All AI news

AIZvi Mowshowitz reports that OpenAI released 722 math manuscripts from an internal frontier model on GitHub, later reduced to 719 after three withdrawals, covering 90 of the top 500 open problems. He says the work came mostly from a single prompt, with an average of three hours of compute per solution. Mathematicians reacted with mixed feelings, and the post highlights concerns about unread papers, cryptography implications, and the role of Lean verification.

Why it matters: The post traces how OpenAI's release of 719 math manuscripts divided mathematicians and reshaped verification, credit, and publication norms in the field.

AISimon Willison says he built a Newsletters index for his blog almost entirely by voice, using the ChatGPT desktop app's Codex voice mode while cooking dinner. The feature imports weekly Substack posts via RSS and undocumented API, monthly newsletters from a GitHub archive repository, and a private sponsors-only newsletter. He says he switched back to typing for review and fixes before deploying the pull request.

AIClément Delangue of Hugging Face says he loves the building-in-public approach, responding to Matth Lapeyre's post about Microduck's battery testing. Lapeyre reports the new custom board ran over 3 hours on a 2600 mAh battery, with average current dropping from about 1.13 A to 0.82 A and peak CPU temperature falling from 114°C to 60°C without throttling.

AIGoogle AI announces that SynthID.com is now available globally in English for verifying AI-generated images, video, and audio. The post also lists Nano Banana 2.1, EmbeddingGemma 2, Gemma 4 with BOTANIC-1, a Gemini business agent, and Guided Vision in Gemini Live.

AISimon Willison says he built a new feature for his blog entirely by voice using Codex Desktop while cooking dinner. The post links to a write-up on his site about the voice-driven workflow.

AIMerve Noyan, of Hugging Face, says a new guide on speculative decoding is in progress, followed by one on quantization. She points readers to the Llama App docs page comparing prefill and decode.

AIAlibaba's Qwen team says Qwen-Image-2.1-Turbo is an accelerated checkpoint of Qwen-Image-2.1 on the same 7B architecture, now with open weights. It generates 2K images from text in 8 denoising steps and supports natural-language edits, with Pro and Turbo APIs also live.

AITian Keyu, the Peking University PhD student known as the "ByteDance poisoning intern," has a 10-person world-model lab valued at $200 million after $30 million from Fivesource Capital and IDG, according to Leiphone. The lab plans to train a foundation model on about 100 million hours of video using a 200,000-symbol visual vocabulary, with a 2027 release targeted. Tian says the approach could cut the cost of generating one second of video by at least an order of magnitude.

AISantiago says a friend at a large company stopped spending an hour each Monday matching Stripe revenue against spreadsheets after adopting Viktor over Slack. Viktor, given access to Stripe and Google Sheets, posts weekly reports of discrepancies and proposes fixes that the user only approves. The post, a paid partnership, promotes Viktor's cloud browser, code execution, 3,200+ integrations, and memory, with $100 in free credits.

AIThe post argues that SaaS will shift from selling access to selling results, which it calls RaaS. It cites Intercom's Fin, which charges $0.99 per resolved conversation and was acquired by Salesforce for about $3.6B, as evidence. The author also reports that Forrester found per-seat pricing adoption fell from 21% to 15% in twelve months.

AIOpenAI told investors its annualized revenue at the end of September was roughly $18 billion below previously reported figures, CNBC confirmed, and AI-linked stocks fell, with CoreWeave down nearly 8%, Oracle down almost 6%, and Nvidia down 3%. The Nasdaq Composite dropped more than 1%, its worst day since mid-August, as the company faces pressure to justify its valuation ahead of a potential IPO.

AIThe post describes a hard-to-hire profile: someone who can explain technical concepts to non-AI-native engineers, keep products simple, and reject unnecessary features or behaviors. A quoted reply from @adelwu_ calls this combination of trend awareness, creative taste, AI and technical understanding, and execution quality a rare niche that every AI startup wants.

AICredo's latest full-DSP module, Cardinal 802, uses a 4×200G design aimed at both 800G and 1.6T, after the company expanded its ZeroFlap optical module line from 800G to 1.6T over the past year. The company also added Kfir200 silicon photonics PIC from its DustPhotonics acquisition and a PILOT diagnostics platform to the module.

AIGoogle's Playground experiment lets users describe a game in ordinary language and have generative AI build a playable browser-based result that can be revised through further prompts. TechRadar's reviewer built a dragon platformer, Mystic Dragon Glide, from a couple of sentences and a satirical puzzle RPG, Red Tape Hero, from a longer prompt. Playground produced working controls, objectives and music, but the reviewer found the results impressive as prototypes rather than games they would want to play for dozens of hours.

AIOpenAI and Anthropic are hiring former Trump administration officials for senior roles as AI companies work to strengthen ties with Washington. Anthropic has appointed at least two former Trump officials, including Chris Liddell to its board and Sihao Huang as head of frontier compute strategy, while OpenAI has hired at least three, including Thomas Lind to lead cyber and strategic risk on its national security policy team. The moves come as frontier AI is increasingly treated as a national security issue.

AIHy Luu, a 29-year-old Houston engineering consultant who lives with his mother, began margin investing in Tesla in 2021 and at one point amassed more than $100,000 of margin debt. CNBC reports that Robinhood's margin book grew to a record $21.6 billion in the second quarter, up 127% year over year, and total margin debt hit around $1.5 trillion in June per FINRA. Luu says his net worth has risen to more than $800,000, but he stresses that his strategy is risky and does not recommend it.

AIRobert Reich argues that liability law can reduce existential risks from the climate crisis and AI, citing the Suncor v Boulder Supreme Court case in which at least four justices questioned oil companies' claim that the Clean Air Act bars such suits. He points to past settlements, including $206bn from the 1998 tobacco agreement and $20bn from BP after Deepwater Horizon, as precedents, and says AI firms could face similar liability for harms caused by escaping AI agents.

AIThree OpenAI safety researchers, Tomek Korbak, Jasmine Wang, and Mikita Balesni, say they were fired and that their terminations are scaring remaining employees. OpenAI says an investigation found they violated policies on handling sensitive information and denies firing anyone for raising safety concerns, without specifying the breach.

AIAutonomous AI agents break traditional security models because their browser-based activity looks identical to a human user's, and signatures prove identity but not intent. The article says organizations should treat agent policy as a commercial question with a security implementation, and recommends short-lived machine credentials, cryptographic verification via Web Bot Auth, browser-layer intent detection, and defenses against prompt injection.

AIMagnific says its Desktop app is offering unlimited usage until October 15. The post links to a sign-up page but gives no further details on terms or pricing.

AIA CTO hiring new graduates at a larger company reports they are generally unfamiliar with AI coding tools and have little hands-on use of them. Many of those who did internships worked at traditional companies that also did not use these tools, so they are more fluent in pre-AI software development methods than the "AI-native" label suggests.

AITRAE has merged its TraeCode and TraeWork products into a unified new TRAE with an Agent mode and an IDE mode. In hands-on tests, multiple agents handled planning, design, coding, testing, and fixes within one project, with outputs saved in a shared 'My Artifacts' area. The tests also found that agents working in parallel produced conflicting specifications, so someone had to coordinate them.

Why it matters: The hands-on tests show how parallel agents split planning, design, coding, testing, and fixing inside one project, and where their outputs conflicted.