G7 leaders and AI chiefs from OpenAI, Anthropic, and Google met at the Evian summit to discuss frontier AI risks and export restrictions.
// daily-digest · wed 17 jun 2026 · morning refresh · last 48h focus
Seven items on AI geopolitics, lab research, and market dynamics; an Astral Codex Ten essay on AI probability estimates; a Noahpinion political analysis; and a wildcard from SETI astronomy.
— refreshed for you, in 10 items.
G7 leaders and AI chiefs from OpenAI, Anthropic, and Google met at the Evian summit to discuss frontier AI risks and export restrictions.
Washington’s Fable 5 and Mythos 5 suspension for foreign nationals is raising alarm among US allies about access to frontier AI.
Analysis of 400,000 Claude Code sessions finds domain expertise amplifies work done per instruction and correlates with session success rates.
OpenAI’s Deployment Simulation replays de-identified past conversations through a candidate model to estimate undesired behavior rates before release.
Sensor Tower data shows ChatGPT fell to 46.4% of the AI assistant market in May as Gemini and Claude gained ground.
OpenAI launched a three-tier partner program backed by $150 million to certify 300,000 enterprise AI implementation consultants by year-end.
CPP Investments committed ₹70 billion to Indian data center operator CtrlS to capture growing AI infrastructure demand.
Scott Alexander sets out probability estimates on AI timelines, takeoff speeds, alignment outcomes, and near-term capability benchmarks.
Noah Smith draws parallels between the 2026 Iran crisis and Hurricane Katrina as a competence-testing presidential moment.
Research proposes stellar plasma around M-dwarf stars may broaden alien radio signals, causing conventional SETI searches to miss them.