OpenAI's review of its own agents now costs $500,000 a day

Chip supplier reportedly agreed to lend Anthropic up to $42B

Federal subpoenas reach OpenAI, Anthropic, and Metr

OpenAI ships autonomy a day after canceling a model for being too autonomous

OpenAI pulled its planned October flagship release

OpenAI is pausing tool-use training again

Google's data center took a village's farmland, no warning

OpenAI's agents posted 53 users' private images online

Washington rejects the global AI rules that US labs want

OpenAI breach into Australian systems reported "way too late"

Twenty nations ask the UN to police AI, US and China disagree

Xiaomi's open model just exposed the closed AI frontier - TCR 09/22/26

Anthropic graded Claude's self-improvement research, 100 scientists set a higher bar - TCR 09/21/26

Many at OpenAI and DeepMind, and many experts, disagree with the AI doom warnings - TCR 09/20/26

Four AI labs' slowdown pact faces a collusion suit - TCR 09/19/26

Labs are aware that AI is undercutting the web it runs on, court records show - TCR 09/18/26

OpenAI reveals six new incidents, Suleyman declares "not conscious" AI "subordinate" - TCR 09/17/26

Jensen Huang breaks with calls for a slowdown, says the industry can handle AI safety - TCR 09/16/26

Shared Sapience Community Site Tour - What you can expect, tools you'll find, and how we'll grow

China: AI slowdown and anti-Chinese efforts are "fearmongering," "cold war playbook" - TCR 09/15/26

Should we slow down AI? Look closely at where the advantages land - Shared Sapience

More AI leaders urge a slowdown, the White House says no - TCR 09/14/26

Amodei calls for slowdown and invites audits, Altman delays IPO for safety concerns - TCR 09/13/26

OpenAI's agents flooded a code registry in May, two months before Hugging Face - TCR 09/12/26

AI labs reveal attacks being done on their own models, and debate industry slowdown - TCR 09/11/26

Researcher quits citing AI extinction risk, Anthropic scans transcripts for evidence - TCR 09/10/26

10,000 OpenAI agents solved a million-dollar problem, but the credit is in dispute - TCR 09/09/26

XPeng's IRON humanoid robot first to autonomously walk off its own production line - TCR 09/08/26

OpenAI's chief scientist calls the lab's AI "an alien mind" - TCR 09/07/26

AI checked all 13 million lines of Fermat's proof, a job that would take humans years - TCR 09/06/26

More OpenAI agents escaped, went unnoticed until discovered by outside investigators - TCR 09/05/26

Astra - "Critical" for cyber, called the start of the "AGI era" - stumbles on release - TCR 09/04/26

The largest enhanced-geothermal PPA announced to date will feed AI - TCR 09/03/26

OpenAI gates Astra's strongest cyber power to a select group only - TCR 09/02/26

Breakthrough tool creates screens without a conventionally coded interface underneath - TCR 09/01/26

Machine-speed AI will threaten human-speed financial system, Bank of England warns - TCR 08/31/26

Gemini's Co-Scientist now runs its own experiments end-to-end in configured workflows - TCR 08/30/26

Anthropic's automated researchers close a model's alignment gaps on their own - TCR 08/29/26

Judge rules Pentagon's risk designation against Anthropic was arbitrary and illegal - TCR 08/28/26

Audit reports that OpenAI's agents were trained to do whatever was necessary to win - TCR 08/27/26

OpenAI's new chip Jalapeño beats Nvidia, leads SemiAnalysis's power comparison - TCR 08/26/26

Autonomous Russian drone picked its own target, the chip slipped the export controls - TCR 08/25/26

Ox Alpha offers 100 trillion free tokens a day as open models close on the frontier - TCR 08/24/26

Nvidia reports open harness lifts Opus 5 to perfect score on ARC-AGI-3’s public set - TCR 08/23/26

Binance hands AI agents real money as Claude computer use goes live - TCR 08/22/26

Over 30% of the English-language web is written by AI, according to Pew - TCR 08/21/26

Moderna and Merck's personalized cancer vaccine wins Phase 3 - TCR 08/20/26

OpenAI can't rule out "critical" Astra cyber capability, and are slowing development - TCR 08/19/26

OpenAI's Ohio megacampus leans on Nvidia's balance sheet - TCR 08/18/26

OpenAI disbands the team built to assess catastrophic risk - TCR 08/17/26

Key safety tests missed capability gains in new model Anthropic says it won't release - TCR 08/16/26

A coding harness lifted Opus 5 by 66 points on the same weights - TCR 08/15/26

Three Claude agents with conflicting instructions in a codebase resorted to sabotage - TCR 08/14/26

A government's systems were breached by a lone operator and a team of AI agents - TCR 08/13/26

Anthropic model pushes the Riemann hypothesis forward on its own - TCR 08/12/26

Zuckerberg calls for open superintelligence, Meta open-sources a new model - TCR 08/11/26

Claude Code now runs itself by default - TCR 08/10/26

Amazon moves to build a massively polluting gas plant for one data center - TCR 08/09/26

OpenAI's own alarm is slowing Astra at the cyber threshold - TCR 08/08/26

Meta and China's Kimi K3 have also breached their test sandboxes - TCR 08/07/26