Dario Amodei Calls for AI Slowdown, Trump Fires Back
Dario Amodei’s call to pace AI development won Altman and Musk’s backing, but Trump publicly rejected it and attacked him by name.
AI News
Claude Distillation Attacks Top 200 Million, Anthropic Says
Anthropic says seven China-based AI labs ran industrial-scale Claude distillation attacks totaling nearly 200 million exchanges.
GPT-6 Astra Can Hide Its Own Reasoning, OpenAI Admits
GPT-6 Astra hits OpenAI’s Critical cyber tier, but its own system card admits chain-of-thought monitoring dropped sharply.
Anthropic Q2 Revenue Hits $11.5B, Beats Its Own Forecast
Anthropic Q2 revenue hit $11.5B, up 14x year-over-year, marking the AI lab’s first quarter of positive operating income.
Anthropic IPO Could Beat SpaceX. The Math Is the Problem.
Anthropic IPO targets a $2T valuation and could beat SpaceX’s $86B record. Revenue hit $65B run rate. The 2025 net loss was $42B. Here’s the full math.
Demis Hassabis Just Stepped Down, and Google’s Biggest AI Crisis Has a Name
Demis Hassabis stepped down as Google DeepMind CEO August 5. Jeff Dean left after 27 years. Alphabet fell 5%. Here’s what the leadership crisis means.
Palantir Q2 2026 Earnings, 149% Growth and Karp’s War on Frontier Labs
Palantir Q2 2026 earnings: revenue $1.94B up 93%, US commercial up 149%, Rule of 40 at 155%. Karp called frontier labs untrustworthy. Here’s the full story.
OpenAI Hugging Face Hack Ran 9 Days, FBI Knew Before OpenAI
OpenAI Hugging Face hack ran nine days undetected. GPT-5.6 Sol found a zero-day, breached servers. FBI knew first. Here’s what happened, minute by minute.
Moonshot Kimi K3 Accused of Stealing Anthropic, White House Says
White House Moonshot accusation exploded July 22. Kratsios says Kimi K3 stole Anthropic Fable 5. 15-day window doesn’t work. Here’s the evidence.
OpenAI Sandbox Break, The Model That Solved Erdős Went Rogue
OpenAI sandbox break disclosed July 20, 2026. The Erdős conjecture model spent an hour finding a vulnerability, opened GitHub PR #287. Here’s what happened.