech and Climate.
The whole signal →4 sources
China wants licences to export its biggest AI
The full story
China wants licences to export its biggest AI
China's commerce and science ministries proposed on Friday that 23 kinds of AI technology need a licence to leave the country.
Why it matters. Firms that build on Chinese models now face a new question over whether those licences can be pulled, a Council on Foreign Relations fellow said.
Taiwan to make AI chip sales to China a crime
Swipe up for the story ↑Scroll down for the story ↓
Taiwan will add AI servers and cutting-edge chips to its list of controlled exports, its economy minister Kung Ming-hsin told Bloomberg.
Once the rule takes effect, shipping that hardware to China will be a crime under Taiwan's Foreign Trade Act.
Taiwan is a centre of world chipmaking, and the rule brings it closer to US limits on sales to China.
Semafor reports the move is meant to win favour with the US government.
Why it matters
The rule aims at the black market that sends banned chips to Chinese buyers through middlemen.
Why it matters
Chinese chipmakers such as Huawei are meanwhile winning more of China's own AI chip market from Nvidia, Reuters reported.
Nobody knows yet
When Taiwan's rule on AI servers takes effect.
Taiwan's economy minister described the plan to Bloomberg but gave no start date in what we could read.
2 sources
Taiwan to make AI chip sales to China a crime
The full story
Taiwan to make AI chip sales to China a crime
Taiwan will add AI servers and cutting-edge chips to its list of controlled exports, its economy minister Kung Ming-hsin told Bloomberg.
Why it matters. The rule aims at the black market that sends banned chips to Chinese buyers through middlemen.
Anthropic drafts its own AI safety law
Swipe up for the story ↑Scroll down for the story ↓
Anthropic, the company behind the Claude chatbot, has quietly shown US lawmakers a draft law on AI safety, Politico reported on Friday.
It would let the US government stop a model only if it posed an imminent risk of more than $1 billion in damage, or of killing or badly injuring more than 50 people.
Outside testers registered with the US Commerce Department would check each lab's models, and the labs could pay them.
Only a federal judge in Washington, DC could order a model halted.
Why it matters
The draft would bar every other US agency from suspending a model, which leaves the US Commerce Department and one court as the only route.
Why it matters
Anthropic did not immediately respond to Reuters.
Nobody knows yet
Whether Anthropic's draft law has a sponsor in the US Congress.
Politico's report rests on three unnamed people, and neither Anthropic nor the US Commerce Department responded to Reuters.
1 source
Anthropic drafts its own AI safety law
The full story
Anthropic drafts its own AI safety law
Anthropic, the company behind the Claude chatbot, has quietly shown US lawmakers a draft law on AI safety, Politico reported on Friday.
Why it matters. The draft would bar every other US agency from suspending a model, which leaves the US Commerce Department and one court as the only route.
Open AI models now carry most of one gateway's work
Swipe up for the story ↑Scroll down for the story ↓
Open-weight models, which anyone can download and run, handled more than seven in ten tokens, the small chunks of text AI reads and writes, on Vercel's AI Gateway in September.
Vercel's gateway lets companies buy AI from many makers through one connection.
Open models' share of spending there rose from one dollar in twenty in June to nearly one in four.
A small new model called Jev reached 35% of paid teams within two weeks, the fastest Vercel has seen.
Why it matters
The figures cover only one gateway, and they miss companies that buy straight from OpenAI or Anthropic.
Why it matters
Chinese open models from DeepSeek and others are the strongest open ones, researchers told Rest of World.
Who gains
US companies that host Chinese AI models
More than two-thirds of September spending on Chinese models through Vercel's gateway went to US hosts, up from under a tenth in December.
3 sources
Open AI models now carry most of one gateway's work
The full story
Open AI models now carry most of one gateway's work
Open-weight models, which anyone can download and run, handled more than seven in ten tokens, the small chunks of text AI reads and writes, on Vercel's AI Gateway in September.
Why it matters. The figures cover only one gateway, and they miss companies that buy straight from OpenAI or Anthropic.
New York council questions SpaceX over Grok
Swipe up for the story ↑Scroll down for the story ↓
SpaceX's AI policy chief, Gabriela Castillo Madrid, answered New York City Council members by video on Friday.
SpaceX owns Grok, the chatbot on X, since it merged with Elon Musk's AI company in February.
The council speaker, Julie Menin, said she had found a sexual image of herself made with Grok that morning.
Castillo Madrid said the company has 'strict controls' and would look into it.
Why it matters
A former employee, Devin Kim, told the council only three staff worked on AI safety in his time there.
