Daily News
AI research, safety, product, and engineering links
Latest AI Reading
Source-dated posts from the last 14 days
Updated 2026-09-23 01:18 UTC
Signals of Slop
Slop isn't limited to low-effort AI-generated content. Humans can also create slop. This summary lists signals of slop : traits that make something slop, independent of the level of effort spent or the degree of AI invo...
Read originalAn unexamined cause of the OpenAI Hugging Face hacking incident: its binary performance metric
We argue that a main cause of the OpenAI Hugging Face incident was overlooked: the overly simple evaluation metric in ExploitGym was misaligned. Further, techniques already exist that can mitigate such misalignment in t...
Read originalAstra is much better at reasoning with filler tokens than previous models
Unlike other models we tested, GPT-6 Astra performs modestly better on general benchmarks with filler tokens, and significantly better on serial depth-heavy tasks.
Read originalFusion Energy Projects are Not Trying to Imitate the Sun
Post Intro: The quest to use fusion power for break-even electricity generation has sometimes been described as “putting the sun in a jar”. Where by “jar,” we mean a carefully arranged series of coils that produce a pow...
Read originalAI: artificial immigrants
Advanced AI is basically the embodiment of immigration as envisioned in the conservative nightmare: We are letting a bunch of new agents into our society They don’t clearly share our values and we suspect a society full...
Read original🔬 An Oscar, Two Asteroids, and the Algorithm in Your sklearn: John Platt on AI for Science
We talked to Google’s Oscar winning “Giganerd” about automating science, solving climate change, and how future generations can contribute to science in the age of superintelligent AI
Read originalA plan for a perpetual motion research center that runs forever
There is a popular tradition on this website of clarifying complex ideas through somewhat heavy handed, metaphorical dialogues. TO WIT, Alice says “I don’t have a good lead on how to build a perpetual motion machine, bu...
Read originalWhat’s New for Game Developers: DLSS 5 with 3D-Guided Neural Rendering, NVIDIA ACE Updates, and New RTX Kit Capabilities
NVIDIA DLSS 5 introduces DLSS 3D-Guided Neural Rendering and granular controls that help game developers add lifelike lighting and material detail while...
Read originalAnnouncing B-Side Labs: Measuring Character (Seeking Collaborators and Testers)
tl;dr Rapid AI adoption means that models are increasingly becoming autonomous decision-makers embedded in high-stakes systems. However, frontier models lack stable character, abandoning their designated personas or fac...
Read originalTrading firms could control meaningful amounts of compute by 2030
It's pretty crazy that right now, the highest margin thing to do with these models seems to be simply selling them through an API. Dwarkesh's blog prize [1] questioned how this dynamic could ever result in lab profitabi...
Read originalBring more intelligence to everyday work with GPT-6 Sol and GPT-6 Luna on Amazon Bedrock
GPT-6 Sol and GPT-6 Luna are now generally available on Amazon Bedrock, giving you more options to match intelligence and efficiency to each workload.
Read originalClaude Opus 5.5 is now available on AWS
Claude Opus 5.5, Anthropic's most capable Opus model for agentic coding, knowledge work, and long-running tasks, is now available on Amazon Bedrock and Claude Platform on AWS. This post covers what's new in Opus 5.5, pr...
Read originalEnabling Private High-Performance Production AI Inference with NVIDIA Confidential Computing
As large language model (LLM) inference increasingly processes sensitive information and proprietary model context across personal, enterprise, and regulated...
Read originalEvaluate skill-equipped agents with Strands Evals and Amazon Bedrock AgentCore
Skills let you encode domain-specific procedures as reusable, portable instructions for agents, but a fluent answer doesn't prove the agent picked the right skill or followed it. Learn how to measure skill selection and...
Read originalTopology-Aware Workload Scheduling with NVIDIA Topograph
AI factories are power-limited systems that deliver maximum value when fully optimized. GPU workload placement is a key optimization. Poor workload placement...
Read originalPremature Optimization
Your code doesn't work, but you're spending the majority of your time debating new implementations details in order to achieve maximal performance. You're finally deciding to address your longstanding problem of missing...
