All AI news

AIDario Amodei has written an essay arguing that the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first step by giving third-party evaluators permanent, employee-level access to its systems. The evaluators can verify adherence to safety measures, report incidents, and assess model alignment during training.

AIVinoo Ganesh, co-founder of Kepler and former leader of Spark at Palantir, where he built Project Frontline, walks through best practices for forward deployed engineers (FDEs). The discussion appears in a Latent Space podcast episode.

AIAlberto Romero argues that AI solving Millennium Prize problems in 2026 threatens mathematics through success, not failure. He draws on Terence Tao's view that struggle shapes mathematicians, and that proof abundance without hard problems could leave fields depleted, like overplanted farmland.

AILogan Kilpatrick, who is associated with Google and Gemini, reflects that AI progress is a long road marked by setbacks, unexpected challenges, and ups and downs. He encourages people to take time to reflect, learn, and keep pushing forward.

AIDario Amodei announced a new essay arguing the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first step by giving third-party evaluators permanent, employee-level access to its systems to verify safety measures, report incidents, and assess model alignment during training.

AIJohn Schulman thanked Dwarkesh Patel, Beren Millidge, and Charlie O'Neill for a conversation about the AI frontier. The episode covers topics including the case against recursive self-improvement, the drivers of Chinese labs' progress, and whether long-horizon RL could elicit AGI.

AIMira Murati highlighted a conversation about where AI is heading and the work humans still need to do. Quoted context from Thinking Machines says John Schulman, in a talk with Dwarkesh, argues human judgment still matters for teaching models messy real-world tasks, applying taste, and specifying what people actually want.

AIThinking Machines shared a Dwarkesh Patel podcast episode with John Schulman discussing where human judgment remains essential as models improve and self-improve. Schulman highlights teaching models to handle messy real-world tasks, applying taste to what works in the long run, and specifying what people actually want. The episode also covers recursive self-improvement, long-horizon RL, and the sim-to-real gap.

AIMckay Wrigley dismissed concerns that AI curing cancer could disrupt mathematicians' "process of understanding" and raise attribution questions, calling the objection one of the dumbest things he has read. He argued that building AI to solve humanity's greatest problems would be a miraculous achievement. The reply responds to a quoted post noting that 25 Fields Medal winners issued a joint declaration warning of severe misalignment between AI companies and the mathematics community.

AIAndrew Ng argues that skilled AI engineers increasingly drive the build loop and make product decisions rather than merely implementing specifications from product managers and designers. He identifies four key skills for shaping the build: driving the build loop, making product decisions, communicating and leading, and high-agency ownership.

AIJohn Schulman, Beren Millidge, and Charlie O'Neill discuss whether current training methods could produce recursive self-improvement. They argue that progress depends on whether models can learn their own objectives and on sample efficiency, and that distillation keeps frontier capabilities from centralizing quickly.

AIA post from hardmaru highlights virtual fruit fly brains as the latest trend in AI. A quoted post from @mattyhempstead describes wireheading a fly by artificially enhancing its dopamine neurons to maximize enjoyment, with the stated goal of creating a fly happier than all other flies combined.

AINathan Lambert has compiled a reading list of open-model writing covering why labs release open weights, the open-versus-closed debate, and US-China competition, last updated 15 September 2026. The list includes pieces on open-model economics, safety and marginal-risk research, and recent Chinese releases such as Kimi K3 and GLM-5.2. It also cites lawmaker inquiries into Western companies' use of Chinese models.

AIJunyang Lin warns that adding complexity to model architecture can cause unknown issues that existing evals, even internal ones, cannot detect. He notes that teams may still accept these small unknown costs for large efficiency benefits as foundation models increasingly target agentic tasks.

AIA former Anthropic researcher's resignation post and a senior Anthropic alignment leader's comments that AI could kill all humans drew wide attention. The column argues public and congressional concern about superintelligence risk is growing, citing the Ban Artificial Superintelligence Act and a Senate probe into an OpenAI-related incident.