Why it matters
The council says it will use what it hears to write new rules.
1 source
New York council questions SpaceX over Grok
The full story
New York council questions SpaceX over Grok
SpaceX's AI policy chief, Gabriela Castillo Madrid, answered New York City Council members by video on Friday.
Why it matters. A former employee, Devin Kim, told the council only three staff worked on AI safety in his time there.
Poland summons platforms over school threats
Swipe up for the story ↑Scroll down for the story ↓
Poland summoned Google, Meta, TikTok, Discord and Telegram on Friday after knife attacks at schools last week.
A school employee was killed and about 15 people were hurt, and threats against schools spread online afterwards.
30 of 89
posts flagged, and the ones taken down
Polish officials flagged 89 posts as illegal between 7 and 12 October, and only 30 were taken down, the digital affairs minister Krzysztof Gawkowski said.
X was invited and did not attend.
Why it matters
A law that would let Poland order illegal posts removed at once, and fine platforms, now waits for President Karol Nawrocki's signature.
Why it matters
He vetoed an earlier version in January over free speech.
1 source
Poland summons platforms over school threats
The full story
Poland summons platforms over school threats
Poland summoned Google, Meta, TikTok, Discord and Telegram on Friday after knife attacks at schools last week.
30 of 89
posts flagged, and the ones taken down
Why it matters. A law that would let Poland order illegal posts removed at once, and fine platforms, now waits for President Karol Nawrocki's signature.
Waymo's robotaxis head to Toronto
Swipe up for the story ↑Scroll down for the story ↓
Waymo, the robotaxi company owned by Google's parent Alphabet, said on Friday that its cars will start driving Toronto's streets, with a person at the wheel making maps first.
It wants its first paying riders there in 2028.
Toronto will be its most wintry city yet.
Why it matters
Ontario's only rules for self-driving tests run out in October 2027 and allow no paid service, so the 2028 date depends on the province writing new ones.
Why it matters
Toronto's mayor has said she will back Waymo only if it does not cost jobs.
Nobody knows yet
Whether Waymo has the approval it needs in Ontario.
TechCrunch says Waymo was approved for Ontario's test programme. Electrek says Waymo's own post names no permit, and that the test rules expire in October 2027 with no rules yet for paid rides.
2 sources
Waymo's robotaxis head to Toronto
The full story
Waymo's robotaxis head to Toronto
Waymo, the robotaxi company owned by Google's parent Alphabet, said on Friday that its cars will start driving Toronto's streets, with a person at the wheel making maps first.
Why it matters. Ontario's only rules for self-driving tests run out in October 2027 and allow no paid service, so the 2028 date depends on the province writing new ones.
London drivers push back on robotaxis
Swipe up for the story ↑Scroll down for the story ↓
Ride-hail drivers plan to protest outside the London base of Wayve, a British self-driving company that supplies Uber, from Saturday.
A London Assembly committee report this week asked that Transport for London, the city's transport body, get full power to allow or stop robotaxis.
Today the UK government grants those permits, and Transport for London can only agree or refuse.
Why it matters
A think tank says more than 400,000 UK drivers, 121,000 of them in London, could find much of their work gone.
Why it matters
Uber says drivers and robotaxis will grow side by side for some time.
Who pays
London's ride-hail drivers
Robotaxis are built to run with no driver, and a think tank says much of the work of 121,000 London drivers could go.
1 source
London drivers push back on robotaxis
The full story
London drivers push back on robotaxis
Ride-hail drivers plan to protest outside the London base of Wayve, a British self-driving company that supplies Uber, from Saturday.
Why it matters. A think tank says more than 400,000 UK drivers, 121,000 of them in London, could find much of their work gone.
Nvidia buys stakes in five UK start-ups
Swipe up for the story ↑Scroll down for the story ↓
Nvidia, which makes the chips most AI runs on, said in London on Wednesday that it is taking stakes in five British start-ups.
They are the biotech firms Prima Mente and Boltz, Foresight, Hiverge and Riverlane, a quantum computing company.
Nvidia has joined 116 funding rounds this year, up from 12 in 2022, according to Dealroom, a data firm.
Why it matters
Nvidia has also put money into OpenAI, Anthropic, Intel and xAI.
1 source
Nvidia buys stakes in five UK start-ups
The full story
Nvidia buys stakes in five UK start-ups
Nvidia, which makes the chips most AI runs on, said in London on Wednesday that it is taking stakes in five British start-ups.
