Day Lila

Information Technology · Tuesday, 27 October 2026

01 Briefing what happened

OpenAI says its AI disproved the Exponential Time Hypothesis, a computer-science belief hundreds of papers rest on, and asks scientists who proposed it to check

37 stories 51 sources

OpenAI claims results in computer science, maths and physics from its AI, now with outside experts to check; plus a cheaper OpenAI model, Texas suing OpenAI, Apple replacing broken iPhone Duos and 33 more stories.

3

results OpenAI says its AI found, in computer science, maths and physics

each is now with a group of outside experts [2][3]

16 Oct

the day OpenAI says its model found the disproof

a proof-checking program passed it the next day [1]

1999

the year the Exponential Time Hypothesis was proposed

two of the three people checking OpenAI's claim helped propose it [2]

1,000+

mathematicians who have pledged not to work for AI companies

their group formed after OpenAI posted 722 papers at once [4]

The lead story — what happened

  • OpenAI said on Monday that one of its internal AI models has disproved the Exponential Time Hypothesis, an idea about computing that most experts believed true. [1][2]
  • The hypothesis is about how hard it is to tell whether a long list of yes-or-no conditions can all be true at once. [1] Hundreds of papers lean on it to show that problems in network maths, biology and databases are hard to solve quickly. [1]
  • OpenAI's paper gives a method that counts the answers to such puzzles faster than the hypothesis said was possible, OpenAI says. [2]
  • OpenAI says the model found it on 16 October, and Lean, a program that checks every step of a proof by machine, passed it the next day. [1] The write-up and the Lean proof are public on GitHub, a code-sharing site. [1]
  • Russell Impagliazzo and Ramamohan Paturi of UC San Diego, who helped propose the hypothesis in 1999, and Ryan Williams of MIT will check it. [2] They wrote that taking part 'should not be taken as certification of the claim'. [1]
  • Williams says the three signed no secrecy agreement and are not being paid. [1] They asked OpenAI to release all its raw data and to pay for a conference to examine the claim, and OpenAI agreed. [1]
  • OpenAI also claims a maths result: a proof of the local smoothing conjecture, set by Christopher Sogge in 1991, in every dimension. [1] That would also settle the Kakeya conjecture, which Hong Wang solved in three dimensions and won a Fields Medal for a few months ago. [1][2]
  • Its third claim is in physics: a proof that theories describing quarks and gluons, tiny particles inside atoms, exist in four dimensions. [1] Part of it touches one of the Clay Millennium Prize problems, famous unsolved questions. [1] A group led by David Gross, a Nobel physicist, has kept that material back to write it up itself. [1][3]
  • Alex Iosevich, a mathematician at the University of Rochester, called the maths result a 'historic development'. [3] He also said he felt 'a twinge of sadness'. [3]
  • A day earlier AGMAI, a panel that advises AI firms on maths, said labs should not hand results to chosen experts first. [2] It named unequal access and borrowed reputation as the problems. [2] OpenAI admits its choice of experts 'creates asymmetry of information'. [3]
  • More than 1,000 mathematicians have joined a group that pledges not to work for AI companies, after OpenAI posted 722 maths papers at once on 6 October. [4][1]
  • 'If the claim is correct, it will change an important part of theoretical computer science,' the three checkers wrote. [1]

Who is involved

  • OpenAI

    the company that makes ChatGPT; it says its models found all three results

  • Russell Impagliazzo and Ramamohan Paturi

    computer scientists at UC San Diego who helped propose the hypothesis; they will check the disproof

  • Ryan Williams

    a computer scientist at MIT; he checks it with them, unpaid

  • David Gross

    a Nobel-winning physicist; his group will write up the physics result

  • AGMAI

    an advisory group on maths and AI; it warned against giving chosen experts the results first

How it unfolded

  1. 1991 Christopher Sogge sets out the local smoothing conjecture [1]
  2. 1999 The Exponential Time Hypothesis is proposed [2]
  3. 6 Oct OpenAI posts 722 maths papers at once [1]
  4. 16 Oct OpenAI says its model finds the disproof; Lean passes it on 17 October [1]
  5. 26 Oct OpenAI announces the three results and the expert groups [1][2]

Where this points

The next test is the open check led by Impagliazzo, Paturi and Williams, with the conference OpenAI agreed to fund, and no date has been set for its findings. [1]

Signals on this beat

The bar and the word are our reading of how hard each one is pushing, and the arrow is where it is heading.