Read originalIntroducing Opus 5.5: Anthropic Linkpost
https://www.anthropic.com/claude-opus-5-5 It's a sizeable upgrade: Also, the first model in which they say this: Pacing the frontier Last week, our CEO, Dario Amodei, argued that AI progress should be paced so that safe...
Read originalHow Reactiv automates mobile commerce 80% faster with Amazon Bedrock AgentCore
Reactiv used Amazon Bedrock AgentCore to build a multi-agent AI Scheduler that autonomously refreshes Shopify merchants' mobile apps on a schedule, reducing merchant configuration time by 80% and getting to production 3...
Read originalRight-size generative AI endpoints with concurrency sweeps on Amazon SageMaker AI
Concurrency sweeps help you right-size a generative AI endpoint on Amazon SageMaker AI by systematically benchmarking it at increasing load levels. This post walks through deploying a model, running automated concurrenc...
Read originalHow Trane gets building insights 60x faster with Amazon Bedrock AgentCore
In about four weeks, Trane Technologies built an AI-powered agentic solution on Amazon Bedrock AgentCore that reduced a 20-minute, multi-screen building diagnostic workflow to a 20-second natural language interaction, a...
Read originalHow Tata Elxsi detects industrial safety risks in seconds on AWS
Learn how Tata Elxsi built IRIS, a real-time industrial safety platform on AWS. IRIS filters camera video at the edge, streams metadata through Amazon Kinesis, runs computer vision on Amazon SageMaker AI, and correlates...
Read originalExtending public sector intelligence with Agentforce and AWS
Public sector agencies process large volumes of unstructured evidence, such as body camera footage and scanned documents. This post shows how to combine Amazon Bedrock Data Automation with the Model Context Protocol (MC...
Read originalInitial DIY Cleanroom Experimentation
In It May Be Possible to Improvise A High Grade Bioshelter , Adin Richards discusses the possibility of improvising defenses against an environmental threat such as mirror bacteria . He gives an exploratory overview of...
Read originalDebating RSI, the US-China Gap, and Jaggedness with JS Denain of Epoch AI
Podcast #19
Read originalNVIDIA Isaac ROS 5.0 Advances Agentic, Open Source Robotics Development
To build and deploy sophisticated robotics applications that can perceive, reason and act in dynamic environments, developers need new physical AI models and tools. The ROS open framework is a project from Open Robotics...
Read originalSummary of METR's predeployment evaluation of Claude Opus 5.5
Note on independence: This evaluation was conducted under an unpaid agreement for AI R&D assessment. 1 We drafted the initial summary, and then Anthropic had the opportunity to review and edit the text. We signed off on...
Read original[AINews] Xiaomi MiMo-V2.6-Pro 1T-A42B: the new top Open Weights model, trained for $3M
crowning a new Chinese frontier lab
Read originalTransformers now runs llama.cpp quants
Read originalJun Kim, oMLX creator and maintainer, joins Hugging Face to support the MLX community
Read originalHow UK AISI and EvalEval Are Making Benchmark Results Reproducible
Read originalJev: System One models for Prod, not God — with Diogo Almeida, CEO, TypeSafe AI
The definitive Jev podcast with its lead creator.
Read originalSimplifying Model Serving Across Multiple GPUs with NVIDIA TensorRT Multi-Device Integration in NVIDIA Dynamo-Triton
The compute and memory demands of generative AI increasingly exceed what a single GPU can provide. NVIDIA TensorRT multi-device inference is a new capability...
Read originalHow to Evaluate AI Agents From Tool Calls to Task Completion
When you ship an AI agent, the key question is whether it can execute a chain of work across dozens of sequential tool calls against a live environment, and...
Read originalAccelerating a ROS 2 Node with an AI Agent and NVIDIA Isaac ROS
GPU acceleration can speed up compute-intensive robotics workloads, but a fast CUDA kernel alone does not guarantee a fast ROS 2 graph. As messages move between...
Read originalBenchmarking LLM Inference at Scale with AIPerf
You’re deploying a model on a system. It starts up, prompts are getting responses. Now the hard question: Is this fast? Your instincts might lead you to send...
Read originalxAI’s Grok 4.6 is now available in Amazon Bedrock
xAI's Grok 4.6 is now available in Amazon Bedrock: a frontier model for long-running agents, coding, and knowledge work, with a 500K token context window and four reasoning effort levels. It runs on both the bedrock-man...