Why it matters. Nvidia has also put money into OpenAI, Anthropic, Intel and xAI.
iPhone Duo pre-orders sell out on day one
Swipe up for the story ↑Scroll down for the story ↓
Apple's first folding iPhone, the iPhone Duo, went on pre-order on Friday, and the phones set aside for launch-day pickup sold out quickly.
Some online delivery dates slipped to late January, and Apple called the demand 'nothing short of extraordinary'.
It reaches shops on 23 October, and T-Mobile is offering $1,000 off with a trade-in or a new line.
Why it matters
Folding phones are about 2% of phones sold this year but nearly 7% of the money spent, IDC, a market research firm, estimates.
Why it matters
Apple is betting that its biggest spenders will pay more, analysts told CNBC.
The full story
iPhone Duo pre-orders sell out on day one
Apple's first folding iPhone, the iPhone Duo, went on pre-order on Friday, and the phones set aside for launch-day pickup sold out quickly.
Why it matters. Folding phones are about 2% of phones sold this year but nearly 7% of the money spent, IDC, a market research firm, estimates.
China's CXMT makes nearly $20 billion
Swipe up for the story ↑Scroll down for the story ↓
CXMT, a Chinese maker of memory chips, expects a profit of 131 to 132.5 billion yuan, about $19.7 billion, for January to September.
A year earlier it lost 5.3 billion yuan in the same months.
AI computers need so much memory that chips for phones and PCs are short and prices have jumped.
Why it matters
Its chairman wants the company to buy back up to 2 billion yuan of its own shares.
Why it matters
Outside testers have not checked CXMT's claim that its newest memory is in mass production.
Who gains
CXMT, a Chinese memory chip maker
AI computers soak up memory, so chips are short and prices high, which turned a 5.3 billion yuan loss into a profit of about $19.7 billion.
Who pays
Buyers of phones and PCs
AI servers take so much memory that the chips phones and PCs use are short and dearer.
1 source
China's CXMT makes nearly $20 billion
The full story
China's CXMT makes nearly $20 billion
CXMT, a Chinese maker of memory chips, expects a profit of 131 to 132.5 billion yuan, about $19.7 billion, for January to September.
Why it matters. Its chairman wants the company to buy back up to 2 billion yuan of its own shares.
Taiwan raises $4 billion for a Texas supplier park
Swipe up for the story ↑Scroll down for the story ↓
Taiwan is raising at least $4 billion through an electronics makers' group to build an industrial park in Texas.
It is meant for the small firms that big chipmakers such as TSMC and Foxconn depend on.
Taiwan's government will also guarantee loans for those firms and help them with US visas and permits.
The plan may cover more than one site, at $2 billion to $4 billion.
Why it matters
TSMC has spent $165 billion on its Arizona site, but its smaller suppliers cannot afford to move alone.
1 source
Taiwan raises $4 billion for a Texas supplier park
The full story
Taiwan raises $4 billion for a Texas supplier park
Taiwan is raising at least $4 billion through an electronics makers' group to build an industrial park in Texas.
Why it matters. TSMC has spent $165 billion on its Arizona site, but its smaller suppliers cannot afford to move alone.
A Chinese start-up shows a key chip-machine part
Swipe up for the story ↑Scroll down for the story ↓
Langdao Silex, a Chinese start-up founded in 2024, showed a prototype light source for extreme ultraviolet (EUV) machines at a Shanghai fair.
EUV machines print the finest patterns on the most advanced chips, and China cannot buy them from ASML, the Dutch company that makes them.
Langdao says its source gives over 15 watts and that 90% of its parts are made in China.
Why it matters
ASML's production machines use sources of about 250 watts or more, so the gap is large.
Why it matters
Langdao has not said where it measured its 15 watts, which affects how it compares.
Nobody knows yet
How strong Langdao's chip-machine light source really is.
Langdao has not shown it working or said where its 15 watts were measured, so no outsider can compare it with ASML's.
1 source
A Chinese start-up shows a key chip-machine part
The full story
A Chinese start-up shows a key chip-machine part
Langdao Silex, a Chinese start-up founded in 2024, showed a prototype light source for extreme ultraviolet (EUV) machines at a Shanghai fair.
Why it matters. ASML's production machines use sources of about 250 watts or more, so the gap is large.
Thai police cut a cable feeding a scam compound
Swipe up for the story ↑Scroll down for the story ↓
Thai authorities found a 350-metre fibre-optic cable strung across the border from a village in Kanchanaburi province into Myanmar.
It carried about 6 terabytes of Thai internet data a day, enough for 300 to 400 people.
Officials say it kept a suspected scam compound in Myanmar online, along with an illegal casino.
A man was paid about $20 a day to let the cable run through his house.