AIClare Liguori published a manifesto with 10 principles for developers who use AI coding tools but feel they are not shipping much faster. Werner Vogels, owner of this account, called it mandatory reading. The principles are drawn from practices she has seen work across Amazon teams.

AIRedwood Research argues that AI companies should regularly report whether their architectures allow latent reasoning or latent communication between agents, and that such reporting should be externally verified. It proposes opaque serial depth as a minimally invasive proxy, with third-party evaluators reviewing near-frontier models, including internal R&D prototypes. The post also calls for published monitorability policies and stress tests on chain-of-thought monitoring.

AIA developer, Sainimatic, is building a simulation where a fruit fly can fly freely forever, as a refuge from other projects that subject the fly to painful, nightmarish environments. The main post itself says only that the author needed a break in fruit fly heaven, so the simulation details are drawn from the quoted background post.

AIJan Leike, an Anthropic researcher, argues there is still time to change the rules of the AI scaling race, but perhaps not much. He says AI will keep improving rapidly and the stakes will rise, and without pacing mechanisms applying to everyone, companies are incentivized to accelerate or risk commercial failure.

AIJan Leike, who the account identifies as an Anthropic researcher, says AI's benefits will be massive but will take many years to materialize. He argues for more patience with these developments so the associated risks can be reduced.

AIJan Leike, who works at Anthropic, says AI company leaders have pushed back against regulation, except Anthropic. He argues regulation serves their interest because accidents will trigger inevitable backlash that could be rushed, excessive, and stifle AI's benefits.

AIJan Leike, who works at Anthropic, points to a statement signed by 1,386 employees of frontier AI companies, including six chief scientists. The statement asks for an option to pace AI development.

AIAnthropic's Jan Leike argues that many effective safety and alignment interventions take substantial time and care to implement. He cites jailbreaking mitigations for Opus 4, which took over a year to develop and could not have been built once the model required them.

AIJan Leike says effective AI regulation is needed and calls the RAISE Act and California's SB 53 a good start. He argues they fail to keep pace with the speed of development, mainly requesting transparency and self-commitments. He says he has personally donated to political candidates from both parties who support such regulation.

AIAnthropic's Jan Leike argues that now is a good time to build institutional mechanisms to pace frontier AI development. He says the industry is locked in an all-out scaling race toward superintelligence, and more time may be needed for safety and alignment mitigations.

AIJohn Schulman argues that training on user data contributes little to frontier math gains, which come mainly from scaling pretraining and RLVR. He says user data is more likely used to find failure modes that hired annotators struggle to recreate. He calls for stronger norms on disclosing how companies train on user data, including the methods and capabilities targeted.

AINathan Lambert argues that a resignation post by AI researcher Jacob Coxon spread widely because public fear of AI extinction risk had been building. He says concrete risks such as cyber attacks and bio-risks deserve debate, while he assigns extinction risk a probability too low to discuss and expects recursive self-improvement to produce only lossy, jagged gains rather than a rapid takeoff.

AICohere CEO Aidan Gomez posted that "Encoders are back," signaling renewed interest in encoder architectures. The post was a short reaction to an image or discussion that scaling01 described as an "alien architecture," with no further technical details given.

AIThomas Dohmke says he will buy the new iPhone Duo but argues it cannot become his intelligent personal hub without an agent that can use the same apps he does. He claims, like the iPad, the hardware is held back by an operating system that treats software as operable only by humans. He concludes that agents using computers is the new paradigm.

AIFormer OpenAI and Anthropic employee Jacob Coxon resigned and posted a viral tweet, which has gathered over 700k likes and 140 million views, accusing AI companies of gambling with our lives. Coxon said people building AI earnestly believe it could kill us all by the end of the decade. The article argues that more insiders may leave, leaving the industry's remaining staff to accelerate development.

AIDwarkesh Patel says Paul Christiano has perhaps the best original track record in the world for predicting the timeline and nature of AI takeoff. The post links to a Christiano article as background, but the post itself gives no further specifics.