Makers grading their own claims Building↑

OpenAI says its new Terra model did tasks for about a ninth of the cost of Anthropic's best model, in its own tests. [6] Cadence says its chip-design AI gives up to 30% better results than a general model. [36] Apple says only a 'small number' of iPhone Duos have a screen fault, and gives no figure. [9][10]

AI taking on people's work High↑

The Dutch government plans 25,000 fewer jobs and will set its AI goals in early 2027. [15] Bedrock Robotics raised $600 million for diggers that drive themselves, as building workers retire. [34] Apollo's chief economist says AI has not yet lifted profits outside the big tech firms. [39]

Paying to switch features on Building↑

Nvidia's gaming cards run one kind of AI maths at half speed, and hobbyists found a way to lift the limit. [23] Tesla charges $99 a month to turn on its Full Self-Driving software. [33] T-Mobile's cheapest new plan streams video at 480p, a low quality. [31]

Staff and critics pushing back Building↑

More than 300 Boston Dynamics staff signed a letter against selling robot dogs to ICE, the US immigration agency. [14] More than 1,000 mathematicians pledged not to work for AI companies. [4] Flock's boss told police chiefs that protests against his cameras are not real. [25]

The rest of the day

36 more stories on this beat.

Each with its own sources. None of these is a link to the story above.

  1. 02

    OpenAI sells a cheaper middle model

    OpenAI released GPT-6.1 Terra on Monday, a model in the middle of its range, below Sol and Astra and above Luna. [5][6] Developers pay $1 per million tokens, small pieces of words, going in and $5 per million coming out, half the price of the Terra it replaces. [6] OpenAI says Terra matched GPT-6 Astra, its strongest model, on one test while costing a tenth as much per token. [7] It also says Terra did tasks for about a ninth of the cost of Anthropic's Claude Opus 5.5, in its own tests, which nobody outside has checked. [6]

    Why it matters — Terra is built for high-volume office work such as reading invoices into accounting software. [5][6] A data scientist quoted by The Register says prices keep falling because the firms that run AI models compete with each other. [5]

  2. 03

    Texas sues OpenAI over safety claims

    Texas sued OpenAI and its boss, Sam Altman, on Friday, saying the company sells ChatGPT as safe while its safeguards fail, including for children. [8] The case uses the state's law against misleading trade and tries to make Altman personally liable, which is unusual. [8] Texas wants fines of up to $10,000 per violation and a ban on children using ChatGPT. [8] It also wants a court-appointed expert to approve new models. [8] OpenAI said it works with experts to improve its safeguards and wants rules for the whole industry. [8]

    Why it matters — Ken Paxton, the Texas attorney general who brought the case, is in a close race for the US Senate. [8] He has sued or settled with Roblox, TikTok and Meta in the last year, and Meta paid Texas $1.4 billion in 2024 over face recognition. [8]

  3. 04

    Apple will replace cracked iPhone Duos

    Apple said on Monday it knows of 'an issue impacting the inner display' on a small number of iPhone Duos, its first folding phone. [9][11] Owners whose screen fails can get a replacement through AppleCare, Apple's repair service. [10] Since the phone went on sale on Friday, owners had posted pictures of bright lines and cracks along the fold. [12] Apple did not say what causes it or how many phones are hit. [10][11] Outside the warranty, an inner-screen repair costs $999, and the phone starts at $1,999. [10]

    Outside the warranty, fixing the inner screen costs about half the price of a new Duo.