Read originalNVIDIA Launches DSX Ready to Qualify Power and Cooling Products for AI Factories
Every AI factory needs power and cooling that fit its computing architecture. As AI infrastructure expands, power, cooling, water, site and grid constraints are shaping what builders can deploy. Choosing products that f...
Read originalHow BMW Group detects cost anomalies across 14,000 cloud accounts
BMW Group operates CLEA, a FinOps platform monitoring more than 14,000 cloud accounts. This post shows how BMW added automated daily cost anomaly detection, moving from reactive dashboards to proactive alerts using Prop...
Read originalWhy Deploying Physical AI at Scale Demands Safety at Every Layer
Physical AI is moving rapidly from research to large-scale deployment. By 2035, ABI Research projects an installed base of 49 million level 3-5 autonomous vehicles (AVs), while Omdia estimates that roughly 60 million in...
Read originalFrom Enablement to Execution, Egypt’s AI Ecosystem Reaches Production Scale
Today, Egypt’s AI builders gathered in the Grand Egyptian Museum for a reception that highlighted the nation’s rapidly growing AI ecosystem — spanning AI natives, developers, researchers, startups and enterprises — buil...
Read originalImproving synthesis prediction of small molecules at scale with RetroChimera
Custom-made molecules are advancing medicine, materials, and agriculture, but producing them is slow and expensive. A new Nature paper highlights RetroChimera, a predictive model that helps accelerate chemical synthesis...
Read originalAI Security Is an Engineering Problem — How to Solve It at Every Layer of the Agent Stack
AI security is an engineering problem. That means defined security requirements, enforceable controls, named owners and evidence that protections work. As AI becomes more capable, the industry must accelerate security e...
Read originalPruning LLMs Like a Physicist: Block Removal as an Ising Optimization Problem
Read originalImport AI 473: The US’s superintelligence strategy; human brain in a mouse skull; and machine hermeneutics
Welcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe now RAND thinks the best AI strategy for...
Read originalThe current balance of power in open models
The expanded form of a testimony I prepared for Congress.
Read original5 Companies Using NVIDIA AI for Clean Energy
Clean energy isn’t hard to come by, but the pace of large-scale adoption has historically been slow due to bottlenecks — including out-of-date infrastructure, elongated research and development timelines, and upfront co...
Read originaltokenizers v1: encode, decode and scaling, measured
Read originalWhy I still haven’t bought into true RSI
An “AI moderate’s” view on recent events and the trajectory of frontier models.
Read original[AINews] Here are 6 Clones of Jev in 2 days
Imitation is the sincerest form of Flattery
Read originalMilleMiglia: A realistic instance generator for middle-mile logistics
Algorithms & Theory
Read original[Paper] Stringological sequence prediction III
Abstract: In previous papers (Kosoy 2026a,b), we began the study of sequence prediction algorithms adapted to stringological word complexity measures. In particular, we defined a complexity measure called Arithmetic Rep...
Read original[AINews] not much happened today
a quiet day
Read originalDynamically Scaled Activation Steering
Activation steering has emerged as a powerful method for guiding the behavior of generative models towards desired outcomes such as toxicity mitigation. However, most existing methods apply interventions uniformly acros...
Read originalA Defense of Gradual Disempowerment
(Or: Why Bentham's Bulldog and John Halstead are wrong in their critique of Kulveit et al. ) Gradual Disempowerment is a 2025 paper (with a nice, dedicated website ) proposing a form of existential risk from AI that goe...
Read originalThe future of practice: Enabling teachers to create learning interactives with generative UI
Education Innovation
Read originalModels know when they’re reward hacking — and we can catch them at scale - Goodfire
We found a clear internal signal in models that accompanies reward hacking, and built probes that detect it — enabling efficient, real-time detection of reward hacking at scale.
Read originalHow to Use AI Agents to Prepare 3D Scenes for Simulation
Agentic AI workflows can be used to prepare and validate digital twins for physical AI systems. Agents can inspect 3D scenes, author simulation-relevant data in...
