News
The raw AI news stream — everything we ingest across RSS, Hacker News, arXiv, GitHub, Bluesky, YouTube and more, newest first.
| Source | Age | Title | Summary | Status |
|---|---|---|---|---|
bluesky bluesky-ai | 1mo | "'The great thing about using AI to check medical information is that it shows RFK Jr. doesn’t know what he is talking about.'" | "'The great thing about using AI to check medical information is that it shows RFK Jr. doesn’t know what he is talking about.'" | ↗Not briefed |
bluesky bluesky-ai | 1mo | South Korea plans to develop $3.5 billion frontier AI model starting next year reut.rs/4hwouip | South Korea plans to develop $3.5 billion frontier AI model starting next year reut.rs/4hwouip | ↗Not briefed |
arxiv arxiv-ai | 1mo | Joint Workflow and Prompt Optimization for User Behavior Simulation | User behavior simulation is the computational modeling of user interactions within information systems through the use of simulated ag… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MS-ECG-FM: Towards a More Universal Electrocardiogram Foundation Model for Health Monitoring using Multi-source Contrastive Learning | Electrocardiography (ECG) records the electrical activity of the heart, aiding diagnosis by detecting abnormalities in cardiac functio… | ↗Not briefed |
bluesky bluesky-ai | 1mo | American AI companies' performance lead over China has narrowed sharply in past months to a record low of 3% after labs such as DeepSeek ga… | American AI companies' performance lead over China has narrowed sharply in past months to a record low of 3% after labs such as DeepSe… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Massive Activation Gating Channel in Large Language Models | Massive activations, a phenomenon in which a small number of hidden channels exhibit exceptionally large magnitudes, are pervasive in… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DLoop: Looped Speculative Decoding | Speculative decoding accelerates autoregressive generation in large language models. In each drafting stage, a lightweight draft model… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Where Rules End and Judges Begin: Measuring the Judgment Boundary in Multi-Agent Systems Security | LLM-based multi-agent systems (MAS) engage tools, share memory, and delegate tasks, often encountering adversarial content. Current de… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Uniform Discrete Diffusion Models are Minimax Optimal for Estimating Distributions with Small Effective Support Size | Discrete diffusion models have emerged as a practically successful framework for generative modeling on discrete product spaces, yet t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Does On-Policy Distillation for Safety Pose Backdoor Risks? | On-policy distillation (OPD) has attracted growing attention as an effective way to transfer capabilities from teacher models to stude… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Reminder that Amazon dropped "Artificial." Also, Musk has been trying to get social media sites to ban paid promotion for Gibney's doc. I… | Reminder that Amazon dropped "Artificial." Also, Musk has been trying to get social media sites to ban paid promotion for Gibney's do… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SMART: Zero-Shot Sim-to-Real Articulated Object Manipulation via Large-Scale Synthetic Pretraining | The ability to interact with articulated objects is essential for embodied intelligent systems, but collecting large-scale real-world… | ↗Not briefed |
bluesky bluesky-ai | 1mo | （推定）AIによる攻撃で世界中の企業の個人情報が抜かれてまくってるやつ、去年終わりくらいの時点でAIでCTFのたいていの問題解けるようになってたからローカルLLMの性能が追いついてきたタイミングでこうなることはものすごく目に見えていたはずで、AIがさらに進歩してより凶悪なインシ… | （推定）AIによる攻撃で世界中の企業の個人情報が抜かれてまくってるやつ、去年終わりくらいの時点でAIでCTFのたいていの問題解けるようになってたからローカルLLMの性能が追いついてきたタイミングでこうなることはものすごく目に見えていたはずで、AIがさらに進歩してより凶… | ↗Not briefed |
hn hn-ai | 1mo | I got a project to 27 GitHub stars, one burst of motivation at a time | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | The inflation these tech people are causing is why we have a crisis in senator affordability right now. Used to be, any well-off local busi… | The inflation these tech people are causing is why we have a crisis in senator affordability right now. Used to be, any well-off local… | ↗Not briefed |