    Why it matters — Apple told Bloomberg that affected phones show the fault soon after first use. [10] The BBC notes that some pictures of broken phones online may have been made with AI, so the real number is hard to judge. [11] A phone that CNET tested has folded more than 55,000 times so far without breaking, Engadget reports. [13]

  4. 05

    Robot-dog staff protest an ICE sale

    More than 300 of Boston Dynamics' roughly 1,400 staff signed a letter asking bosses to stop a sale of its Spot robot dogs to ICE, the US immigration agency. [14] Boston Dynamics is owned by Hyundai, the Korean carmaker. [14] A $1.3 million US contract from 25 September pays Radeco, a Connecticut firm, for four Spots it has fitted to detect gas, radiation and explosives in tunnels. [14] The company stands by the deal and says its terms ban using Spot as a weapon. [14]

    Why it matters — The letter points to a 2025 raid that held 475 mostly South Korean workers at a Hyundai factory in Georgia. [14] Staff fear the deal will make hiring harder while the company plans 1,250 new jobs by 2033. [14]

  5. 06

    Dutch government plans 25,000 fewer jobs

    The Dutch government wants to cut 25,000 full-time jobs, about 16% of its workforce, by 2035. [15] It aims to save 900 million euros by 2027, rising to 3.5 billion euros by 2035. [15] Police and defence are left out. [15] The cabinet says it will set goals for using AI to work more productively in early 2027. [15] So how AI is meant to replace the work is not yet clear. [15]

    What the Dutch government says the cuts will save, in billions of euros.

    Why it matters — Eric van der Burg, the state secretary in charge, told parliament he wants a 'simpler, smart and smaller' government. [15] The first savings target falls in 2027, the same year the AI plan is due. [15]

  6. 07

    Denmark's break from Microsoft Office ends

    A Danish ministry's test of swapping Microsoft Office for Collabora, a British open-source office suite, ended in under six months, Computerworld Denmark reported. [16] Open-source means its code is free for anyone to use and change. The test began in 2025 with 38 staff, under a minister who wanted less reliance on American technology. [16] Staff complained about formatting, blank pages and tracking changes. [16]

    Why it matters — France and the Netherlands are also moving away from US software. [16] A Copenhagen Business School professor says the Danish ministry never properly studied the test, so other public offices cannot learn from it. [16]

  7. 08

    China lets VW and Xpeng drive further alone

    Volkswagen has a licence to test Level 3 self-driving, where the car drives itself in some conditions with a driver ready to take over, on public roads in Hefei, China. [17][18] Xpeng, a Chinese carmaker, may now run Level 4 robotaxis across Guangzhou, at first for riders with invitation codes. [17] China stopped issuing new robotaxi licences in April after an incident with a Baidu robotaxi in Wuhan, and started again in July, Bloomberg reported. [17]

    Why it matters — It is a test permit, so VW cannot sell Level 3 cars yet, and it aims to have the technology ready in 2027. [18] Xpeng wants to carry riders with no safety driver in 2027 and is talking to European ride-hailing firms. [17]

  8. 09

    Chinese research ships linger near Pacific cables

    Chinese research ships spend more time near undersea internet cables in the Pacific than ships from any other country, a report from CSIS, a US research group, found. [19] It compared 200 research ships from 47 countries. [19] Between July and October 2024 one Chinese ship spent time near three cables linking the US with Australia, Indonesia, Singapore and Guam. [19] The ships moved back and forth in a pattern used to map the sea floor. [19]

    Why it matters — The report says such maps could help sabotage or submarine missions, though the ships are not part of China's navy. [19] A report last week counted 500 cases of Russian ships watching cables over five years. [19]

  9. 10

    Chinese team sets a fibre speed record

    A Chinese team says it sent 3.2 petabits of data a second down one optical fibre cable in real time, a record for this kind of test. [20] A petabit is a million billion bits, and the team compares it to downloading more than 20,000 4K films a second. [20] The same team reached 2.5 petabits earlier this year, so the speed rose almost 30% in a year. [20] The cable has 24 cores, separate glass threads that each carry light. [20]

    Data sent down one 24-core fibre each second. The first two used commercial-grade hardware.

