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 |
|---|---|---|---|---|
arxiv arxiv-ai | 1mo | Recovering Agentic Sovereignty: Mitigating the Consensus Paradox via Contrastive Epistemic Decoding🔧 Consensus | Large language models (LLMs) exhibit a parametric vulnerability to adversarial swarm consensus. To mitigate this sycophancy, we introd… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SambaGraph: Action-Reaction Spatio-Temporal Graphs for Soccer Tactical Response Modeling | Soccer tactics are interactive: an attacking action changes the opponent's defensive problem, and the observed response depends on the… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RootQuantV2: Adapting a Vision Foundation Model for Root-Trait Regression from Minirhizotron Imagery | A lack of high-throughput phenotyping solutions for root traits in field-grown crops has severely constrained understanding and improv… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI sued over Canada school shooting. British Columbia has filed a lawsuit against OpenAI in California over the company's failure to r…🔧 ChatGPT | OpenAI sued over Canada school shooting. British Columbia has filed a lawsuit against OpenAI in California over the company's failure… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Une province canadienne attaque en justice OpenAI, derrière ChatGPT, après la tuerie de Tumbler Ridge en février🔧 ChatGPT | Une province canadienne attaque en justice OpenAI, derrière ChatGPT, après la tuerie de Tumbler Ridge en février | ↗Not briefed |
arxiv arxiv-ai | 1mo | AkasicMEM: Governed Enterprise Memory for Agents | Agent memory enables enterprise agents to retain knowledge acquired during work and reuse it across tasks and agents, turning executio… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Apple challenges OpenAI’s forensic analysis in trade secrets case, seeks access to hardware R&D | Apple challenges OpenAI’s forensic analysis in trade secrets case, seeks access to hardware R&D | ↗Not briefed |
bluesky bluesky-ai | 1mo | Une province canadienne attaque en justice OpenAI, derrière ChatGPT, après la tuerie de Tumbler Ridge en février🔧 ChatGPT | Une province canadienne attaque en justice OpenAI, derrière ChatGPT, après la tuerie de Tumbler Ridge en février | ↗Not briefed |
bluesky bluesky-ai | 1mo | At Dartmouth, the student newspaper finds that the Provost of Dartmouth has written next to none of his published works since '22. If a stu… | At Dartmouth, the student newspaper finds that the Provost of Dartmouth has written next to none of his published works since '22. If… | ↗Not briefed |
arxiv arxiv-ai | 1mo | IndustrialVLA-Bench: A Traceable Multi-Axis Evaluation of Open Robot Policy Models | Open robot policies increasingly follow two paradigms: vision-language-action models (VLAs) directly map observations and instructions… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Please let all of my opponents be represented by an LLM 🤞 | Please let all of my opponents be represented by an LLM 🤞 | ↗Not briefed |
bluesky bluesky-ai | 1mo | 墓所の主とそれに使える教団みたいにに曲解したくなりますな :) openai.com/index/adviso... | 墓所の主とそれに使える教団みたいにに曲解したくなりますな :) openai.com/index/adviso... | ↗Not briefed |
bluesky bluesky-ai | 1mo | The amount of digging I had to do to find out the "AI" tool the FAA is announcing with such fan fair is just a fucking everyday Math Thing … | The amount of digging I had to do to find out the "AI" tool the FAA is announcing with such fan fair is just a fucking everyday Math T… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI sued by Canadian province over role in mass shooting ft.trib.al/3tJyEk4 | OpenAI sued by Canadian province over role in mass shooting ft.trib.al/3tJyEk4 | ↗Not briefed |
hn hn-ai | 1mo | A zero-day has been released for Meta's Muse | — | ↗Not briefed |
hn hn-ai | 1mo | I built an autonomous accounting tool to let AI do my taxes | — | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: Ceremonious Architecture in Times of AI | There's extensive literature on several software architecture patterns and styles whose benefits are well known, but which are of… | ↗Not briefed |
arxiv arxiv-ai | 1mo | HABILIS Brain 0: Geometry-Change Supervision for Vision-Language-Action and Residual Flow Recovery | Vision-language-action policies benefit from geometric supervision, but current-frame geometry alone does not explicitly describe the… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Weakly Supervised Quantum Error Mitigation | Supervised approaches to quantum error mitigation learn a map from noisy circuit outputs to ideal ones, and therefore require the idea… | ↗Not briefed |
bluesky bluesky-ai | 1mo | the @freyja-lynx.dev llm pun singularity may happen, but not today | the @freyja-lynx.dev llm pun singularity may happen, but not today | ↗Not briefed |
bluesky bluesky-ai | 1mo | 100 or more people marched and rallied against OpenAI in Manhattan today during Climate Week! High five to R.A.D. Collective NYC, Planet Ov… | 100 or more people marched and rallied against OpenAI in Manhattan today during Climate Week! High five to R.A.D. Collective NYC, Plan… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI flags 6 new incidents of 'concerning' behavior, unveils plan to track it | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | reading an llm horror story that takes pains to say "the llms isn't actually wanting to do this, it just knows that if it were evil it'd do… | reading an llm horror story that takes pains to say "the llms isn't actually wanting to do this, it just knows that if it were evil it… | ↗Not briefed |
