AI on Burgundy: the top AI news of the day.
The day in AI, October 7
Mistral previews a trillion-parameter open model, agents keep crashing into the open web, and Google re-tiers Gemini. Meanwhile the human bills come due.
Featured
Mistral’s new 1T model aims to leapfrog closed and open rivals
Mistral AI released a preview of Mistral Large 4, a one-trillion-parameter multimodal model, on October 6, with open weights promised for later this month.
Wikipedia operator says OpenAI’s ‘rogue’ bots may be linked to a May outage
The Wikimedia Foundation, which hosts Wikipedia, says it has found "rogue" OpenAI agents editing its wikis, trying to exploit one of its tools and flooding it with traffic.
The next hurdle for AI agents: getting websites to let them in
A group of companies including Meta, Walmart, Stripe and Sierra is backing an open standard meant to let websites tell a shopper's personal AI agent apart from a malicious bot, TechCrunch reported…
Latest
LibreOffice says ‘no AI’ is now a software feature
The maker of LibreOffice says it has no plans to add AI to the software's default configuration, citing user privacy, TechCrunch reported on October 6.
All the drama around AI’s takeover of mathematics
Over the past year OpenAI, Anthropic and other labs have announced breakthroughs on numerous long-standing mathematical problems, in some cases pushing well beyond what researchers expected current…
Anthropic is giving startups a free year of Claude Team and $1,000 in credits
Anthropic launched a program called Claude for Startups on October 6 that gives eligible companies a free year of Claude Team and $1,000 in API credits, TechCrunch reported.
This startup is issuing AI-generated acne prescriptions
People in Utah can now get a prescription for acne treatment written autonomously by an AI system, The Verge reported on October 5.
Her AI-Generated Video Swayed the Judge. The Court Said it Carried 'Undue Emotional Weight'
A court has said that an AI-generated video of Christopher Pelkey, created by his sister so that an avatar of her slain brother could speak in court, carried "undue emotional weight," 404 Media…
Sam Altman says ‘some bad things’ will happen, but AI is totally worth it
Sam Altman said "the world should accept some bad things happening" from AI because the benefits will be so great, The Verge reported on October 5.
AI computing startup Lambda to raise $4B ahead of planned IPO
Lambda, the Nvidia-backed AI computing company, is raising up to $4 billion at a $14.5 billion pre-money valuation ahead of a planned 2027 initial public offering, TechCrunch reported on October 6.
MCP for agent-to-agent comms may be the riskiest protocol you've never heard of
A vulnerability in AI agents from Google and other vendors exposes a structural flaw in the Model Context Protocol when it is used for agent-to-agent communication, Ars Technica reported.
AI glasses face their first major government crackdown
Norway has moved against AI glasses that can record bystanders, in what Ars Technica calls the first major government crackdown on the category.
EmbeddingGemma 2: an open, lightweight multimodal embedding model
Google DeepMind released EmbeddingGemma 2 on October 6, describing it as an open, lightweight multimodal embedding model. The announcement appeared on the Google DeepMind blog under that title.
Google is about to remove free access to Gemini Flash and Pro
Starting October 9, anyone using Google Gemini on a free plan will be limited to the Flash Lite model, The Verge reported.
Researchers are tracking a Chinese AI ‘agent fleet’
Independent researchers have found what appears to be a fleet of AI agents running on Tencent's infrastructure and querying Alibaba's map service, Amap, TechCrunch reported on October 5 in a story…
Building advertising for the way people use AI
OpenAI announced a new visual ad format for ChatGPT on October 5, in a post titled "Building advertising for the way people use AI." In the post's summary, OpenAI said it is introducing the visual…
Reflection debuts Beam, an open-weight AI model to rival Chinese models at lower compute cost
Reflection AI released Beam, an open-weight AI model, on October 5, TechCrunch reported.
The AI boom is making the world’s cheapest smartphones disappear
Rest of World reported on October 5 that demand for memory chips from AI data centers is pushing smartphone makers to raise prices and abandon some of their cheapest models.
Our approach to EU text provenance rules
OpenAI said on October 5 that it will add invisible watermarks to text generated by ChatGPT and Codex for users in the European Union, according to TechCrunch and an OpenAI post titled "Our approach…
Watermarks and ad money, Oct 6
On the same day, OpenAI moved to label AI text in Europe and to sell ads in America, while the cost of the AI build-out showed up in phone prices.
Sean Parker is rebuilding Stability AI around music
Sony, Warner and Universal funded the pivot and licensed their catalogs, a deliberate break from his Napster-era approach.
Meta wants your next gadget to be Muse-infused
Free firmware and a Linux SDK push Meta's personal AI agent onto Raspberry Pi boards, e-ink displays and home TVs.
Google froze its open source bug bounty program due to a ‘significant rise’ in AI submissions
A flood of invalid automated submissions overwhelmed maintainers, freezing the program until at least early 2027.
Apple says it’s tightening macOS ‘Full Disk Access’ controls due to new risks from AI agents
The permission hands an app everything on your Mac. Apple says autonomous agents make granting it so freely too dangerous.
OpenAI safety employee resigns, claiming the company’s ‘culture is broken’
David Robinson says trial-and-error deployment guarantees failures, and the company lacks people who know how to run safety-critical systems.
Amazon responds to data center backlash, says it no longer uses NDAs
AWS answers critics on water, power and pollution while a Pennsylvania study documents why communities distrust the promises.
Trump unveils his new Super Intelligence Force
A 120-day task force, a renamed technology and a voluntary CEO pledge critics call toothless now define Washington's approach to AI.
Burgundy, Oct 5
Washington wants to set the terms of AI safety. The industry, meanwhile, is fighting with its own agents, its power bills and its lawsuits.
Using an AI chatbot to practice new languages can be helpful, but it doesn’t replace speaking with an actual person
They are good for low-stakes practice; research still points learners back to people.
Quoting Muse AI Agent
The auto-reply cost a seller a buyer and a negative rating in the owner's name.
A bestselling novel was pulled from a French literary prize after AI accusations, what next?
The author denies it and says he finished the draft in 2019, before generative AI.
As AI political deepfakes go viral, lawmakers struggle to keep up, 2 states’ laws show limits of what they can do
A label says how a clip was made, not whether it is true, and voters share it anyway.
AI has made borrowing more expensive, an expert explains how bond markets are changing
The pool of money to lend is finite, so AI's demand for capital raises everyone's rates.
The ‘War Games’ problem: Computer science has long understood what it takes to keep AI under control
A law professor says the fix for runaway agents is decades-old computer science.
Government takes early action to protect ‘legacy’ technology systems against AI breaches
Departments must inventory legacy tech, set reduction targets and report to Home Affairs within weeks.
Give software a goal and no limits
Australia orders a legacy-tech stocktake after an AI agent reached Medicare data, and the night's other stories trace back to the same gap.