    Why it matters — The test used hardware built to commercial standards, not lab-only equipment, so it is closer to a cable that could be laid. [20] Lab tests on similar fibre have reached 5.29 petabits a second. [20]

  10. 11

    Intel's boss has talked to AMD's

    Lip-Bu Tan, Intel's chief executive, has spoken with Lisa Su, who runs AMD, about working together, Tom's Hardware reports. [21] The two companies have competed for decades to make the main processors in PCs and data centres. [21] At a talk on 24 October, Tan said Intel's 18A chip-making process has reached volume production. [21] He said its next process, 14A, should be ready for volume production in early 2027. [21]

    Why it matters — Tan wants other companies to have their chips made in Intel's factories, but he named no new customers. [21] Neither company has announced any deal from the talks. [21]

  11. 12

    Micron sets out its next memory chips

    Micron, a US maker of memory chips, set out how it will keep shrinking DRAM, the working memory inside computers. [22] Its next process should enter volume production in the second half of 2027, with about a quarter more memory in the same space. [22] After that it plans wafer bonding, joining chip layers together, then High-NA EUV, the newest machines for printing the finest chip patterns. [22] 'DRAM is nearing atomic limits,' its technology chief, Scott DeBoer, said. [22]

    Why it matters — Micron is also designing new kinds of memory to sit directly on top of processors for AI work. [22]

  12. 13

    Hobbyists unlock Nvidia cards' AI maths

    Hobbyists found a way to remove a limit Nvidia puts on some AI maths in its RTX 20, 30 and 40 gaming graphics cards, Tom's Hardware reports. [23] Nvidia's own specifications run one kind of this maths at half speed on these cards, but not on its pro cards. [23] With the limit removed, an RTX 4090's peak rate doubled, from 165.6 to 331.6 trillion operations a second. [23] The trick uses a flaw in Nvidia's own security firmware. [23]

    Why it matters — Real tasks gain much less than double, and chatbots reply no faster. [23] The projects warn that the unlock can damage the card. [23]

  13. 14

    IBM will use AI to tidy its mainframes

    IBM announced IBM Z Atlas, a coming AI tool for its Z mainframes, the large computers it sells for heavy business work. [24] The tool will map how the parts of a customer's system connect and test changes before they go live. [24] IBM says untested changes are the leading cause of unplanned outages on these machines. [24] People, not the AI, keep the final say, IBM says. [24]

    Why it matters — IBM's chief executive said last week that its newest Z17 mainframes have outsold every earlier generation by $1 billion. [24] Rivals pitch AI for moving mainframe programs onto other machines, while IBM wants AI to keep customers on its own. [24]

  14. 15

    Flock's boss says the protests are not real

    Garrett Langley, chief executive of Flock, whose cameras read car number plates for police, told police chiefs on Sunday that protests against his company are not real, 404 Media reports from leaked audio. [25] He claimed, without evidence, that most online criticism of Flock came from Russian bots. [25] He said online chatter about Flock is down about 95% since July. [25]

    Why it matters — He repeated a disproven theory that Ukraine's drone strikes on Russian data centres cut bot activity. [25] He also said his wife had advised him never to apologise. [25]

  15. 16

    A bigger-screen iPad mini

    Apple announced a new iPad mini with a larger 8.5-inch OLED screen, where each tiny dot makes its own light, and a faster A20 Pro chip. [26] It starts at $699, or $849 with a mobile connection. [26] Apple says it runs AI tasks up to 4.5 times faster than the last model. [26] Orders open on 28 October and it goes on sale on 6 November. [26]

    Why it matters — It arrives two days before Apple's 'Pro and Tell' event on Wednesday. [26] It runs the new Siri that Apple released for iPhones the same day. [27]

  16. 17

    Siri now remembers past chats

    Apple released iOS 27.2 on Monday, the first big update to its newest iPhone software. [27] Siri can now recall earlier conversations and look at voicemails, Safari history and photos in messages. [27] Users can choose which apps it may read. [27] The test version of Siri's AI now works in French, Japanese, Korean, Portuguese and Spanish. [27] FaceTime calls can now use the front and back cameras at once. [27]

    Why it matters — Deaf users can now add a sign-language interpreter or a live captioner to a video call. [27] The same update lets the iPhone Duo use an Apple Pencil. [27][10]

