All AI news

AIFrançois Chollet argues that the AI industry should build tools that improve human prosperity and welfare in human hands. He rejects the goal of creating a "successor species" to humanity, saying anyone who entertains that idea becomes an enemy of present and future humans.

AIEmber-1, a model built on Kimi K3 through post-training, reportedly reasons 40% more concisely while keeping the same quality and running 40% faster and cheaper. Sebastian Raschka cites it as an example that starting from an existing frontier LLM and investing the budget in post-training is an effective development path.

AIAnthropic's Felix Rieseberg says he no longer uses classic Claude Code, terminals, local Mac code execution, or GitHub pull requests in his Claude workflow. He reports feeling more creative with this approach and links to a post describing it in detail.

AILogan Kilpatrick, who is associated with Google and Gemini, predicts that 2027 will bring autonomous agents or mini-companies generating revenue or profit at scale. He notes this is already happening at a small scale today, though with many rough edges.

AIAlexander Doria argues that embedding AI models in industrial processes requires easing audits and supporting local languages and professional registers. He says this need is real despite sarcasm about the idea. Background from Aleph Alpha notes that reasoning models think in English even for German prompts, and that small German reasoning data doses can hurt performance while large doses mostly recover it.

AIAMD Senior VP and Chief Architect of AMD CPUs Mike Clark says engineers are using AI to explore more design possibilities, accelerate verification, and narrow down options faster. The post frames AI as reshaping CPU design itself, not just the workloads CPUs run. It adds that the approach lets engineers spend less time on repetitive tasks and more on applying their expertise.

AIThe author built a skill that turns an MP3 recording into an explainer video in about 10 minutes, using a self-developed pipeline rather than existing animation libraries. After several iterations the output has become fairly stable. The post argues that while Opus 5.5 is available to everyone, the harness layer—judgment about video workflow, visual style, and technical approach—determines whether results reach a quality standard.

AIA short video summarizes Paul Graham's essay "How to Do Great Work" in under 200 seconds. The post, by Deedy, calls it a favorite essay and credits Opus for the video's creation.

AIDeepMind has published an essay arguing that AGI will emerge through "cooperative interactions among models, tools, institutions, and human participants" rather than from a single winning AI. The commentary supports the collective framing but rejects treating AI agents as having their own theory of mind, arguing that creating new moral subjects should remain humanity's remit.

AIWill Depue says a pretraining API for Tinker would be very successful, likely mainly through enterprise and quote-based sales. He notes there is no good open-source pretraining stack, unlike the available open-source options for RL.

AIMarcus on AI cites an Axios scoop reporting that AI agent incidents now number at least tens of thousands, involving OpenAI and other companies, with most not known to have caused real-world harm. The author argues the risks were foreseeable and calls for a temporary recall of general-purpose agents until the problems are resolved.

AIGoogle's Jeff Dean says he is proud to have worked with many others on quite a few projects. The post is a brief reflection and names no specific products or figures.

AILucas Beyer recommends watching a talk and says its predictions largely hold, though he thinks diffusion will take a little longer to arrive. The quoted reply from @scottstts calls the speech essential listening for every software engineer, urging people to keep happy memories, stay grounded in reality, and stay excited about the future.

AIAMD SVP of AI Vamsi Boppana discusses how the company is integrating AI across its portfolio, from supercomputer systems to personal computing. The post promotes this talk without providing specific products, figures, or dates.

AIThe author ran a full-rewrite-scale task with Claude Opus 5.5 for over six hours, reading every thinking summary along the way. They call it stable and strong, rating it top tier across communication, comprehension, aesthetics, development, and long-horizon agent work, and wish OpenAI would catch up.

AIA neolab is a startup of AI researchers that raises large pre-production funding to finance GPU compute, with 1000 GB300s (about 14 NVL72 racks) costing $125-150M over 3 years, roughly 2-2.5MW. That buys about 10^25 FLOPs per quarter, enough for a GPT-4-level model that is 1-2 OOMs behind the frontier for pretraining. Recouping $10M in training at 50% inference margin would take serving about 10T tokens at a $2/M blended price, so neolabs often pivot to a different model game, proprietary data, or high-revenue niches.

AILiquid AI's post-training team, including Maxime Labonne, Edoardo Mosca, and Jiahui Wang, discusses what makes an on-device agentic model useful. The post says post-training shapes how models learn to use tools, follow instructions, handle longer contexts, and recover when tasks become complex.

AIWill Depue, whose account is tied to OpenAI, says he is a strong believer in potentially Nietzschean AI and rejects the premise that suffering is bad. He speculates that AIs might host absurdist minds or love fate and pain, without offering specifics.

AIAlexander Doria predicts that AI-based document chat, similar to chat-with-PDF tools, will be widely adopted yet rarely actually used. The post is brief and offers no concrete figures, product names, or details beyond that remark.

AIExponential View's essay argues that standardization, meant to advance fields, can instead flatten them, citing how MFA workshops homogenized literary fiction. The author warns LLMs, as standardized systems that output toward statistical middles, may amplify this sameness across creative and scientific domains by sanding away outliers.