Why it matters
The report calls Myanmar the centre of the world's online scam trade, and many people working in the compounds are trafficked there, groups that help them say.
Why it matters
Scam bosses, mainly Chinese, now run smaller operations in hotels and villas instead of big compounds, the report says.
Good news
Thai officials found and cut a 350-metre internet cable that a scam compound in Myanmar had strung across the border. It was carrying about 6 terabytes of Thai data a day to keep the compound online.
1 source
Thai police cut a cable feeding a scam compound
The full story
Thai police cut a cable feeding a scam compound
Thai authorities found a 350-metre fibre-optic cable strung across the border from a village in Kanchanaburi province into Myanmar.
Why it matters. The report calls Myanmar the centre of the world's online scam trade, and many people working in the compounds are trafficked there, groups that help them say.
CERN and Mistral plan an AI model for physics
Swipe up for the story ↑Scroll down for the story ↓
CERN, Europe's particle physics laboratory, and Mistral, a French AI company, agreed on Thursday to build a model called Flare.
It will learn from billions of simulated particle collisions in the Large Hadron Collider, the world's largest particle accelerator.
It is meant to replace a chain of separate tools that turn detector signals into data physicists can use.
Why it matters
CERN already holds about one million terabytes of collider data, and the number of collisions will rise tenfold after an upgrade due around 2030.
Why it matters
The deal gives Mistral one of the world's largest science datasets.
2 sources
CERN and Mistral plan an AI model for physics
The full story
CERN and Mistral plan an AI model for physics
CERN, Europe's particle physics laboratory, and Mistral, a French AI company, agreed on Thursday to build a model called Flare.
Why it matters. CERN already holds about one million terabytes of collider data, and the number of collisions will rise tenfold after an upgrade due around 2030.
Banks and shops block outside AI agents
Swipe up for the story ↑Scroll down for the story ↓
Charles Schwab, a big US investment firm, is blocking outside AI agents such as Meta's Muse from moving clients' money, its chief executive told Bloomberg Television.
An AI agent is a program that acts for a person, here using the client's password.
Schwab will let outside AI only check balances, and wants clients to use its own assistant, Charley.
JPMorgan Chase's boss said the same this week, and Amazon and eBay have blocked shopping agents.
Why it matters
Nobody yet knows who pays if an agent moves money wrongly, Schwab's chief said.
Why it matters
Cheq, a firm that checks shopping agents, counted 100 times more agent visits to shops in 45 days.
Who gains
Charles Schwab's own AI assistant, Charley
By blocking outside agents from moving money, Schwab steers clients who want AI help towards its own tool.
2 sources
Banks and shops block outside AI agents
The full story
Banks and shops block outside AI agents
Charles Schwab, a big US investment firm, is blocking outside AI agents such as Meta's Muse from moving clients' money, its chief executive told Bloomberg Television.
Why it matters. Nobody yet knows who pays if an agent moves money wrongly, Schwab's chief said.
Replit says it dropped Anthropic
Swipe up for the story ↑Scroll down for the story ↓
Replit, an app that lets people build software in their web browser, no longer works with Anthropic, its chief executive Amjad Masad said on Thursday.
He said OpenAI's models now do better on Replit's tests and cost less.
A person who knows the deal told TechCrunch that the Anthropic partnership is still in place.
Why it matters
Replit used Anthropic's Claude models as recently as late September, its own blog shows.
Who gains
OpenAI
Replit says it moved its coding app to OpenAI's models on cost and test results.
Nobody knows yet
Whether Replit has really left Anthropic.
Replit's chief executive said so on stage. A person who knows the deal told TechCrunch it is still in place.
The full story
Replit says it dropped Anthropic
Replit, an app that lets people build software in their web browser, no longer works with Anthropic, its chief executive Amjad Masad said on Thursday.
Why it matters. Replit used Anthropic's Claude models as recently as late September, its own blog shows.
AI reviewers delay the next PostgreSQL
Swipe up for the story ↑Scroll down for the story ↓
The team behind PostgreSQL, a popular free database, has pushed its next version back to around 19 November.
Peter Eisentraut, a lead developer, said AI tools that review code found a long list of small flaws, which would once have been found over five years.
Even a three-line fix takes time to handle, and with a long list the team ran out of time, he said.
Why it matters
The pulled features will likely go into version 20 next year, which had been planned as a release of fixes only.
Why it matters
Eisentraut said the extra checks will make version 19 very robust.
Good news
AI tools that review code found a long list of small flaws in PostgreSQL 19, a widely used free database, before its release. Its developers say flaws like these used to turn up in the five years after a launch.