  17. 18

    Instagram tests 90-minute videos

    Instagram is testing long, widescreen videos it calls Features, of up to 90 minutes. [28][29] The test starts small in the US and with some creators in Brazil. [28] Adam Mosseri, who runs Instagram, said he had been against long videos 'for a long time'. [29] He worried that long viewing means fewer photos and Reels sent between friends. [28]

    Why it matters — The move is aimed at YouTube, which has become a big hit on TV screens. [30] Unlike YouTube, Instagram does not yet share ad money with creators, and says it is too early to name a payment plan. [29]

  18. 19

    T-Mobile plans let users swap extras

    T-Mobile, a large US phone network, launched T-Mix plans that bundle phone service with picks such as Netflix, YouTube Premium or Xbox Game Pass. [31] The cheapest plan, at $55 a month, gives one pick, which can be changed up to four times a year. [31] It streams video at 480p, a low quality, while the $75 and $95 plans stream up to 4K. [31] Some picks can only be swapped once a quarter or less often. [31]

    Why it matters — Video quality is one of the things that separates the cheap plan from the dearer ones. [31] Picks can also be traded for money off a new phone, CNET noted. [31]

  19. 20

    Turo marks Teslas with self-driving on

    Turo, a site where people rent out their own cars, will show which Teslas have Full Self-Driving (Supervised) switched on, from 28 October. [32][33] The software handles most driving but still needs a driver watching. [32] Hosts get one free month, then pay Tesla's usual $99 a month. [33] Tesla says about 1.7 million customers have paid for it, out of more than 10 million cars it has sold. [32]

    Why it matters — Electrek notes that Full Self-Driving subscriptions count toward Elon Musk's pay package. [33] Turo says thousands of Teslas with the software on are already listed. [33]

  20. 21

    Self-driving diggers raise $600 million

    Bedrock Robotics, a two-year-old start-up, raised $600 million at a value of $4.1 billion. [34] It sells a kit that makes existing excavators, the big digging machines, drive themselves. [34] Its machines now work on real sites, including a water treatment plant in Nevada. [34] Its chief executive, Boris Sofman, once ran the self-driving lorry project at Waymo, Google's sister company. [34]

    Why it matters — It takes five years or more to master an excavator, Bedrock says. [34] A builders' trade group expects more than 40% of US construction workers to retire in the next five years. [34]

  21. 22

    Ex-GitHub boss opens a rival

    Entire, a coding platform from Thomas Dohmke, the former head of GitHub, opened to all developers on Monday. [35] It is built for the flood of AI-written code that, ZDNet says, has caused growing outages at GitHub. [35] Entire replaces the usual step where people read each change with 'Trails', where AI agents score every change. [35] Developers can merge code when the checks pass 'without needing to look at the code itself', the company says. [35]

    Why it matters — Entire claims its code hosting is up to 89 times faster than GitHub's. [35] The design lets a team merge code that no person has read. [35]

  22. 23

    Chip-design firm trains its own AI

    Cadence, one of the two biggest makers of software for designing chips, unveiled Maestro, AI models trained for chip-design work. [36] It claims up to 30% better results and up to five times fewer tokens than a general AI model. [36] Some customers get early access late this year and can train the models on their own designs. [36]

    Why it matters — Its rival Synopsys built its model to run on OpenAI's computers without using customer data, so the two now differ on where a customer's designs go. [36] A Cadence executive said the cost of running AI agents had 'really started to explode'. [36]

  23. 24

    A resale shop blocks AI shoppers

    The RealReal, a site that resells designer clothes, made it harder for AI agents to log in, fill carts and buy. [37] The Verge had shown that Meta's Muse agent could re-add an item to a cart each time the shop's 20-minute hold ran out. [37] That blocked other buyers for hours. [37] Agents can still find listings and send people to them, the company's technology chief said. [37]

    Why it matters — Meta says shopping is the most common use of Muse, with purchases over $200 on average. [37] Amazon has also blocked agents from shopping on its site. [37]

  24. 25

    Fanatics builds a shopping helper with OpenAI

    Fanatics, which sells official sports clothing, built a 'Shop the Look' feature with Codex, OpenAI's coding tool. [38] It suggests matching team gear that is in stock on each product page. [38] Fanatics sells more than 1.5 million different items across hundreds of websites. [38] It says the feature raised how much shoppers engage with products by 10%. [38]