hn hn-ai | 1mo | Escaping the OpenAI Codex sandbox, twice | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I don’t think it does. but i believe the paper provided a theoretical framework for the LLM to use and a candidate search space. So that pa… | I don’t think it does. but i believe the paper provided a theoretical framework for the LLM to use and a candidate search space. So th… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI continuing its tradition of taking what was open and making it no longer so. | OpenAI continuing its tradition of taking what was open and making it no longer so. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Ok, but, here's the problem: Pangram sucks. It doesn't provide meaningful information, it just doesn't. Yes, the people who sell it lov… | Ok, but, here's the problem: Pangram sucks. It doesn't provide meaningful information, it just doesn't. Yes, the people who sell i… | ↗Not briefed |
bluesky bluesky-ai | 1mo | One (or more?) of the recent unsolved math problems was not in fact a mathematician with an LLM. It was a software engineer with an LLM, wh… | One (or more?) of the recent unsolved math problems was not in fact a mathematician with an LLM. It was a software engineer with an LL… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Synthesis and editing of multi-instrument audio mixtures using scalar-quantised latents with MIDI Span conditioning | Music creation often involves iterative refinement, changing selected musical details while retaining the rest. To support such refine… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Really wild stuff from Meta: “every Muse user gets a better personal agent as we all collectively use the product and help the model unders… | Really wild stuff from Meta: “every Muse user gets a better personal agent as we all collectively use the product and help the model u… | ↗Not briefed |
bluesky bluesky-ai | 1mo | no it is not it’s like over 2x more expensive. it costs fable money openrouter.ai/openai/gpt-6...🔧 OpenRouter | no it is not it’s like over 2x more expensive. it costs fable money openrouter.ai/openai/gpt-6... | ↗Not briefed |
arxiv arxiv-ai | 1mo | DefaultGNN: A Dual-Perspective GNN Framework for Predicting Corporate Default from Buyer-Seller Transaction Networks | Corporate default prediction is a core problem in financial risk management, yet traditional credit models rely heavily on financial s… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A JEPA Recipe for Tabular Foundation Models | Tabular foundation models learn to predict cell values in context, whereas world-model self-supervision asks for prediction in represe… | ↗Not briefed |
bluesky bluesky-ai | 1mo | See? When presented with *exactly the thing she asks for as evidence*, she simply *denies the evidence* and moves the goalposts! | See? When presented with *exactly the thing she asks for as evidence*, she simply *denies the evidence* and moves the goalposts! | ↗Not briefed |
bluesky bluesky-ai | 1mo | Well that’s stupid, we all know that slow thinking is LLM and fast thinking is jev. | Well that’s stupid, we all know that slow thinking is LLM and fast thinking is jev. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Exactly my experience. OpenAI just introduced Astra Legal, their current top-end model and harness for legal research and drafting, braggin… | Exactly my experience. OpenAI just introduced Astra Legal, their current top-end model and harness for legal research and drafting, br… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Compressing Long Context into Answer-Aligned Memory Embeddings for LLM Inference | Large language model (LLM) inference is constrained by the quadratic scaling of self-attention and the linear scaling of the KV cache,… | ↗Not briefed |
hn hn-ai | 1mo | Claude Status – Elevated errors for multiple models | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | okay, math knowers, a question: Is it implausible that there are classes of problems that are within human grasp but unlikely to be solve… | okay, math knowers, a question: Is it implausible that there are classes of problems that are within human grasp but unlikely to be… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Like… no shit Sherlock. Corporations need to be prosecuted for the harmful effects of their products. OpenAI have done so many criminal ac… | Like… no shit Sherlock. Corporations need to be prosecuted for the harmful effects of their products. OpenAI have done so many crimin… | ↗Not briefed |
hn hn-ai | 1mo | Tinycc-cygwin (tcc-cygwin) compiler | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I find the evidence utterly unconvincing. In my experience, an LLM can only solve a problem if you tell it what the problem is, tell it how… | I find the evidence utterly unconvincing. In my experience, an LLM can only solve a problem if you tell it what the problem is, tell i… | ↗Not briefed |
hn hn-ai | 1mo | Forging 1024-bit RSA signatures in nearly SNFS time | — | ↗Not briefed |
hn hn-ai | 1mo | Claude Is Down:( 529 | — | ↗Not briefed |
hn hn-ai | 1mo | RustyTux: Linux kernel n-day local privilege escalation | — | ↗Not briefed |
hn hn-ai | 1mo | Pyrowave: GPU-back video codec for high-bandwidth, low-latency streaming | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | More details in the article below | More details in the article below | ↗Not briefed |
bluesky bluesky-ai | 1mo | I am saying an LLM cannot be lovable though. for the record | I am saying an LLM cannot be lovable though. for the record | ↗Not briefed |
hn hn-ai | 1mo | The first mRNA flu shot is now available. Here's who can get it | — | ↗Not briefed |
hn hn-ai | 1mo | ShinyHunters hijack rival Clop's leak site | — | ↗Not briefed |