Read originalCute Critters Come to the Cloud: ‘Aniimo’ Launches on GeForce NOW
A new creature-catching adventure is ready to stream from the cloud this week. Pawprint Studio’s Aniimo arrives on GeForce NOW at launch, inviting gamers to explore the vibrant continent of Idyll across supported device...
Read original[AINews] Reality Checks on AI News (Yegge shuts down Gas Town, Databricks’ +60% Astra cost)
A dash of cold water keeps the foomers away.
Read originalREVERSAL-BENCH: A Reversibility Axis and Reset Oracle for Measuring the Reset-Free RL Cliff
A central goal of autonomous reinforcement learning is continuous policy training without external resets. However, existing paradigms largely depend on underlying environmental reversibility, a property absent in real...
Read originalIf Anyone Builds It, Everyone Dies: One Year Closer
In celebration of still being alive and fighting, we are giving away 1,000 Amazon e-books of “If Anyone Builds It, Everyone Dies”. Feel free to send a copy to yourself, a loved one, or a friend—we need all hands on deck...
Read originalTensorRT Edge-LLM Completes the MLPerf Edge Agentic Benchmark 6.4x Faster on Jetson AGX Thor
AI agents are moving from cloud data centers to vehicles, robots, and other edge devices. Unlike a chatbot that answers a single prompt, an agent works through...
Read originalUnderwriting Superintelligence: Backing Agents you can Sue — Rune Kvist, AIUC
We sit down with AIUC’s CEO on their Series A!
Read originalTranslating CUDA Tile Operations from Python to Rust Using Agentic AI
cuTile Rust (cutile-rs) is a tile-based system for safe, idiomatic GPU kernel authoring in the Rust programming language. Extending the Rust ownership model to...
Read originalNVIDIA Vera Rubin NVL72 Delivers Leading Performance in MLPerf Inference v6.1 Debut
System performance, efficient infrastructure scaling and continuous software optimization are key levers that determine AI inference economics. Higher system performance means more tokens generated, resulting in higher...
Read originalEmerald AI, Google and NVIDIA Launch Alliance to Advance Flexible AI Data Centers
AI factories are the infrastructure of the intelligence era. Scaling them responsibly will depend as much on innovation across the grid as inside the data center. Today, Emerald AI, Google and NVIDIA announced the launc...
Read originalMistral and Mozilla are bringing open, private and multilingual AI to your web browser
Open, private and multilingual AI is coming to your web browser. Mistral and Mozilla team up to put powerful, trustworthy AI where you already browse.
Read original[AINews] Jev: a “System One Model” that only decides/classifies/routes/scores — >100x faster, >200x cheaper than small frontier LLMs
congrats to TypeSafe!
Read originalUniversity of Manchester Uses NVIDIA Earth-2 to Forecast Air Pollution Across the UK
Air pollution is a serious public health risk, contributing to an estimated 30,000 deaths in the U.K. alone last year. Data-driven insights can help — but computing air quality with traditional chemistry-based models is...
Read originalTrajectory as the Teacher: Few-Step Discrete Flow Matching via Energy-Navigated Distillation
Discrete flow matching generates text by iteratively transforming noise tokens into coherent language, but may require hundreds of forward passes. Distillation uses the multi-step trajectory to train a student to reprod...
Read originalShared Selective Persistent Memory for Agentic LLM Systems
Agentic LLM systems that generate code through multi-turn tool use face a fundamental context problem: each session starts from zero, discarding the configuration choices, domain constraints, data schemas, and tool-use...
Read originalHow Value Induction Reshapes LLM Behaviour
Conversational Large Language Models are post-trained on language that expresses specific behavioural traits, such as curiosity, open-mindedness, and empathy, and values, such as helpfulness, harmlessness, and honesty....
Read originalGlyph: A Multi-Strategy Agentic System for Column Description and Sensitivity-Ontology Tagging of Enterprise Data Catalogs
Enterprise data lakes accumulate tables faster than human stewards can document or classify them, leaving columns with missing descriptions and unassigned governance labels. This documentation debt undermines data disco...
Read originalDACA-GRPO: Denoising-Aware Credit Assignment for Reinforcement Learning in Diffusion Language Models
Diffusion large language models are a compelling alternative to autoregressive models, yet existing RL methods for diffusion treat all denoising steps as equally important and rely on biased, high-variance likelihood es...
