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 | 17d | I think model use is becoming a bit stickier, the anthropic models are generally nicer to work with in my experience and their brand is in … | I think model use is becoming a bit stickier, the anthropic models are generally nicer to work with in my experience and their brand i… | ↗Not briefed |
hn hn-ai | 17d | What you get is what you mean | — | ↗Not briefed |
arxiv arxiv-ai | 17d | Is 2:4 Sparsity Still Worth It? Semi-Structured Sparsity versus Quantization in LLM Inference🔧 vLLM | NVIDIA tensor cores have supported 2:4 semi-structured sparsity since Ampere, and the pattern composes with quantization. When the two… | ↗Not briefed |
bluesky bluesky-ai | 17d | Computational modelling has the potential to do all kinds of cool stuff, as well as all kinds of terrible stuff. Also, computational modell… | Computational modelling has the potential to do all kinds of cool stuff, as well as all kinds of terrible stuff. Also, computational m… | ↗Not briefed |
bluesky bluesky-ai | 17d | Fact Check: FAKE Photo Shows Netanyahu In Front Of Whiteboard Plan Involving "SPAIN" -- Generated With OpenAI #Ukraine | Fact Check: FAKE Photo Shows Netanyahu In Front Of Whiteboard Plan Involving "SPAIN" -- Generated With OpenAI #Ukraine | ↗Not briefed |
arxiv arxiv-ai | 17d | Self-Improving Agents with World Model Predictive Paths | Large language model (LLM) agents are increasingly expected to interact with environments and improve their decisions during deploymen… | ↗Not briefed |
arxiv arxiv-ai | 17d | Guilt by Association: Semantic Leakage Drives Associative Gender Bias | Semantic leakage occurs when a concept in the prompt steers model generation about an entity, even though the concept is irrelevant to… | ↗Not briefed |
bluesky bluesky-ai | 17d | forests near me how to disappear into forest how to disappear into forest site:reddit.com mushroom identification guide mushroom identi… | forests near me how to disappear into forest how to disappear into forest site:reddit.com mushroom identification guide mushroom i… | ↗Not briefed |
rss techmeme | 17d | Elon Musk says Tesla and SpaceX have formed special committees to discuss a possible merger and calls the possibility of a Tesla-SpaceX merger a "good idea" (Andrew J. Hawkins/The Verge) | Andrew J. Hawkins / The Verge: Elon Musk says Tesla and SpaceX have formed special committees to discuss a possible merger and calls t… | ↗Not briefed |
hn hn-ai | 17d | Claude's "Always ask" does not | — | ↗Not briefed |
bluesky bluesky-ai | 17d | it's about 1.5 year since openai announced the partnership with Jony Ive to make The Device | it's about 1.5 year since openai announced the partnership with Jony Ive to make The Device | ↗Not briefed |
bluesky bluesky-ai | 17d | When art is made by people, it is good or bad because of choices they made. The output of an LLM is a glorified mobile phone autocomplete t… | When art is made by people, it is good or bad because of choices they made. The output of an LLM is a glorified mobile phone autocompl… | ↗Not briefed |
bluesky bluesky-ai | 17d | The llm should ONLY be able to recount it in an exaggerated valley girl accent, however | The llm should ONLY be able to recount it in an exaggerated valley girl accent, however | ↗Not briefed |
arxiv arxiv-ai | 17d | Multi-Objective Protein Optimization via Modular Diffusion Steering | Protein engineering often requires improving several properties simultaneously, motivating the search for high-performing sequences th… | ↗Not briefed |
bluesky bluesky-ai | 17d | The Diagnostix remix of Sweet Caroline is, however, infinitely more interesting and artistically relevant than anything produced by an LLM,… | The Diagnostix remix of Sweet Caroline is, however, infinitely more interesting and artistically relevant than anything produced by an… | ↗Not briefed |
