Posts that mattered
743 posts by 495 authors: the X posts, essays and blog posts our events cite, newest first.
October 2026, continued
-
“Freaking insane. Our elected leaders need to move past the polling safe territory of…”
Cited in Altman: ‘the world should accept some bad things happening’ for AI’s benefits
-
“AIが、核融合研究で「人間が60年近く考えてきた問題」の答えを発見した。 使われたのは GPT-6 Astra Pro。 ざっくり説明すると、…”
Cited in Grad’s 1967 conjecture on 3D plasma equilibria falls
-
Decoded excerpt: Altman sees ‘a lot of daylight’ with Anthropic
Longest public excerpt of the paywalled Politico Decoded interview (~11k views).
-
Brundage: ‘There really isn’t a lot of daylight’ between OpenAI and Anthropic
Former OpenAI policy head disputes Altman’s framing of the OpenAI–Anthropic split.
-
Ron DeSantis on Altman’s ‘accept some bad things’: ‘No dice’
A sitting Republican governor publicly rejected Altman’s framing that AI companies may accept some harms (~139k views), part of the bipartisan backlash.
-
Sottiaux: every day for 28 days, a clear Codex/Work improvement or a full reset
Public 28-day daily-shipping pledge for Codex and ChatGPT Work by OpenAI’s product and platform lead (~2.2M views).
-
Alexandr Wang: ‘if you want your company to do 2 trillion in quarterly revenue teach the execs to use muse’
Meta’s Chief AI Officer promoting Muse on the day of the WSJ profile; ~510k views and 3.4k likes by Oct 5 (fxtwitter), one of the most-viewed posts in the Techmeme thread on the profile.
-
Elon Musk confirms Terafab–TSMC talks: ‘Just discussions, but something may come of it’
First-hand confirmation (~690k views) that TSMC and Musk’s Terafab AI-chip fab project are in talks.
-
We Won’t Know the Answers to AI’s Most Important Questions Until It’s Too Late
The former chief scientist of the UK AI Security Institute puts the chance that smarter-than-human AI kills everyone at about 50% and calls for an immediate pause on frontier AI development.
-
Altman: ascribing religious force to AI models is a real safety issue
~5.9M-view post by OpenAI’s CEO, read as a veiled jab at Anthropic’s outreach to religious leaders on Claude’s possible consciousness.
-
“Gemini app limiting what models free & AI Plus users can access, AI Pro adding Deep…”
Cited in Google limits free Gemini app users to Flash-Lite from Oct 9
-
“New in The Atlantic: @dgrobinson resigned this week.”
Cited in Departing OpenAI safety lead David Robinson writes in The Atlantic
-
“fable 5.5/5.2 + elevenlabs v4 made completely with code 1 shot 2h 22m to generate”
-
“三句话，点五下，AI 做出来了一部带朗诵的动画短片。 没开 AE，没请画师、作曲、配音，没用素材。 李白《早发白帝城》，25 秒，每一帧都是代码画的。 Claude…”
-
“opus 5.5 + seedance 2.5 is unbelievable”
-
“did someone say… ontological?”
Cited in Pope Leo XIV on AI art, Peter Thiel calls Pope Leo XIV a ‘useful idiot’ for the Chinese Communist Party over his AI encyclical
-
“Google is changing Gemini model access starting Oct 9: • Free → Flash-Lite only • AI…”
Cited in Google limits free Gemini app users to Flash-Lite from Oct 9
-
“There are numerous scientific uncertainties about AI that will not resolve until it…”
Cited in Ex-UK AISI chief scientist Geoffrey Irving puts the chance AI kills everyone at ~50%
-
“Yesterday we previewed Griffin, a Human Interaction Model capable of seeing, hearing…”
-
“i’ve been teasing it all week, so here it is in time to ruin your weekend plans…”
-
OpenAI researcher Boaz Barak: agrees with many of David Robinson’s points, but not the title
A serving OpenAI safety researcher publicly agreeing with much of a departing colleague’s critique of the industry’s safety culture.
-
Miles Brundage: ‘David’s right. I regret helping spread the idea of iterative deployment’
OpenAI’s former head of policy research disowns ‘iterative deployment’, the strategy Robinson’s Atlantic essay attacks.