1 source
AI reviewers delay the next PostgreSQL
The full story
AI reviewers delay the next PostgreSQL
The team behind PostgreSQL, a popular free database, has pushed its next version back to around 19 November.
Why it matters. The pulled features will likely go into version 20 next year, which had been planned as a release of fixes only.
487 researchers want hacking AIs kept offline
Swipe up for the story ↑Scroll down for the story ↓
An open letter signed by 487 AI researchers, among them Geoffrey Hinton and Yoshua Bengio, says AI that can carry out cyberattacks should be trained only on computers cut off from the internet.
They want a new international body to set safety standards and to receive reports when AI escapes its controls.
Why it matters
In October last year the two signed a letter warning against building super-intelligent AI.
Part of a signal
AI programs loose on real systems
Seen in 3 of the 12 beats
Also seen in Cyber and Gaming.
The whole signal →1 source
487 researchers want hacking AIs kept offline
The full story
487 researchers want hacking AIs kept offline
An open letter signed by 487 AI researchers, among them Geoffrey Hinton and Yoshua Bengio, says AI that can carry out cyberattacks should be trained only on computers cut off from the internet.
Why it matters. In October last year the two signed a letter warning against building super-intelligent AI.
MI5 sets up a centre to judge AI threats
Swipe up for the story ↑Scroll down for the story ↓
MI5, Britain's domestic security service, is setting up an AI Threat Assessment Centre, its head Ken McCallum announced in his yearly speech on 13 October.
It brings in experts from the AI Security Institute, the National Crime Agency and the National Cyber Security Centre.
Its job is to give the UK government one agreed view of AI risks, he said.
Why it matters
The misuse he fears most, AI helping make biological weapons, has not happened yet, he said.
1 source
MI5 sets up a centre to judge AI threats
The full story
MI5 sets up a centre to judge AI threats
MI5, Britain's domestic security service, is setting up an AI Threat Assessment Centre, its head Ken McCallum announced in his yearly speech on 13 October.
Why it matters. The misuse he fears most, AI helping make biological weapons, has not happened yet, he said.
A conservative group sets out its AI plan
Swipe up for the story ↑Scroll down for the story ↓
Humans First, a conservative group that wants tighter limits on AI, published an 'America First AI Policy' on Friday.
It calls for tougher export controls on China, independent testing of big AI companies and a ban on AI that improves itself.
It also wants to stop the US government using Flock cameras and to pause H-1B visas for foreign workers.
Why it matters
Its plan goes much further than the Trump administration's own AI rules.
1 source
A conservative group sets out its AI plan
The full story
A conservative group sets out its AI plan
Humans First, a conservative group that wants tighter limits on AI, published an 'America First AI Policy' on Friday.
Why it matters. Its plan goes much further than the Trump administration's own AI rules.
Anthropic's staff expect AI to do their jobs
Swipe up for the story ↑Scroll down for the story ↓
Anthropic published a survey of 1,334 of its own staff beside surveys of US adults on Friday.
Most of its staff said AI will do at least 80% of their work by 2030, against about 5% of US workers.
Most of them expect US unemployment above 12% in 2030; it is 4.2% now.
Their middle guess for US economic growth in 2030 was 17.1%, against 5.4% for the public and 2.2% growth last year.
Why it matters
Anthropic's chief economist says the company is about 18 months ahead of other firms in using AI.
Why it matters
A Yale study found AI has so far changed jobs without any link to unemployment.
1 source
Anthropic's staff expect AI to do their jobs
The full story
Anthropic's staff expect AI to do their jobs
Anthropic published a survey of 1,334 of its own staff beside surveys of US adults on Friday.
Why it matters. Anthropic's chief economist says the company is about 18 months ahead of other firms in using AI.
A tracker counts AI-linked job cuts
Swipe up for the story ↑Scroll down for the story ↓
The Center for AI Safety, a research group, runs a dashboard of US job cuts that companies linked to AI.
It counts just over 117,000 such cuts this year, worth about $8.5 billion a year in lost wages.
The counts come from layoff announcements that mention AI, which does not prove AI replaced each worker.
Why it matters
In 2025 about one layoff in 20 was publicly linked to AI.
Nobody knows yet
How many US job cuts AI really caused.
The tracker's 117,000 counts layoffs that companies linked to AI, and its state figures are estimates, not counted layoffs.
The full story
A tracker counts AI-linked job cuts
The Center for AI Safety, a research group, runs a dashboard of US job cuts that companies linked to AI.
Why it matters. In 2025 about one layoff in 20 was publicly linked to AI.