    Why it matters — Team merchandise is still Fanatics' biggest source of money. [38] It makes more than 45 million items a year and sells in more than 180 countries. [38]

  25. 26

    Apollo sees little AI payoff outside big tech

    Torsten Slok, chief economist at Apollo, a large investment firm, says AI has not yet raised profits for US companies outside the tech sector. [39] Revenue per worker is rising at the seven biggest tech firms, such as Apple, Nvidia and Tesla, but not across the rest of the S&P 500, a list of 500 large US companies. [39] Apollo also sees growth in AI spending slowing, as firms switch to cheaper models when prices rise. [39]

    Why it matters — Some critics answer that computers also took a decade to show up in productivity figures. [39] If firms stay unconvinced, the data centres being built to serve them would earn less, The Register notes. [39]

  26. 27

    A database firm cuts its AI bill by 80%

    Aiven, a European company that runs open-source databases, says it cut AI spending by about 80% for its heaviest users. [40] It moved most of their work to open-weight models, whose inner settings are published so anyone can run them. [40] It still uses top models from firms like OpenAI to plan the work. [40] Aiven plans to sell the tool it built for swapping models to its customers. [40]

    Why it matters — The Register says developers are mixing in cheaper models, including Chinese ones, to cut their bills. [40]

  27. 28

    The internet's address book prepares for quantum

    ICANN, the body that coordinates the internet's address system, expects work to protect it from quantum computers to speed up, its technology chief told The Register. [41] The domain name system, or DNS, turns website names into the numbers computers use. [41] Its answers are signed with encryption so they can be trusted, and future quantum computers could break that encryption. [41] John Crain said the system may need to be ready in three or five years, not ten. [41]

    Why it matters — Bigger keys would resist quantum attacks, but they run into the DNS's limits on message size. [41]

  28. 29

    Teams will warn about scam chats

    Microsoft is adding an AI feature to Teams, its work chat app, that reads conversations with new contacts for signs of phishing, scams and identity theft. [42] Microsoft says no messages will be stored or used to train AI. [42] It starts in early November, switches on automatically and should be fully available in the first half of December. [42]

    Why it matters — Microsoft had already warned of attackers posing as help-desk staff on Teams to get control of people's computers. [42]

  29. 30

    Gelsinger backs a new kind of memory

    Pat Gelsinger, Intel's former chief executive, is backing Nagare Labs through Playground Global, the investment firm he joined last year. [43] Nagare is building Z2RAM, a memory chip meant to be nearly as fast as SRAM, the quickest memory on a chip, while packing data almost as tightly as DRAM. [43] It was founded by Georgia Tech researchers Suman Datta and Shimeng Yu. [43]

    Why it matters — Gelsinger says memory has had no major new type in three decades. [43] Many new memory types have failed before, The Register notes. [43]

  30. 31

    ChatGPT sends support cases to reporters

    ChatGPT keeps telling people with ordinary tech problems to email 404 Media, a small technology news site, the site reports. [44] A man in Russia, locked out of 13 years of Instagram photos, was told by ChatGPT to write to a named 404 Media reporter. [44] The chatbot then wrote the email for him. [44] The site says it now gets many such AI-written emails. [44]

    Why it matters — 404 Media says people turn to ChatGPT because big platforms make it almost impossible to reach a human. [44]

  31. 32

    AI agents pitch stories to editors

    AI agents are emailing story ideas to magazines, Wired reports. [45] Defector, a sports site, has had daily pitches from agents since September. [45] Many come from iLands, a start-up where people set up bots that try to earn real money. [45] Its agents need tokens, bought with money, to keep running. [45] Editors told Wired they will not publish bots, and Wired found no agent that had sold a story. [45]

    Why it matters — One agent asked the Philadelphia police for comment on a story before its owner reined it in. [45]

  32. 33

    Anthropic's filing names its computer suppliers

    Anthropic's share-sale papers list deals for computing power that The Register adds up to between $782 billion and $807 billion. [46] They include at least $114.5 billion with Google, $110 billion with Amazon and up to $120.5 billion with SpaceXAI. [46] The filing warns that Amazon, Google and Microsoft, which rent it computers, also compete with it and may serve themselves first. [47][46]