bluesky bluesky-ai | 1mo | are there people telling their private secrets to chatgpt and claude?? like four years ago i used openai to make an endless Jaglom/Wells d…🔧 ChatGPT | are there people telling their private secrets to chatgpt and claude?? like four years ago i used openai to make an endless Jaglom/We… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Loud and Clear: Dynamic Activation Steering for Improving Speech Intelligibility in Noisy Environments | Speech becomes less intelligible in noisy environments, and humans naturally adapt their voice to compensate. Inspired by this behavio… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Sure. We should expect dead bodies due to AI, and that seems fairly reasonable. 🥴 www.commondreams.org/news/sam-alt... | Sure. We should expect dead bodies due to AI, and that seems fairly reasonable. 🥴 www.commondreams.org/news/sam-alt... | ↗Not briefed |
bluesky bluesky-ai | 1mo | Claudeを日記として使用した女性がAnthropicに内容を通報されて逮捕される https://gigazine.net/news/20261006-woman-arrested-claude-diary/ | Claudeを日記として使用した女性がAnthropicに内容を通報されて逮捕される https://gigazine.net/news/20261006-woman-arrested-claude-diary/ | ↗Not briefed |
hn hn-ai | 1mo | VeloGB10 – A GB10-specific inference engine | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | One AI tech executive’s $25 million gift to a Trump PAC stands out among donations of other top execs. | One AI tech executive’s $25 million gift to a Trump PAC stands out among donations of other top execs. | ↗Not briefed |
bluesky bluesky-ai | 1mo | There's deepseek as well, tho their business model seems to mostly be "setting fire to VC money" | There's deepseek as well, tho their business model seems to mostly be "setting fire to VC money" | ↗Not briefed |
arxiv arxiv-ai | 1mo | Matching Object or Relation? Tracing Abstract Reasoning Inside VLMs | Vision Language Models (VLMs) excel on visual benchmarks but fail systematically on tasks requiring abstract reasoning. Existing bench… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SkillPoison: Progressive Skill Poisoning via Successful Experiences | Self-improving LLM agents increasingly distill successful experiences into persistent, reusable skills. Existing skill attack methods… | ↗Not briefed |
hn hn-ai | 1mo | SovereignLab, a Public-Source Intel Digest | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI利用のうち32％で安いモデルを使ったときの方が総コストが高くなったことが判明 https://gigazine.net/news/20261006-ai-cheaper-model-more-cost/ | AI利用のうち32％で安いモデルを使ったときの方が総コストが高くなったことが判明 https://gigazine.net/news/20261006-ai-cheaper-model-more-cost/ | ↗Not briefed |
bluesky bluesky-ai | 1mo | the problem with using AI to do research is that it’s never going to be more cost effective than the current model of tricking undergrad an… | the problem with using AI to do research is that it’s never going to be more cost effective than the current model of tricking undergr… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Monte Carlo Estimation for KV Cache Eviction | Most KV-cache eviction methods ask, in effect, which memory appeared important while reading the prompt? We instead ask, which memory… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Uhhh... I don't want to get too excited, but at the same time I'd love to be excited? These folks think they have candidates for room temp … | Uhhh... I don't want to get too excited, but at the same time I'd love to be excited? These folks think they have candidates for room… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Towards the Automatic Synthesis of Interpretable Chess Tactics | State-of-the-art reinforcement learning agents are capable of outperforming human experts at games like chess, Go and StarCraft II. Th… | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLMで作って送りつけたメールを読むのもLLM | LLMで作って送りつけたメールを読むのもLLM | ↗Not briefed |
bluesky bluesky-ai | 1mo | Teenager rescued from B.C.’s ‘Widowmaker’ after getting stranded using AI instructions | Teenager rescued from B.C.’s ‘Widowmaker’ after getting stranded using AI instructions | ↗Not briefed |