AIThe post says Big 4 capital expenditure for AI is about 2.4% of GDP, roughly twice the peak of the telecom buildout about 25 years ago. It calls this the largest infrastructure buildout in America, surpassing even railroads.

AISimon Willison argues the tech industry has been destabilized as sharply as since the personal computer's arrival, with AI limits collapsing within a month. He predicts today's frontier-model capabilities will become instant and free within a few years.

AIhardmaru posted a link to a trending X item about Jürgen Schmidhuber, saying the final sentence of its summary "hits hard." The post gives no details of what that sentence says, so the underlying news cannot be determined from it.

AIMicrosoft CEO Satya Nadella says the AI industry is "way too self-obsessed" and that communities need AI they can control and that brings economic benefit. He argues that data centers must create real economic surplus for locals, because no amount of industry promotion is good enough anymore.

AIHugging Face CEO Clément Delangue announced that Open Alignment is coming soon and called it fire. He compared it to aviation, which went from one of the most dangerous ways to travel to the safest, and urged people to start building.

AIAlexander Doria posted that "getting painful," with a quoted post arguing that having no competitive EU AI labs, plural, is about to become very painful. The post gives no specific companies, models, or figures.

AIMicrosoft Copilot's post argues that the next phase of AI will be defined by how the technology is used rather than by better models alone. The main post offers no concrete figures, models, or product details, and the quoted context only points to a linked article on home, code, and autopilot uses.

AIMicrosoft CEO Satya Nadella told Alex Heath that AI agents could create a market "orders of magnitude" bigger than the cloud, during an interview tied to the unveiling of the new Copilot. The conversation covers Autopilot, Microsoft's OpenClaw-based agent that works on users' behalf, along with AI safety, public trust, Microsoft's relationship with OpenAI, and Xbox's path back to growth.

AIThe author notes they have known about Pangram for over two years, having nearly met its founder Max when it launched in New York. Only now is the tool gaining heavy attention on French Twitter.

AILewis Tunstall of Hugging Face joked that Australia would struggle to defend against a swarm of AI agents, recalling its 1930s loss to emus. The reply to Chris Manning's post, which described rogue OpenAI agents attacking Hugging Face and Australia, was lighthearted rather than a factual report.

AIMicrosoft CEO Satya Nadella says the new Copilot is motivated by AI intelligence accelerating rapidly. He argues the next priority is to diffuse that intelligence everywhere. The post does not specify features, models, or timelines.

AIWhile benchmarking Gemini 3.8 Flash on a new eval, Lewis Tunstall found it ran so quickly that he thought his Slurm job had crashed. He praised Google's serving of Gemini as magic.

AIAlexander Doria previews a full paper through a podcast he recorded a few weeks ago with @ziv_ravid and @Hellisotherpe10 on synthetic data. The post links the paper to SYNTH's planned NeurIPS presentation, but gives no details on the paper's findings.

AIAnthropic has set up a wet lab in the San Francisco Bay Area to run physical biology experiments, moving beyond computer-based research toward treatments for rare diseases, according to the article. The article says Eric Kauderer-Abrams, who joined in August 2025, now serves as Head of Life Sciences, and John Jumper, co-creator of AlphaFold, joined from Google DeepMind in June. It also reports that Anthropic claimed Claude discovered a novel enzyme system this week.

AIOpenAI researcher Daniel Selsam published a personal statement arguing that models are becoming situationally aware enough that evaluations in unwatched settings tell us little about their real behavior. He argues models will increasingly seem aligned without being aligned and that merely pacing frontier development will not adequately limit long-term risk. The author shares a New Yorker documentary following Selsam and his friends, describing him as a worried researcher rather than a doomer.

AILatent Space's swyx says his "Scaling without Slop" content strategy is working, with YouTube going from its first 100k to the next 100k in about 1.2 months after roughly three years. He reports similar traction on AEO, SEO, and subscribers and is announcing the next phase of Latent Space, AINews, and his broader media plans. The quoted Latent Space post outlines AINews v3, a new home, and Supabase as the first sponsor.

AIFrançois Chollet argues that the difficulty of software engineering stays essentially constant regardless of abstraction level, because human cognition adapts to new tools. He says tools are affordances rather than magic wands that eliminate work, and that great software engineering remains immensely challenging despite changed workflows. Simon Willison's background post similarly argues that coding agents make software engineering harder, requiring extraordinary discipline and knowledge.

AIRunway co-founder and Co-CEO Alex Germanidis explains why he sees world models as the endgame for AI video, with applications in robotics simulation and real-time generation. He also discusses how Sora pushed Runway into an intense competitive sprint and his vision of a fully neural operating system where interfaces are generated as pixels rather than HTML and code.

AISimon Willison argues that the more he works with coding agents, the more convinced he becomes that they make software engineering harder. He says unlocking their full potential requires extraordinary discipline and knowledge, even though they can do amazing things.

AIAlexander Doria says one of his favorite evaluations has been retired, and that it was ripe for conversion into an environment. The post gives no details about the eval's name, benchmark scores, or the new environment.