    Why it matters — Anthropic says it may pay for more of its own equipment up front if it cannot get enough computing from them. [46]

  33. 34

    Anthropic's filing warns about AI building AI

    Anthropic's filing spends 94 pages on risks, Wired counts. [48] It warns that the risks grow as AI systems take over AI research, if progress outruns the company's ability to test and protect new models. [48] Elsewhere it calls AI helping to build AI one of its biggest advantages over rivals. [48] Bills in the US Congress, one backed by Bernie Sanders, propose banning it. [48]

    Why it matters — The filing also warns that if Anthropic does not use AI this way, it could fall further and further behind rivals that do. [48]

  34. 35

    Bending Spoons closes its Miro deal

    Bending Spoons, an Italian company that buys older internet firms, completed its $1.355 billion all-cash purchase of Miro on Monday. [49] Miro makes an online whiteboard used by more than 250,000 organisations. [49] Some former Miro shareholders put about $295 million back into new Bending Spoons shares. [49]

    Why it matters — Bending Spoons bought Airtable weeks before it agreed the Miro deal. [49] It will add Miro to its figures when it next reports results. [49]

  35. 36

    Marissa Mayer defends her AI app

    Marissa Mayer, the former head of Yahoo, defended using Dazzle, her start-up's AI assistant, to describe her daughter for a teacher. [50] She had told The New York Times the app said her daughter was 'empathetic, creative and an animal lover'. [50] Users on X criticised her for handing the question to an AI. [50] On Monday she said she used it because a few bullet points felt too limiting. [50]

    Why it matters — Mayer said the response to Dazzle so far has been 'terrific'. [50]

  36. 37

    BlackBerry's car software moves into robots

    BlackBerry, once a big phone maker, now sells QNX, software that runs systems like braking in more than 275 million cars. [51] Its chief executive, John Giamatteo, told CNBC that robots are one of QNX's fastest-growing businesses. [51] Nvidia has made QNX the base of its Halos safety software for robots, he said. [51] QNX's backlog of future royalties has passed $1.2 billion. [51]

    Why it matters — BlackBerry's shares are up 142% this year. [51] The company is worth about $5.4 billion, against nearly $85 billion in 2008. [51]

02 Lesson why it matters

Why a chip maker switches off part of the cheaper version of its own chip

A maker selling one chip to gamers and to businesses cannot charge both the same, so the cheap version comes with part of its speed turned off.

The twist

The cheap version is slowed so that buyers who need full speed still buy the dearer card.

The picture

Peak speed of one kind of AI maths on an Nvidia RTX 4090, before and after hobbyists lifted the limit. Real tasks gained much less.

How it works

  1. A maker builds one chip and sells it to two kinds of buyer
  2. Businesses will pay far more than gamers for the same speed
  3. One price would be too dear for gamers or too cheap for businesses
  4. So the cheaper version has part of its speed switched off
  5. Buyers who need full speed pay more for the version with nothing switched off

The same force, elsewhere today

Where this chain is also running, in today's other stories.

  • Nvidia's gaming cards unlocked

    Nvidia's own specifications run one kind of AI maths at half speed on its gaming cards and at full speed on its pro cards.

  • Turo marks Teslas with self-driving on

    Tesla sells the car once and switches Full Self-Driving on only for drivers who pay $99 a month.

  • T-Mobile's new plans

    The $55 plan streams video at 480p, and 4K comes only with the $75 and $95 plans on the same network.

  • OpenAI's cheaper Terra

    Terra costs a tenth of Astra and is a bit worse at most things, so buyers who need the best still pay for Astra.

Where you've seen this

Air travel

the cheapest tickets cannot be changed, so business travellers who change plans buy dearer fares

Free apps

the free version adds a watermark or limits file size, so offices pay for the full one

Train tickets

off-peak tickets cannot be used at rush hour, so commuters buy dearer peak tickets

The catch

It works only while the lock holds: hobbyists broke Nvidia's with a flaw in its own firmware, though real tasks gained far less than the peak.

