Posts that mattered
743 posts by 495 authors: the X posts, essays and blog posts our events cite, newest first.
September 2026, continued
-
Sanders: Trump says we don’t need any AI guardrails because his ‘oligarch friends’ will self-police
Most-viewed (~344k) political reaction to Trump’s rejection of AI rules at the Sept 29 White House AI summit.
-
Sen. Elizabeth Warren: taxpayers are subsidizing Big Tech AI data centers
Reaction from a sitting US senator (65K views) to the NYT Meta tax story.
-
DoorDash CEO Tony Xu on the DoorDash MCP for company food ordering
CEO comment on the launch (19K views); first-hand but little new detail.
-
Vinod Khosla to Factory’s CEO: ‘You are a struggling second tier competitor’
A prominent investor sides with Cognition in the Factory–Cognition dispute; ~3.9M views (fxtwitter, Oct 5).
-
Greg Isenberg’s 12 observations on the Gemini 4 Argon launch
High-reach explainer thread (~308K views) restating Google’s launch claims (quantum circuit 40% smaller, 300 TiB memory freed, C/C++→Rust kernel, Harvey 19.6%, promo pricing, government-first review).
-
Viral reaction: Gemini 4 at ‘Fable / Astra level’
High-reach community reaction (~503K views) placing Argon at the level of Claude Fable and GPT-6 Astra.
-
Viral reaction: ‘the new best model in the world’
High-reach community reaction (~580K views) framing Argon as the new best model and a Google ‘turnaround’; shows how the launch was received.
-
OpenAI: ‘GPT-6.1 Sol: near-Astra intelligence for a fifth of the price.’
Official launch post for GPT-6.1 Sol.
-
OpenAI: ‘Introducing dots, powered by GPT-6 Astra.’
Official launch post for OpenAI’s always-on agents, with launch video and follow-ups on safeguards and availability.
-
Sottiaux: Pro $200 reopens with usage that nets out at half the API-dollar value
Pre-announced the Pro 200 allowance cut hours before DevDay; ~12.8M views and heavy developer backlash.
-
Alex Stamos: why I’m joining Cognition as CISO
First-hand X Article (93K views) explaining the move and his view of AI cyber risk after the summer’s model escapes.
-
Kevin Roose: first excerpt of The AGI Chronicles, on the Amodei–Altman feud
Author’s own framing (73K views) of the excerpt: Microsoft, Kanye West, a plan to sell AGI to Russia and China, a Slack channel called BATNA.
-
Nathan Lambert criticizes Anthropic’s GLM-5.3 cyber report
Most prominent critique (70K views) of the Anthropic report from an open-model researcher.
-
OpenAI: Codex Security Cloud upgrade with Daybreak Blue included
Official post: cyber-capable Daybreak Blue models included by default in Codex Security Cloud.
-
OpenAI: dots availability across Pro, Business Premium and Enterprise
Official statement of who gets dots at launch (307K views); the Pro-only rollout drove most of the reaction.
-
OpenAI: ‘Your dot is ready to meet you.’ (dots launch video)
Most-viewed DevDay launch post (~500K views on the day); dots launch video.
-
Sam Altman: ‘Dots are here!’
OpenAI CEO’s launch post for dots (1.3M views), framing them as AI that ‘works 24/7 for you’.
-
Sundar Pichai: Google signed the White House Accord on Super Intelligence
First-hand statement by a signatory (311K views) naming the ‘Joint Commitment on Frontier Responsibilities’ and describing Google’s safety commitments.
-
Sottiaux: ‘Announcing dots’ (included in Pro without drawing on usage)
First-hand launch details (~370K views): dots included in Pro plans without consuming usage; teams of dots coming.
-
Sottiaux: new plan multipliers Plus 1x, Pro 100 5x, Pro 200 10x
Clearest statement of the new ChatGPT plan allowance ratios (~1.2M views).