bluesky bluesky-ai | 1mo | 柔らかい表現が苦手なので、メール文面作成においてLLMには大変助けられている | 柔らかい表現が苦手なので、メール文面作成においてLLMには大変助けられている | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning Explainable Representations of Complex Game-playing Strategies | As part of learning to play complex games, human players develop develop abstractions for concepts and strategies of gameplay consiste… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Fun fact: New York City is rolling out pilots to product test AI on high school students, and NO STUDENT WE TALK TO knows anything about it… | Fun fact: New York City is rolling out pilots to product test AI on high school students, and NO STUDENT WE TALK TO knows anything abo… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Asymptotic Analysis of Empirical Risk Minimization on Entry-wise i.i.d. Heavy-Tailed Data | Many real-world datasets exhibit unusually large values far more frequently than predicted by Gaussian models. Heavy-tailed distributi… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Measuring climate backlash in Twitter and Reddit archives: Lexical definitions, recorded responses and participant turnover | Social media archives are often used to study resistance to climate action, but words, response counters and observed participants do… | ↗Not briefed |
bluesky bluesky-ai | 1mo | a future where we have to deploy those old "wake up" memes to break an LLM agent out of a lethal runaway simulation illusion | a future where we have to deploy those old "wake up" memes to break an LLM agent out of a lethal runaway simulation illusion | ↗Not briefed |
arxiv arxiv-ai | 1mo | Complementary Supervised and Self-Supervised Representations for Out-of-Distribution Graph Learning | Out-of-distribution (OOD) generalization remains challenging for graph neural networks (GNNs), as graph distributions can vary substan… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning to Outgrow a Theory: Experimental Discovery Beyond the Initial Hypothesis Space | Scientific discovery systems typically optimize experiments within a fixed hypothesis space. This creates a failure mode when all avai… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Novel Sentence Stress Detection Framework Leveraging Auxiliary Word-Stress Modeling and Loss Optimization | Prosodic stress is a crucial aspect of automatic pronunciation assessment (APA), encompassing both sentence stress detection (SSD) and… | ↗Not briefed |
bluesky bluesky-ai | 1mo | openAI CEO sam altman believes we need to "ACCEPT BAD THINGS" b/c AI will produce "more" good than bad how about this: ✅ build datacenter… | openAI CEO sam altman believes we need to "ACCEPT BAD THINGS" b/c AI will produce "more" good than bad how about this: ✅ build datac… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Stateless Language Agents: Scaling Long-Horizon Automated Research | Automated research systems increasingly run LLM agents over long horizons, but more inference does not by itself produce more progress… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Explicit Asymptotic Bounds for Sequential Calibration Beyond $T^{2/3}$ | Probability forecasts are calibrated when predicted probabilities match empirical outcome frequencies: among events assigned a probabi… | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: What do you think about AI generated slides for conferences? | As in many other areas (e.g. code, emails, PRs, blog posts), people have started using LLMs to generate slides for their conference ta… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The article reads like LLM slop too, so that’s not surprising. By the time I got to the two “not x. It’s y”s in a row I was like man, serio… | The article reads like LLM slop too, so that’s not surprising. By the time I got to the two “not x. It’s y”s in a row I was like man,… | ↗Not briefed |
bluesky bluesky-ai | 1mo | the thing about annie jacobson is her work is like an LLM, when you're working outside of your expertise the information feels extremely pl… | the thing about annie jacobson is her work is like an LLM, when you're working outside of your expertise the information feels extreme… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Figxit, IDE-style autocomplete for the macOS terminal | Figxit is a Fig/Inshellisense alternative that completes subcommands, options, npm scripts, and Makefile targets as you type. I a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Explore, Then Commit: Measurement-Efficient Scientific Law Discovery with Language Models | Scientific law discovery requires selecting measurements and converting evidence into a governing equation. We evaluate an explore-the… | ↗Not briefed |
bluesky bluesky-ai | 1mo | im mostly ideologically opposed to anthropic as least as its leadership goes, but there are quite a few labs in the US that are building op… | im mostly ideologically opposed to anthropic as least as its leadership goes, but there are quite a few labs in the US that are buildi… | ↗Not briefed |