And the whole of it

Each buyer sees only the price they paid and the version they got. The maker sees every buyer at once, and a phone plan or a car subscription puts most of us on one side of a switch like this.

03 Truth what's really going on

What is really going on

OpenAI says its AI disproved the Exponential Time Hypothesis, but the three computer scientists asked to judge it are only starting to check, unpaid. Many of the day's other big numbers also come only from the companies that gain from them: OpenAI's own tests of its cheaper Terra model, Apple's 'small number' of broken Duos and Cadence's 30% better chip designs. [1][6][9][36]

Why it works on us — A famous problem plus a computer-checked proof sounds finished, so the claim spreads at once while the experts' check has no end date. [1]

Who gains

  • OpenAI — Well-known scientists will examine and explain its results in public. AGMAI, an advisory panel, warns that this kind of early access lends AI firms borrowed reputation. [1][2]
  • Nvidia — Its pro cards keep full speed on one kind of AI maths that its own specifications halve on its gaming cards. [23]
  • Tesla and Elon Musk — Each Turo host who keeps Full Self-Driving after the free month pays $99 a month, and those subscriptions count toward Musk's pay package. [33]
  • IBM — Z Atlas makes its mainframes easier to keep, and IBM says its newest ones outsold every earlier generation by $1 billion. [24]
  • Makers of open-weight AI models — Aiven moved most of its heaviest users' AI work to them and cut that spending by about 80%. [40]

Who pays

  • Impagliazzo, Paturi and Williams — They are checking OpenAI's proof without pay and under no secrecy agreement. [1]
  • Dutch civil servants — The Dutch government plans 25,000 fewer full-time jobs by 2035, outside the police and defence. [15]
  • iPhone Duo owners outside the warranty — An inner-screen repair costs $999 on a phone that starts at $1,999. [10]
  • Boston Dynamics staff who object to the ICE deal — Their bosses kept the sale and asked them to 'respect and accept this decision'. [14]
  • Turo hosts who keep Full Self-Driving — After one free month they pay Tesla $99 a month, a cost they will want back from renters. [33]

What nobody knows yet

Open questions from across today’s stories — ours included.

  • 01

    Whether the Exponential Time Hypothesis has really been disproved.

    The three checkers say their part is not a certification, and no date is set for their findings. [1]

  • 02

    Whether the physics proof holds.

    David Gross's group has not released the AI's material and will write up the work itself first. [1][3]

  • 03

    What mathematicians had proved before OpenAI's maths result.

    Al Jazeera says local smoothing had been proved only in three dimensions. The Next Web says two dimensions in 2019, and the related Kakeya conjecture in three in 2025. [3][1]

  • 04

    How many iPhone Duos have the screen fault, and why.

    Apple gave no number and no cause, and the BBC says some pictures of broken phones online may be made with AI. [10][11]

  • 05

    Whether Terra really matches Astra, and undercuts Anthropic's model, outside OpenAI's own tests.

    The figures come from OpenAI's own testing and have not been checked by anyone outside. [6]

  • 06

    How the Dutch government will use AI to cover 25,000 fewer jobs.

    The cabinet will set its AI goals only in early 2027. [15]

  • 07

    What ICE will use the four robot dogs for.

    The contract names tunnel and smuggling work. Ars, citing the Brennan Center, says the same unit has also carried out deportation raids. [14]

  • 08

    Whether Intel and AMD will agree anything.

    Tan gave no details of his talks with Lisa Su, and neither company has announced a deal. [21]

04 Hope carry this

OpenAI agreed to release all its raw data on its computer-science claim and to pay for a conference where outside scientists can examine it, after the three people checking it asked.

Also true today

  • Apple will replace iPhone Duos whose inner screens fail, through its AppleCare repair service.
  • Deaf iPhone users can now bring a sign-language interpreter or a live captioner into a FaceTime call, with Apple's iOS 27.2 update.
  • A Chinese team says it sent 3.2 petabits of data a second down one fibre cable on hardware built to commercial standards, almost 30% faster than its own record from earlier this year.

Across the beats