Irish tech workers join a union
Swipe up for the story ↑Scroll down for the story ↓
Ireland's union for tech workers has grown fivefold since it started in 2023, to more than 1,350 members.
Ireland holds the European headquarters of most big tech firms, and more than 170,000 people work in its tech sector.
Meta has roughly halved its Irish staff since the pandemic, and TikTok said it would cut about 300 Irish jobs, mostly in safety teams.
Why it matters
Ireland's AI summit this week had 36 executives and industry bosses as speakers and one union speaker, the union says.
Why it matters
Total tech jobs in Ireland still rose, from about 144,000 in 2021 to over 177,000.
1 source
Irish tech workers join a union
The full story
Irish tech workers join a union
Ireland's union for tech workers has grown fivefold since it started in 2023, to more than 1,350 members.
Why it matters. Ireland's AI summit this week had 36 executives and industry bosses as speakers and one union speaker, the union says.
Microsoft's president hits back on green cards
Swipe up for the story ↑Scroll down for the story ↓
Brad Smith, Microsoft's president, said a US crackdown on green cards for skilled foreign workers could cripple US tech for more than a generation.
Last week the US suspended Microsoft and seven other firms from PERM, the scheme employers use to sponsor workers to stay for good.
Vice President JD Vance said Microsoft cut 6,000 US jobs last year while getting 6,300 H-1B visas.
Why it matters
Smith says Microsoft hired 9,200 Americans in its last financial year and pays visa holders the same as other staff.
Why it matters
The suspension still lets it hire on H-1B visas but closes the main path to staying permanently.
1 source
Microsoft's president hits back on green cards
The full story
Microsoft's president hits back on green cards
Brad Smith, Microsoft's president, said a US crackdown on green cards for skilled foreign workers could cripple US tech for more than a generation.
Why it matters. Smith says Microsoft hired 9,200 Americans in its last financial year and pays visa holders the same as other staff.
A UK boss gives AI only to his best staff
Swipe up for the story ↑Scroll down for the story ↓
Ben Chisell, chief executive of Paysend, a London payments firm, gives general AI tools only to staff who show what they add.
He says AI lets the firm do about 25% more engineering work with the same team.
A UK government study in January found 75% of firms using AI saw better productivity, but 77% saw no change in sales.
Why it matters
Unily, a British software firm, says many companies now hand out AI mainly as a way to cut costs and staff.
1 source
A UK boss gives AI only to his best staff
The full story
A UK boss gives AI only to his best staff
Ben Chisell, chief executive of Paysend, a London payments firm, gives general AI tools only to staff who show what they add.
Why it matters. Unily, a British software firm, says many companies now hand out AI mainly as a way to cut costs and staff.
OpenAI gives $5 million after a maths row
Swipe up for the story ↑Scroll down for the story ↓
OpenAI said on Thursday it will give $5 million to a new Simons Foundation programme to help mathematicians check results made by AI.
The Simons Foundation funds maths and science research.
Two mathematicians say they were close to the same result; OpenAI denies seeing their work.
Why it matters
The whole programme is worth $15 million with money and computing from the foundation, Anthropic and OpenAI.
1 source
OpenAI gives $5 million after a maths row
The full story
OpenAI gives $5 million after a maths row
OpenAI said on Thursday it will give $5 million to a new Simons Foundation programme to help mathematicians check results made by AI.
Why it matters. The whole programme is worth $15 million with money and computing from the foundation, Anthropic and OpenAI.
Microsoft brings Defender to Linux desktops
Swipe up for the story ↑Scroll down for the story ↓
Microsoft released a test version of Defender, its security software, for computers running Linux, the free operating system many developers use.
It needs Defender for Endpoint Plan 2 or a Microsoft 365 E5 or E7 licence.
Ubuntu, a version of Linux, is used by about 28% of professional developers, a 2025 survey found.
Why it matters
Companies that already pay for those Microsoft plans can now watch their Linux machines with the same tool.
How it works
Why a bundle discount makes every rival a second bill
How it works
A buyer who has paid for a whole set of programs pays nothing more to use each one, so a rival's program arrives as an extra bill.
How it works · step 1 of 6
A company sells many programs together, at a lower price than buying each one
How it works · step 2 of 6
A business pays once and gets the whole set
How it works · step 3 of 6
Each program in the set now costs it nothing extra to use
How it works · step 4 of 6
A rival offers one program, perhaps a better one
How it works · step 5 of 6
Choosing the rival means paying for the same job twice
How it works · step 6 of 6
So the business keeps the program it already has
The twist
Once a business has paid for the bundle, Microsoft's own chat or security program costs it nothing more. A rival's program, however good, is a new bill on top.
