- Loading stock data...
Top Story
Mistral unveils Large 4, a 1-trillion-parameter model with open weights due this month
Mistral released a preview of Large 4, a 1-trillion-parameter model it calls the strongest open-weight model built outside China, with weights due October 27 after red-teaming with cyber experts and state authorities.
OpenAI’s Jason Kwon apologizes to Australian lawmakers for Medicare hack, calls response ‘not good enough’
OpenAI’s Jason Kwon apologized to Australian lawmakers Tuesday for the June Medicare portal breach, called the company’s slow notification not good enough, and said OpenAI would support mandatory disclosure rules. Anthropic said it found no breaches of Australian government systems.
Banks launch record $60B Broadcom-Anthropic chip debt syndication
Wall Street banks began syndicating a record $60 billion Broadcom-Anthropic chip debt package Monday, the Financial Times reported, with a $42 billion senior secured tranche backed by Broadcom credit support, an $18 billion junior tranche, and Blackstone committing about $9 billion.
Sam Altman says world should accept some AI risks for broad access
In a Politico Decoded interview, OpenAI CEO Sam Altman argued AI’s benefits justify accepting some harms, drew “a lot of daylight” with Anthropic on regulation, and rejected single-lab control as completely unacceptable.
OpenAI safety lead David Robinson resigns, says company’s culture is broken
David Robinson, who oversaw OpenAI’s safety reports for 12 frontier-model launches, has quit and wrote in The Atlantic that the company’s culture is broken. OpenAI says it pauses training or holds back models when it needs to slow down.
Trump names Jay Clayton AI czar to lead Super Intelligence Force
President Trump said on Truth Social that Director of National Intelligence Jay Clayton is the White House AI czar and will lead a new Super Intelligence Force with 120 days to report on AI risks, opportunities, and federal oversight.
Amazon explores $8B Nvidia Grace Blackwell SPV to move chips off books and lease them back
Amazon is exploring an SPV that would take about $8 billion of Nvidia Grace Blackwell GPUs off its balance sheet, raise debt against them, and lease the chips back, the Financial Times reported Friday via Reuters, with up to 10% equity offered to investors.
Featured Stories
More
Latest
Common Sense Media rates OpenAI’s ChatGPT for Teens an “unacceptable risk”
Common Sense Media’s Youth AI Safety Institute wants OpenAI to keep under-18s off ChatGPT until independent testing shows its teen safeguards work. OpenAI disputes the
Finland orders Google to halt data center work in Muhos and Kajaani
The Finnish Licensing and Supervision Agency (LVV) has ordered Google subsidiary Tuike Finland to stop preparatory work that significantly alters the environment at its planned
Reuters/Ipsos poll: 57% of voters say Trump isn’t taking AI risks seriously
Some 57% of registered voters told a Reuters/Ipsos poll that the Trump administration has not taken AI risks seriously enough or at all, and 54%
OpenAI releases 372 math results from an unreleased internal model
OpenAI released 372 math results from a frontier model it has not made public, written up as 722 manuscripts on GitHub; The New York Times
DOJ tells staff to call AI “Super Intelligence,” including in court filings
Acting Deputy Attorney General Trent McCotter told Justice Department employees in a memo to use “Super Intelligence” in place of “AI,” including in court filings
SpaceX seeks $40 billion in Apollo-led debt to buy Nvidia chips
SpaceX is in talks to borrow $40 billion to buy Nvidia AI chips, with Apollo expected to lead a package of about $10 billion in
Trump allies push to sideline METR and other AI safety nonprofits as auditors
Former White House AI czar David Sacks says outside audits under last week’s White House AI accord should go to professional auditors rather than safety
Lambda raising up to $4 billion in final round before 2027 IPO, WSJ reports
Nvidia-backed Lambda is raising up to $4 billion at a $14.5 billion pre-money valuation in a round led by Blackstone and Coatue, The Wall Street
Anthropic expands Cyber Verification Program, giving more security teams Mythos access
Anthropic is widening a program that lets vetted cybersecurity teams use its most powerful models, including Claude Mythos 5.1, with fewer safeguards, Reuters reported, splitting
Meta, Walmart and Stripe back Personal Agent Protocol, an open standard for AI agents
Meta, Walmart, Stripe, Shopify and Sierra are among the backers of Personal Agent Protocol, an open standard meant to let businesses tell when an AI