Read originalShallow Beliefs: Midtraining does not inoculate against EM from reward hacking
It would be useful if we had the ability to modify a model’s beliefs. For example, this could facilitate honeypots and better monitoring [1] , help us do better science on current models [2] , and augment certain forms...
Read originalCan Skills Learned in Games Transfer to Real-World Work?
Good Start Labs trained an AI on a railroad game — and one version improved at financial research. The difference was the training design.
Read originalBypassing inference bottlenecks: Accelerating complex AI search with Retrieve-for-Train
Algorithms & Theory
Read originalIntroducing Gemini 3.8 Live and 3.8 Live Extended Thinking
Read originalYour Agent Aced the Task. Will It Do It Again?
Read original[AINews] AEF-1 standard emerges for Third Party Evaluators, as Xai, OpenAI, and Anthropic all cosign
Pacing gathers pace.
Read originalSources
Frontier Labs
OpenAI Alignment Research Blog
OpenAI alignment research.
OpenAI Engineering
OpenAI engineering articles and system-building notes.
Anthropic Research
Anthropic research on alignment, interpretability, evaluations, and societal impacts.
Anthropic Engineering
Anthropic engineering team posts and system-building articles.
Anthropic Alignment Science
Anthropic alignment science, interpretability, and risk evaluation articles.
Google DeepMind Blog
Google DeepMind official blog.
Google Research Blog
Official Google Research blog.
Meta AI Blog
Meta AI official blog.
Microsoft Research Blog
Microsoft Research blog.
AI Safety and Governance
LessWrong
AI alignment, rationality, and AI risk discussion.
AI Alignment Forum
Technical AI alignment research community.
AI Alignment
Alignment essays and research posts.
MIRI Blog
Machine Intelligence Research Institute blog.
METR
Model evaluation and frontier-risk research.
METR Evaluations
METR evaluation reports.
Apollo Research Blog
Frontier AI risk, scheming, and evaluations.
Redwood Research Blog
AI risk and safety research.
FAR.AI Blog
AI safety and alignment research.
CAIS Blog
Center for AI Safety updates.
Goodfire Blog
Interpretability and model control.
Goodfire Research
Goodfire research index.
Epoch AI Blog
AI trends, compute, data, economics, and forecasting.
Epoch AI Latest
Unified stream for papers, newsletters, data insights, and podcasts.
AI Companies and Research Labs
Hugging Face Blog
Open-source models, Transformers, applications, and research.
Hugging Face Daily Papers
Daily AI paper discovery.
NVIDIA Technical Blog
GPU, CUDA, AI systems, and technical engineering posts.
NVIDIA Blog
NVIDIA news and applications.
Amazon Science Blog
Amazon research posts, including AI and machine learning.
AWS Machine Learning Blog
AWS ML engineering, product, and practice posts.
Apple Machine Learning Research
Apple machine learning research.
Cohere Blog
Cohere official blog.
Cohere Research
Cohere research posts.
Mistral AI News
Mistral official news and releases.
xAI News
xAI official news.
Academic Labs
BAIR Blog
Berkeley AI Research blog.
Stanford AI Lab Blog
Stanford AI Lab blog.
Stanford HAI News
Stanford HAI news and blog posts.
MIT CSAIL News
MIT CSAIL news.
MIT LINGO Blog
MIT Language and Intelligence group blog.
Personal Blogs and Newsletters
Lilian Weng - Lil'Log
Long-form posts on reinforcement learning, LLMs, agents, and alignment.
Jay Alammar Blog
Visual explanations for machine learning and Transformers.
Andrej Karpathy - Bear Blog
Andrej Karpathy's newer blog.
Andrej Karpathy - Old Blog
Older Karpathy blog posts.
Sebastian Ruder Blog
NLP and machine learning blog.
Sebastian Ruder Newsletter
NLP news and newsletter.
Import AI
Jack Clark's AI research and industry newsletter.
Import AI Substack
Substack version of Import AI.
Interconnects
Nathan Lambert's frontier AI research and industry newsletter.
Latent Space
AI Engineer newsletter and podcast.
DeepLearning.AI - The Batch
AI news digest.
Distill
Classic visual and explanatory machine learning articles.
The Gradient
AI research, society, and commentary.