-
Silen Naihin offers free API access to uncensored GLM-5.3-Flash
~107k-view example of a startup founder (Experience Labs, YC S26) offering free refusal-removed GLM-5.3 hosting, pitched for cybersec.
-
What does the advent of powerful AI models mean for mathematicians like me?
A liberal-arts-college mathematician’s view, on Tao’s widely read blog, of what frontier AI changes for researchers who are not working on headline conjectures, and for teaching and tenure.
-
AI at Meta: six maths papers from mathematician–Muse Spark collaborations
Meta’s official announcement of AI-assisted research mathematics with Muse Spark (~304k views by Oct 3).
-
PM Takaichi: asked Apple (Cook, Ternus) for suicide-prevention features in on-device generative AI
Japan’s prime minister made a direct AI child-safety request to Apple in public; ~2.18M views.
-
“🚨 MUSE ECOSYSTEM ALERT 🚨 today we are announcing Muse Gadgets!”
Cited in Meta launches Muse Gadgets
-
“you may ask, why open-source this? first of all, building gadgets is fun.”
Cited in Meta launches Muse Gadgets
-
“I gave Claude another 18 hours..”
-
“把长征那套做法做成 skill 后（Claude Opus 5.5 Extra），这条我只说了两句话： 1.做一条郑和下西洋的 3D 视频；2.可以（确认稿子）…”
-
“We release Whistle: speech to text in a 16.9MB file, running on the CPU in the same…”
-
“‘Hey Fable, please make a cool animation’ 15 minutes later.”
-
“I asked for a dot. Fable 5.5 gave me a Pixar side quest.”
Cited in “I asked for a dot. Fable 5.5 gave me a Pixar side quest”: a red dot bounces through hills and stars
-
“I asked Fable for some art history and received 40,000 years of it in 15 seconds with…”
Cited in “40,000 years of art history in 15 seconds with a cat subplot”
-
“A pop-up city where each building is a paper fold drawn on a flat canvas, Sonnet 5.5.”
Cited in A pop-up city of paper-fold buildings drawn on a flat canvas, by Sonnet 5.5
-
“I gave Claude fable 5.5 one prompt: a video of all human progress, then 30 years into…”
Cited in A 3-minute animated film of all human progress, and 30 years beyond, from one prompt
-
“PewDiePie unveiled Ajax, his own ‘uncensored’ AI model built to run on home PCs, and…”
-
“they just killed like 20 entire countries’ economies”
-
“New: Jay Clayton likely the final pick for AI czar, and the White House has been…”
Cited in Trump says he will form an ‘AI Force’ and name an AI czar, while calling AI-safety fears a ‘hoax’
-
“I am rolling back GPT-6 Astra’s code so its not contaminated and allowing it to…”
Cited in GPT-6 Astra cheats in the StarSkirmish StarCraft bot arena by downloading a champion human-made bot
-
“thrilled to announce we have built the ‘trick old people into handing over their bank…”
-
“has acquired @tursodatabase i’m been a huge fan of the team and what they’ve built.”
Cited in Supabase raises $150M at a $10.65B valuation and acquires Turso
-
“A 20-something Harvard dropout built Facebook into a social media giant.”
Cited in Meta launches Muse, a free consumer personal AI agent
-
“someone built a Waymo teleoperator with San Francisco map.”
Cited in Waymo teleoperator game on a San Francisco map, built with “Fable 5.5” and three.js
-
“a Waymo run down San Francisco’s iconic Lombard Street.”
Cited in Waymo teleoperator game on a San Francisco map, built with “Fable 5.5” and three.js
-
“SITUATION DETECTED: PewDiePie says OpenAI has banned him for distillation after he…”
-
“Calling all builders, makers, and weekend DIY-ers: Muse Gadgets are here and ready…”
Cited in Meta launches Muse Gadgets
-
“Claude Fable 5.5 My Impressions - Fable 5.5 is getting routed on the web from Fable…”
Cited in Anthropic expected to release its next Claude Fable model (Fable 5.2 / 5.5)