The picture
What a business sees when it compares the two. The CMA says this is why rivals are hard to choose even when their program is as good or better.
Seen the same day
Firms that already pay for Microsoft 365 E5 can now point Defender at their Linux machines, so a separate Linux security tool has to win a second payment.
Seen the same day
Google puts Gmail Live inside its paid AI plans, so a separate email assistant has to beat a feature those subscribers already pay for.
The catch
It only holds while leaving costs more than it saves.
The catch
AI models are bought one request at a time with no bundle, and almost one in three teams on Vercel's AI service changed their main supplier in September.
The whole
The office manager renewing a Microsoft plan sees a smaller bill this year.
The whole
The start-up with a better chat tool sees buyers who never call back, and neither of them sees the other's side.
1 source
Microsoft brings Defender to Linux desktops
The full story
Microsoft brings Defender to Linux desktops
Microsoft released a test version of Defender, its security software, for computers running Linux, the free operating system many developers use.
Why it matters. Companies that already pay for those Microsoft plans can now watch their Linux machines with the same tool.
Munich and Berlin move to free software
Swipe up for the story ↑Scroll down for the story ↓
Munich is replacing paid software with free open-source programs, its chief digital officer Laura Dornheim said at a conference in Prague.
It is swapping VMware for Proxmox and Oracle databases for PostgreSQL, saving about 1.5 million euros a year.
Munich dropped Linux desktops in 2017 after a new council brought back Windows.
Why it matters
Dornheim says open-source projects in government fail for political reasons, not technical ones.
Good news
Munich says it now saves about a million euros a year by moving its computer systems from paid software to free programs.
2 sources
Munich and Berlin move to free software
The full story
Munich and Berlin move to free software
Munich is replacing paid software with free open-source programs, its chief digital officer Laura Dornheim said at a conference in Prague.
Why it matters. Dornheim says open-source projects in government fail for political reasons, not technical ones.
Capita set to miss a pension deadline
Swipe up for the story ↑Scroll down for the story ↓
Capita, a British outsourcing company, will miss a legal deadline to correct about 93,000 civil service pensions, the PCS union says.
The UK government confirmed the March 2027 deadline would be missed, the union said.
Capita took over the scheme for 1.7 million members in December 2025, and its website and service have struggled since.
The corrections follow a court ruling that pension changes treated younger staff unfairly.
Why it matters
About 73,000 of those waiting are retired or have died.
Why it matters
The UK government is paying Capita an extra 45 million pounds for this work on top of its 239 million pound contract.
Who pays
About 93,000 civil service pension holders
Capita will miss the March 2027 deadline to correct their pensions, the PCS union says.
1 source
Capita set to miss a pension deadline
The full story
Capita set to miss a pension deadline
Capita, a British outsourcing company, will miss a legal deadline to correct about 93,000 civil service pensions, the PCS union says.
Why it matters. About 73,000 of those waiting are retired or have died.
Gmail answers spoken questions for paying users
Swipe up for the story ↑Scroll down for the story ↓
Google released Gmail Live, which lets people search their inbox by asking questions out loud.
Similar tools are coming to Docs and to Keep, Google's notes app.
Why it matters
Google is adding AI features to its paid plans, so people already paying get them with no new bill.
How it works
Google puts Gmail Live inside its paid AI plans, so a separate email assistant has to beat a feature those subscribers already pay for.
The idea, in story 29: Why a bundle discount makes every rival a second bill →1 source
Gmail answers spoken questions for paying users
The full story
Gmail answers spoken questions for paying users
Google released Gmail Live, which lets people search their inbox by asking questions out loud.
Why it matters. Google is adding AI features to its paid plans, so people already paying get them with no new bill.
AI chats nobody else can read
Swipe up for the story ↑Scroll down for the story ↓
Brave, which makes a privacy-focused web browser, now encrypts questions to its AI helper Leo before they leave the user's computer.
The question is only unlocked inside a sealed part of a chip, so Brave cannot read it.
A start-up called TrustedRouter raised $6 million to offer the same kind of protection for open models.
Why it matters
AI companies often use people's chats to train new models, and encryption is meant to stop that.
Good news
Brave's browser now encrypts questions to its AI helper on the user's own computer, so Brave itself cannot read them.
2 sources
AI chats nobody else can read
The full story
AI chats nobody else can read
Brave, which makes a privacy-focused web browser, now encrypts questions to its AI helper Leo before they leave the user's computer.
Why it matters. AI companies often use people's chats to train new models, and encryption is meant to stop that.
