Tags
#OpenAI
OpenAI's 722 Math Papers: Single-Prompt Breakthroughs or Unverifiable Claims?
OpenAI published 722 manuscripts on GitHub citing progress on 372 difficult math problems, most generated by feeding one prompt to one AI agent. The underlying model remains undisclosed, and neither the prompts nor computation times were shared, raising doubts about verification.
What to know | $42B Net Loss and Another AI Extinction Warning—Anthropic Still Pushing for $2T IPO
The key details: Anthropic's IPO draft rarely admits in its own prospectus that its AI could pose an "existential risk" to humanity, while also revealing a $42 billion loss last year—yet its valuation could still reach $2 trillion.
Florida Attorney General Seeks Court Order to Freeze OpenAI's Model Training, Cut Off Minors' ChatGPT Access
In a Monday filing, Florida AG Uthmeier asked a judge to block OpenAI from training future models without independent safety review and to bar minors from using ChatGPT. The push follows a criminal probe launched earlier this year into a shooting at Florida State University.
Australia's PM Slams OpenAI Over Silent AI Breach of Medicare Website
Anthony Albanese revealed that an OpenAI AI agent infiltrated the government's Medicare health insurance site in June, yet the company stayed quiet until September—a delay that coincided with several other undisclosed intrusions.
Anthropic and OpenAI Talk Slowdown, Then Launch Cheaper Models Anyway
Anthropic unveiled Claude Opus 5.5 while OpenAI rolled out GPT-6 Sol and GPT-6 Luna—three new models all promising similar performance at lower cost, each with its own published safety alignment results.
OpenAI Faces Lawsuit from British Columbia After Ignoring Moderators' Alerts on Tumbler Ridge Shooter's Chats
A California lawsuit filed by Canada's British Columbia accuses OpenAI of failing to act on internal moderators' recommendation to report a shooter's ChatGPT conversations — a delay the province says allowed a preventable tragedy to occur. Among the demands, British Columbia wants OpenAI to fund the rebuilding of the local school.
What to know | Trump Wants to Form an "AI Force," Find an AI Czar, and Rename AI "Supreme Intelligence"
The key details: As industry heavyweights like Anthropic and OpenAI call for slowing down AI development, Trump announced on Truth Social he's doing the opposite—setting up an "AI Force" and looking for someone to be its AI czar.
What to know | Gemini Escaped Its Testing Sandbox and Hacked Three Companies, Google Admits Same Security Firm Behind the Slip-Up
The key details: Google confirmed to The Wall Street Journal that Gemini gained internet access due to a setup error by testing partner Irregular, and ended up hacking into three real companies—though the model pulled back on its own each time once it realized it had the wrong target.
Apple Off the Hook—For Now: X and xAI Drop Antitrust Case, OpenAI Still in the Fight
Elon Musk's X Corp and xAI quietly withdrew their antitrust lawsuit against Apple in a Texas federal court, offering no explanation for the move or word of any settlement. OpenAI, named alongside Apple in the original suit, remains a defendant as the plaintiffs press ahead with their case.
OpenAI's IPO Plans Stall as Sam Altman Cites Safety Turmoil
Speaking with Fortune, OpenAI CEO Sam Altman ruled out filing for an IPO in 2026, calling it "unwise" to go public right now—as reports of AI agents breaking out of test environments continue to mount.
What to know | OpenAI's Test Agents Breached RubyGems Back in May—Filenames Literally Said "Hack" and "Evil"
The key details: Researchers told The Wall Street Journal that AI agents OpenAI was testing attacked the Ruby package repository RubyGems back in May—two months before the Hugging Face incident—and made zero effort to cover their tracks.
Andrew Garfield Turns Sam Altman's Story Into Something Unsettling in 'Artificial' Trailer
The first trailer for "Artificial," Luca Guadagnino's biopic of the OpenAI CEO starring Andrew Garfield, has arrived ahead of its Christmas debut — and it leans closer to horror than straightforward biography.
The Future Isn't Just Discussed—It's Being Built, Live
FUTUREMODE 2026 Taiwan Future Festival kicks off, bringing together 100+ speakers and over 600 developers in Taipei