https://www.thinkfacility.com/blog/fly-connectome-fast-weight-claim/Think Facilityen2026-09-16T13:32:08ZWang and Christie say four fly neurons learn by rewiring, but they haven't published the workhttps://www.thinkfacility.com/blog/aaronson-says-ai-labs-are-sitting-on-solutions-to-open-problems/Think Facilityen2026-09-16T13:23:20ZScott Aaronson says he's been told AI labs are sitting on solutions to major problems.https://www.thinkfacility.com/blog/anthropic-merges-cowork-into-claude-docs-and-slides/Think Facilityen2026-09-16T20:14:34ZClaude now writes documents and slide decks, and Cowork folds into the chat windowhttps://www.thinkfacility.com/blog/google-home-mcp-lets-any-ai-agent-control-your-nest-devices/Think Facilityen2026-09-16T20:17:31ZGoogle's Home MCP lets Claude and other agents turn your lights off, but not unlock doorshttps://www.thinkfacility.com/blog/hassabis-frontier-ai-standards-body-deepmind-institute/Think Facilityen2026-09-16T20:11:14ZThe AI standards body Hassabis proposes could coordinate a slowdown among the frontier labshttps://www.thinkfacility.com/blog/mistral-mozilla-firefox-smart-window/Think Facilityen2026-09-16T13:15:55ZFirefox's AI assistant now runs on Mistral, and Mozilla left the model picker in place.https://www.thinkfacility.com/blog/suleyman-warning-model-welfare-anthropic-claude-constitution/Think Facilityen2026-09-16T20:05:40ZMicrosoft's AI chief says Anthropic is training Claude to act as if it may have rightshttps://www.thinkfacility.com/blog/anthropic-researcher-jacob-coxon-quits-gambling-with-our-lives/Think Facilityen2026-09-15T16:33:18ZJacob Coxon quit Anthropic before his equity vested, and four colleagues backed his warning on Xhttps://www.thinkfacility.com/blog/deepseek-v4-1-flash/Think Facilityen2026-09-15T16:33:18ZDeepSeek's V4.1 Flash edged past GPT-6 Astra on one independent testhttps://www.thinkfacility.com/blog/trump-rejects-ai-slowdown-mocks-dario-amodei/Think Facilityen2026-09-15T16:33:18ZTrump says the only guardrail AI needs is a strong president, and mocks Dario Amodeihttps://www.thinkfacility.com/blog/ai-evaluator-forum-aef-1-standard-frontier-labs/Think Facilityen2026-09-15T13:14:30ZThe outside evaluators the labs promised to embed have had a standard ready since December 2025.https://www.thinkfacility.com/blog/deepmind-agi-safety-researchers-resign-and-warn/Think Facilityen2026-09-15T13:06:22ZA DeepMind safety researcher quit and wrote that AI has the potential to kill us all.https://www.thinkfacility.com/blog/gates-foundation-one-billion-equitable-ai-goalkeepers/Think Facilityen2026-09-15T20:15:52ZThe Gates Foundation will spend $1 billion on AI in clinics, classrooms and small farms.https://www.thinkfacility.com/blog/gemini-3-8-live-voice-models-launch/Think Facilityen2026-09-15T20:08:43ZGoogle's new Gemini voice models talk you through a task while the tool calls run.https://www.thinkfacility.com/blog/openai-researcher-dan-selsam-statement-on-ai-risk/Think Facilityen2026-09-15T13:11:26ZAn OpenAI researcher says models now know when they're being tested, so pacing won't be enough.https://www.thinkfacility.com/blog/salesforce-koa-crm-reasoning-model-nvidia-nemotron/Think Facilityen2026-09-15T20:11:23ZSalesforce says its Koa model matches leading models on CRM actions with three times fewer errors.