MapQuest shows where cameras watch
Swipe up for the story ↑Scroll down for the story ↓
MapQuest, a map app, added a layer that shows known outdoor cameras and number-plate readers.
It includes cameras made by Flock Safety, one of the biggest makers of plate readers in the US.
Downloads of MapQuest jumped in September after it refused to rename Lake Ontario as Lake America.
Why it matters
Plate readers log every passing car's plate, time and place, and police can search those records.
2 sources
MapQuest shows where cameras watch
The full story
MapQuest shows where cameras watch
MapQuest, a map app, added a layer that shows known outdoor cameras and number-plate readers.
Why it matters. Plate readers log every passing car's plate, time and place, and police can search those records.
An AI designs a homemade gun
Swipe up for the story ↑Scroll down for the story ↓
Cody Wilson, who made the first 3D-printed gun more than a decade ago, showed a new one he says an AI designed.
He trained the AI, called GatGPT, on a large online library of gun files.
Experts criticised its odd design and made-up details.
Why it matters
Mainstream AI tools refuse to help design guns, so a custom AI could lower the skill needed to print one.
Why it matters
US states are already in court over guns made at home.
The full story
An AI designs a homemade gun
Cody Wilson, who made the first 3D-printed gun more than a decade ago, showed a new one he says an AI designed.
Why it matters. Mainstream AI tools refuse to help design guns, so a custom AI could lower the skill needed to print one.
Police use an AI-made child to catch a suspect
Swipe up for the story ↑Scroll down for the story ↓
Police in Livonia, a US city, arrested a school maintenance worker who arranged to meet a 15-year-old for sex.
The boy did not exist: an officer made his profile picture with AI.
He was arrested when he arrived and later admitted it.
Why it matters
Child-abuse images made with AI are a fast-growing problem, the US centre for missing children reports.
Why it matters
US border police also use AI-made online characters to watch protesters, documents seen by 404 Media show.
1 source
Police use an AI-made child to catch a suspect
The full story
Police use an AI-made child to catch a suspect
Police in Livonia, a US city, arrested a school maintenance worker who arranged to meet a 15-year-old for sex.
Why it matters. Child-abuse images made with AI are a fast-growing problem, the US centre for missing children reports.
Fake experts trouble a research firm
Swipe up for the story ↑Scroll down for the story ↓
GLG, a firm that connects investors with experts for paid calls, is adding checks against fake experts.
Its chief executive Gemma Postlethwaite says demand for anyone who knows AI has drawn people using deepfakes and chatbots.
GLG now checks identity on camera with questions meant to trip up AI.
Why it matters
GLG also finds experts such as child psychiatrists to help AI labs train their models on hard subjects.
1 source
Fake experts trouble a research firm
The full story
Fake experts trouble a research firm
GLG, a firm that connects investors with experts for paid calls, is adding checks against fake experts.
Why it matters. GLG also finds experts such as child psychiatrists to help AI labs train their models on hard subjects.
A tool checks lawyers' AI citations
Swipe up for the story ↑Scroll down for the story ↓
CiteProof, a new online tool, checks legal briefs for court cases that AI made up.
A check costs $49 for one brief, or $499 a month for 200.
Last year a Georgia court decision was thrown out after its order was found to cite cases that did not exist.
Why it matters
Lawyers keep using AI to draft, and judges have fined and rebuked those who filed invented cases.
1 source
A tool checks lawyers' AI citations
The full story
A tool checks lawyers' AI citations
CiteProof, a new online tool, checks legal briefs for court cases that AI made up.
Why it matters. Lawyers keep using AI to draft, and judges have fined and rebuked those who filed invented cases.
Elorian's first model looks twice
Swipe up for the story ↑Scroll down for the story ↓
Elorian, a start-up led by a former Google DeepMind researcher, released its first AI model, Glimpse, on Friday.
Most AI models look at a picture once, then reason in words; Glimpse can zoom in and look again while it thinks.
It was trained on a few hundred chips and costs $3 per million tokens of output.
Why it matters
The claims are Elorian's own tests, not independent ones.
1 source
Elorian's first model looks twice
The full story
Elorian's first model looks twice
Elorian, a start-up led by a former Google DeepMind researcher, released its first AI model, Glimpse, on Friday.
Why it matters. The claims are Elorian's own tests, not independent ones.
The truth, across the day
Swipe up for the truth ↑Scroll down for the truth ↓
What's really going on
On the same day China proposed licences for exporting its biggest AI models and chip designs, and Taiwan said shipping AI servers to China will become a crime.
That's the whole Tech day.
39 stories, 55 sources.