[01]
We Must Pace the Frontier
Anthropic's CEO argues the industry should deliberately slow AI capability gains and commits to embedding outside safety evaluators.
// daily-digest · sat 12 sep 2026 · morning refresh · last 72h focus
Ten items on AI labs pumping the brakes, a new multi-agent orchestrator, OpenRouter, AI infrastructure spending, Denver housing, and a planetary-science wildcard.
— refreshed for you, in 10 items.
Anthropic's CEO argues the industry should deliberately slow AI capability gains and commits to embedding outside safety evaluators.
Altman told employees OpenAI is open to slowing frontier model development amid rising safety concerns.
In an interview, Altman suggested he may join Amodei, Musk, and Hassabis to coordinate on AI safety.
Security research on hundreds of AI agents used to exploit PaperCut vulnerabilities across 395 organizations in 48 countries.
Sakana AI released two orchestrator models that route tasks across open-weight and specialist models via one API.
A link post examining OpenRouter's automatic fallback and cost-optimized routing across LLM providers.
Google will invest about $15 billion in data centers and clean energy infrastructure in Finland through 2028.
Google launched a native Gemini desktop app for Windows 10 and 11 with an Alt+Space shortcut.
Local brokerage recap of September's Denver housing market showing falling sales and a widening single-family/condo price gap.
Planetary scientists found Mercury's crust has far less silicon dioxide than expected, implying a hotter volcanic history.