https://www.thinkfacility.com/blog/ai-evaluator-forum-letter-embedded-evaluator-conditions/Think Facilityen2026-09-19T13:14:03ZMore than 100 AI experts say an embedded evaluator shouldn't have other business with the labhttps://www.thinkfacility.com/blog/anthropic-accenture-first-embedded-evaluator-faculty/Think Facilityen2026-09-19T13:08:21ZAnthropic's first embedded evaluator is the consulting firm Accenture, but Anthropic is paying for the workhttps://www.thinkfacility.com/blog/claude-code-reads-agents-md-openai-spec/Think Facilityen2026-09-19T13:35:36ZClaude Code reads AGENTS.md as of version 2.1.277, but only when a project has no CLAUDE.mdhttps://www.thinkfacility.com/blog/meta-muse-ai-agent-arrives-on-mac/Think Facilityen2026-09-19T13:21:14ZMeta's Muse agent now reaches the files and messages on your Mac and acts on themhttps://www.thinkfacility.com/blog/musk-says-ai-doubles-us-gdp-growth-next-year/Think Facilityen2026-09-19T13:39:10ZMusk says AI roughly doubles US growth next year, but the Fed's 2027 median is 2.4%https://www.thinkfacility.com/blog/noam-brown-air-gap-temperature-sensors/Think Facilityen2026-09-18T22:40:00ZCan air-gapped computers talk through CPU heat? What OpenAI's Noam Brown actually saidhttps://www.thinkfacility.com/blog/anthropic-claude-leads-26-percent-of-its-own-ai-research/Think Facilityen2026-09-18T13:10:01ZClaude leads 26% of Anthropic's AI research, but no measured part of it is fully autonomoushttps://www.thinkfacility.com/blog/google-cc-shared-ai-agent-for-families/Think Facilityen2026-09-18T20:18:24ZGoogle's family AI agent has its own Google account, and up to six people share ithttps://www.thinkfacility.com/blog/grok-voice-transcribe-2-speech-to-text-accuracy/Think Facilityen2026-09-18T20:22:21ZGrok's new speech model cut short-phrase word error rate from 20.6% to 6.8%https://www.thinkfacility.com/blog/hacktron-used-claude-opus-to-reach-openais-private-monorepo/Think Facilityen2026-09-18T13:18:10ZOpus 4.8 couldn't write a working exploit, but Opus 5 did it in three hourshttps://www.thinkfacility.com/blog/instinct-concierge-ai-agent-makes-phone-calls/Think Facilityen2026-09-18T20:25:04ZInstinct's agent now makes phone calls, but only the early access group has ithttps://www.thinkfacility.com/blog/newsom-executive-order-ai-kill-switch-california/Think Facilityen2026-09-18T20:10:29ZNewsom gave experts two months to recommend an AI kill switch and auditors inside frontier labshttps://www.thinkfacility.com/blog/noam-brown-says-10000-agents-spent-88-hours-on-navier-stokes/Think Facilityen2026-09-18T13:14:29ZNoam Brown wouldn't give the 10,000-agent swarm 10% of the credit for solving Navier-Stokeshttps://www.thinkfacility.com/blog/spanberger-virginia-data-center-order-ai-task-force/Think Facilityen2026-09-18T20:14:09ZSpanberger wants local approval for data centers over 25 megawatts, and created an AI task forcehttps://www.thinkfacility.com/blog/typesafe-jev-system-one-model-structured-decisions/Think Facilityen2026-09-18T20:16:35ZTypeSafe says its new model can't hallucinate, because it never writes a sentencehttps://www.thinkfacility.com/blog/anthropic-life-sciences-verification-program-claude-biology/Think Facilityen2026-09-17T20:14:10ZAnthropic will stop blocking vetted labs' biology requests in real time and review them offline insteadhttps://www.thinkfacility.com/blog/claude-code-projects-cloud-agents-relaunch/Think Facilityen2026-09-17T20:11:48ZClaude Code projects run parallel cloud agents, but Anthropic says they'll hit usage limits fasterhttps://www.thinkfacility.com/blog/king-charles-ai-summit-dumfries-house-humanity-sacred/Think Facilityen2026-09-17T20:09:01ZKing Charles hosted the AI labs in Scotland and asked for control before it's too latehttps://www.thinkfacility.com/blog/snap-specs-intelligence-anticipatory-ai-iphone-mac/Think Facilityen2026-09-17T20:19:20ZSnap's new AI assistant runs on iPhone and Mac, but the anticipatory part is invite-only