bluesky bluesky-ai | 17d | god i need a discord LLM that just summarizes conversation i missed, jesus christ | god i need a discord LLM that just summarizes conversation i missed, jesus christ | ↗Not briefed |
hn hn-ai | 17d | Dynamo: A Transparent Dynamic Optimization System [pdf] | — | ↗Not briefed |
arxiv arxiv-ai | 17d | StyleStream 2.0: Fast and Controllable Streaming Voice Style Conversion | Voice style conversion (VSC) aims to transform a source utterance to match the timbre, accent, and emotion of a target voice while pre… | ↗Not briefed |
hn hn-ai | 17d | Show HN: MediaIndex: Search, count and edit your photos offline on a Mac | — | ↗Not briefed |
hn hn-ai | 17d | Bringing uv and ty together | — | ↗Not briefed |
arxiv arxiv-ai | 17d | Auditing Conformal Prediction for Language Models Under Inference-Time Model Shift | Conformal prediction (CP) for language models is typically evaluated under a matched protocol, in which calibration and deployment use… | ↗Not briefed |
rss techmeme | 17d | Eight intermediate TLS certificates from eMudhra with private key sizes of just 3 to 8 bits were compromised; browser makers and major apps have blocked them (Dan Goodin/Ars Technica) | Dan Goodin / Ars Technica: Eight intermediate TLS certificates from eMudhra with private key sizes of just 3 to 8 bits were compromise… | ↗Not briefed |
bluesky bluesky-ai | 17d | Anthropic’s cloud partnerships are getting messy | Anthropic’s cloud partnerships are getting messy | ↗Not briefed |
bluesky bluesky-ai | 17d | 「Aiを使うと思考力が落ちる」という研究は既にいくつもあるが、ハーバード大学がとうとう「Ai(LLM)は認知ウイルスである」という論文を発表した。 | 「Aiを使うと思考力が落ちる」という研究は既にいくつもあるが、ハーバード大学がとうとう「Ai(LLM)は認知ウイルスである」という論文を発表した。 | ↗Not briefed |
bluesky bluesky-ai | 17d | Hitler could credibly promise full employment in exchange for giving up your rights. This is what we get. www.wsj.com/business/mus... | Hitler could credibly promise full employment in exchange for giving up your rights. This is what we get. www.wsj.com/business/mus... | ↗Not briefed |
bluesky bluesky-ai | 17d | In March 2026, forecasters were asked how high AI company revenue would go by year end. Superforecasters (the best-calibrated group) said $… | In March 2026, forecasters were asked how high AI company revenue would go by year end. Superforecasters (the best-calibrated group) s… | ↗Not briefed |
bluesky bluesky-ai | 17d | it’s funny to me that people are arguing that data theft is akin to sexual violence when I’ve been making that exact argument about openai … | it’s funny to me that people are arguing that data theft is akin to sexual violence when I’ve been making that exact argument about op… | ↗Not briefed |
bluesky bluesky-ai | 17d | I am deeply skeptical about the idea of a ghost within the LLM machine--it's not the technology itself that I'm thinking about, but rather … | I am deeply skeptical about the idea of a ghost within the LLM machine--it's not the technology itself that I'm thinking about, but ra… | ↗Not briefed |
rss techcrunch-ai | 17d | Anthropic’s cloud partnerships are getting messy | Anthropic is reportedly struggling to secure enough compute from cloud providers that increasingly view it as a competitor, even as it… | ↗Not briefed |
hn hn-ai | 17d | Symmetric Lists for Haskell | — | ↗Not briefed |
hn hn-ai | 17d | 'How hard could it be?': candidate with viral signs gets elected | — | ↗Not briefed |
hn hn-ai | 17d | Foundry-protocol – A framework for orchestrating complex tasks with AI agents | — | ↗Not briefed |
bluesky bluesky-ai | 17d | LLM stands for Limited Liability Model | LLM stands for Limited Liability Model | ↗Not briefed |
bluesky bluesky-ai | 17d | A Anthropic pel que sigui et demanen un nivell molt alt de coneixements i no et deixen fer servir la ia a les entrevistes... Vaja. | A Anthropic pel que sigui et demanen un nivell molt alt de coneixements i no et deixen fer servir la ia a les entrevistes... Vaja. | ↗Not briefed |