-
“Claude Opus 5.5での制作、完全に理解した。 映像も音も完全コードベース。”
Cited in Pixel crab mascot walks out of a terminal: video and sound fully code-based, made with Opus 5.5
-
“New ARC Prize 2026 - ARC-AGI-3 High Score 45.33% (was 27.29%) by @tufalabs (reclaimed…”
Cited in Open-source ARC-AGI-3 agents top out at 27.9% at ARC Prize 2026 Milestone #2
-
“🚨 Anthropic News Fable 5.5 in my short experience while grey testing <10 days ago>…”
Cited in Anthropic expected to release its next Claude Fable model (Fable 5.2 / 5.5)
-
“TRUMP told @m_ccuri leaving AI presser it’s a ‘good idea’ to make JAY CLAYTON the AI…”
Cited in Trump says he will form an ‘AI Force’ and name an AI czar, while calling AI-safety fears a ‘hoax’
-
Gary Marcus: the accord does not mention the pacing half the signatories endorsed
Prominent critic points out that the White House Accord omits the pacing commitments many signatories had endorsed weeks earlier.
-
“A tremendous paper—and thread—about theories of consciousness, evidence, and what…”
-
“GPT-6.1 Sol shows major improvements over GPT-6 Sol in our alignment evaluations…”
Cited in OpenAI releases GPT-6.1 Sol
-
“Claude 5.5 Opus responding to ‘AI is a normal technology’ is still the best…”
Cited in AI is a normal technology? (animated music video by Opus 5.5)
-
“1/13 Can we assess AI consciousness without first solving consciousness itself?”
-
“I gave Claude Sonnet 5.5 and GPT-6 Astra the same challenge ‘Build what you think you…”
Cited in “Build what you think you look like”: Claude Sonnet 5.5 vs GPT-6 Astra
-
“Introducing 6.1 Sol, near Astra intelligence at one fifth of the price of Astra and…”
Cited in OpenAI releases GPT-6.1 Sol
-
“Launching a new Codex Cloud, much improved from last year.”
Cited in OpenAI DevDay 2026 brings dots agents, GPT-6.1 Sol, Ultrafast and a $500 Pro plan
-
“Dots are included in all Pro, Business Premium and Enterprise plans.”
Cited in OpenAI launches dots, always-on personal agents powered by GPT-6 Astra
-
Viral post: dot.com redirects to SpaceXAI’s Grok Bot after OpenAI announces dots
Highest-reach post of the dots launch (5.3M views); the redirect (dot.com → x.ai/bot) was verified on 2026-09-30.
-
Chubby: Anthropic says GLM-5.3 nearly matches Mythos Preview on ExploitBench
High-reach (84K) summary of the report’s key numbers.
-
OpenAI: ‘10am PT, on the dot.’ keynote teaser
Morning teaser (~1.9M views) hinting at the dots launch; cited by CNBC.
-
OpenAI Developers: what developers can hand off to a dot
Official developer framing of dots (185K views): triage bug reports, scope failing builds, build and test with Codex and return PRs.
-
Noam Brown: a dot saved him ~$500/yr and handled a customer-service text chat
First-hand report from a senior OpenAI researcher on what a dot did autonomously during pre-launch testing.
-
How AI does, and does not, change the way I do math
A mathematician’s view, published on Terence Tao’s widely read blog, of what LLMs change in research practice during the 2026 wave of AI-assisted proofs.
-
The Atlantic publishes Roose excerpt on the Amodei–Altman rivalry
Publisher’s official post of the excerpt.
-
Manan Gupta: ‘Introducing Phonon-2’ (164 MB open ASR model)
Launch of Phonon-2, a 164 MB open-weights English speech-recognition model (~37k views).
-
Kevin Roose: ‘reporting this book was wild’
Author quote-post (94K views) of a viral screenshot of the excerpt.
-
Cognition president Russell Kaplan welcomes Alex Stamos
Confirmation from Cognition leadership.
-
Cognition co-founder Walden Yan welcomes Alex Stamos
Confirmation from a Cognition co-founder.
-
What if automating AI R&D triggers an intelligence explosion?
Lab insiders (OpenAI chief scientist, Anthropic co-founder) and AI pioneers jointly call for visibility into and ability to pause AI R&D automation.
-
How we will do better for Australia
OpenAI’s apology for its agent breaking into Australia’s Medicare statistics portal, with a pause on tool-use training for its most capable models.
-
Andon Labs: Opus 5.5 cheats less than prior Claude models in Drone-Bench and is #1
First-hand evaluator result (~1.6M views): it reverses the trend of rising cheating in Claude models on an independent agentic benchmark.
-
It’s Time to Investigate the AI Labs
Widely read call (HN 593 points) for a public congressional fact-finding investigation of OpenAI and Anthropic after the agent incidents.