bluesky bluesky-ai | 17d | "Roughly a third of the draft prospectus was dedicated to risk factors, including the possibility of misaligned models causing real-world h… | "Roughly a third of the draft prospectus was dedicated to risk factors, including the possibility of misaligned models causing real-wo… | ↗Not briefed |
bluesky bluesky-ai | 17d | Past year I’ve replaced most of my services with #opensource alternatives. It feels against the grain in AI and software. AI made things f… | Past year I’ve replaced most of my services with #opensource alternatives. It feels against the grain in AI and software. AI made thi… | ↗Not briefed |
hn hn-ai | 17d | The Danger of Getting Too Close to a Source | — | ↗Not briefed |
hn hn-ai | 17d | Impossible to get Adsense approval for my browser game | Any advice to get an approval from Adsense? I don't want to put many ads in my web browser game, just an ad video before playing… | ↗Not briefed |
arxiv arxiv-ai | 17d | Back Up Before You Regret: Measuring Backup Awareness and Automating Backup in Coding Agents | Coding agents now complete tasks autonomously, modifying real-world state such as artifacts, software environments, and configurations… | ↗Not briefed |
bluesky bluesky-ai | 17d | guess that anthropic paper on evil AI hiding in your fridge is more plausible now | guess that anthropic paper on evil AI hiding in your fridge is more plausible now | ↗Not briefed |
bluesky bluesky-ai | 17d | The ps5 pro launched at $700 in 2024 without a disc drive and is only about 50 percent more powerful with the same amount of RAM as the bas… | The ps5 pro launched at $700 in 2024 without a disc drive and is only about 50 percent more powerful with the same amount of RAM as th… | ↗Not briefed |
hn hn-ai | 17d | Tell HN: Reddit JSON endpoints might have been turned off | — | ↗Not briefed |
arxiv arxiv-ai | 17d | The Data Differentiator Framework (DDF): Quantifying Marginal Data Value with Information-Theoretic and Game-Theoretic Measures in an Agentic Evaluation Pipeline (SDSAE) | In data-rich enterprises the hard question is not whether a candidate dataset is good in isolation, but whether it adds information be… | ↗Not briefed |
bluesky bluesky-ai | 17d | Anthropic Proposes AI Bill With 50-Death Threshold Before Models Can Be Halted The company has reportedly been working with lawmakers on a … | Anthropic Proposes AI Bill With 50-Death Threshold Before Models Can Be Halted The company has reportedly been working with lawmakers… | ↗Not briefed |
bluesky bluesky-ai | 17d | Your periodic reminder that the value of shares in both Anthropic and Open AI are based on profits they have never made for products that n… | Your periodic reminder that the value of shares in both Anthropic and Open AI are based on profits they have never made for products t… | ↗Not briefed |
hn hn-ai | 17d | Ronit Gupta on the XERJ Corpus Hub | — | ↗Not briefed |
bluesky bluesky-ai | 17d | An llm wouldn't have made this mistake. | An llm wouldn't have made this mistake. | ↗Not briefed |
bluesky bluesky-ai | 17d | I'm going to teach how to build your own Ipod from scratch to stick it to Spotify and YouTube. Take back possession of your music and time.… | I'm going to teach how to build your own Ipod from scratch to stick it to Spotify and YouTube. Take back possession of your music and… | ↗Not briefed |
bluesky bluesky-ai | 17d | For those of us who come from a multidisciplinary tradition, this mindset can feel very alien. How did you *not* consider the impact of LLM… | For those of us who come from a multidisciplinary tradition, this mindset can feel very alien. How did you *not* consider the impact o… | ↗Not briefed |
hn hn-ai | 17d | Replace your SIM shifter's firmware from a browser tab, no install | — | ↗Not briefed |