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 | FORALL-LEAN-AGENT for Auditable Reasoning in Formal Mathematics and Software Verification🔧 GPT-5 | Coding agents increasingly automate Lean proof development, but successful compilation alone does not establish that a candidate prove… | ↗Not briefed |
hn hn-ai | 1mo | If Robots Keep Getting Better, Will Humans Still Have an Edge? | — | ↗Not briefed |
hn hn-ai | 1mo | Tech to Replace Animal Testing Is Almost Ready. Scientists Are Not | — | ↗Not briefed |
hn hn-ai | 1mo | AI safeguards are slowing developers down | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | The FTC confirmed an industry-wide investigation into frontier AI labs over agent incidents, the first US enforcement action aimed at auton… | The FTC confirmed an industry-wide investigation into frontier AI labs over agent incidents, the first US enforcement action aimed at… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DeBERTa-ConPara: Attack-Aware and Deployment-Realistic Detection of AI-Generated Text | Robust detection of AI-generated text under deployment conditions is challenging: distribution shifts across domains and generators, a… | ↗Not briefed |
hn hn-ai | 1mo | Someone Torturing LLMs in a Robot Prison Has Triggered the Dumbest Debate in AI | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Google alone plans roughly $175–185B in 2026 capital spending; Meta $130–145B. Anthropic has disclosed $518B in long-term AI infrastructure… | Google alone plans roughly $175–185B in 2026 capital spending; Meta $130–145B. Anthropic has disclosed $518B in long-term AI infrastru… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The companies signing it, Google, Meta, OpenAI, Anthropic, Nvidia and xAI, have enormous financial stakes in the hundreds of billions in th… | The companies signing it, Google, Meta, OpenAI, Anthropic, Nvidia and xAI, have enormous financial stakes in the hundreds of billions… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The companies signing it, Google, Meta, OpenAI, Anthropic, Nvidia and xAI, have enormous financial stakes in the hundreds of billions in th… | The companies signing it, Google, Meta, OpenAI, Anthropic, Nvidia and xAI, have enormous financial stakes in the hundreds of billions… | ↗Not briefed |
bluesky bluesky-ai | 1mo | If I had a time machine, OpenAI wouldn't exist and Anthropic would mean pictures of anthropomorphic animals | If I had a time machine, OpenAI wouldn't exist and Anthropic would mean pictures of anthropomorphic animals | ↗Not briefed |
hn hn-ai | 1mo | California Legalizes Balcony Solar | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Rethinking Data Augmentation under Covariate Shift: Invariant-Guided Diffusion and Prototype Reweighting | In many industrial applications, 1) tabular data is scarce and imbalanced and thus requires synthetic expansion; 2) input distribution… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MemFit: Efficient Long-Term Agentic Memory | Long-term memory systems for large language models (LLMs) have gained popularity for extending reasoning capabilities across applicati… | ↗Not briefed |
hn hn-ai | 1mo | The Man Who Built OpenAI's First Chip | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think we're going to get a random assortment of really quite impactful technological breakthroughs in the next 5 years as scientists run … | I think we're going to get a random assortment of really quite impactful technological breakthroughs in the next 5 years as scientists… | ↗Not briefed |
bluesky bluesky-ai | 1mo | If you want to talk about AI systems, the tools, classifiers, harnesses, agents, and yes, the models too, but the other person insists the … | If you want to talk about AI systems, the tools, classifiers, harnesses, agents, and yes, the models too, but the other person insists… | ↗Not briefed |
bluesky bluesky-ai | 1mo | So what drugs are that Greg Brockman guy on because that’s either drugs, a stroke or the OpenAI president is a toddler. | So what drugs are that Greg Brockman guy on because that’s either drugs, a stroke or the OpenAI president is a toddler. | ↗Not briefed |
arxiv arxiv-ai | 1mo | An Educator-Guided LLM Pedagogical Agent for Scaffolded Feedback in Conceptual Database Design | We present an educator-guided LLM pedagogical agent for scaffolded feedback in conceptual database design. Integrated into an entity--… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Editorial: As Albanese says he supports coordinated principles for AI, he is considering an opt-out copyright system in the hope of luring … | Editorial: As Albanese says he supports coordinated principles for AI, he is considering an opt-out copyright system in the hope of lu… | ↗Not briefed |
bluesky bluesky-ai | 1mo | That's what I mean by coherent. The words read like they make sense, but it doesn't land anywhere. In that snippet there's this epistemic r…🔧 Modal | That's what I mean by coherent. The words read like they make sense, but it doesn't land anywhere. In that snippet there's this episte… | ↗Not briefed |
rss techmeme | 1mo | Artificial Analysis says Gemini 4 Argon (high) matches GPT-6 Astra (max) on its Intelligence Index and has a 15% hallucination rate, compared with 51% for Astra (Artificial Analysis) | Artificial Analysis: Artificial Analysis says Gemini 4 Argon (high) matches GPT-6 Astra (max) on its Intelligence Index and has a 15%… | ↗Not briefed |
bluesky bluesky-ai | 1mo | > a huge conspiracy to make LLMs seem smarter than they are by running inputs through a classifier, parsing math queries, and handing them … | > a huge conspiracy to make LLMs seem smarter than they are by running inputs through a classifier, parsing math queries, and handing… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It is perhaps no coincidence that having my mind shaped by George Saunders's short fiction and later his writing about writing had me yelli… | It is perhaps no coincidence that having my mind shaped by George Saunders's short fiction and later his writing about writing had me… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Kinematic MeanFlow: One-Step Action Generation Policy for Robotic Foundation Models | In this paper, we study how to achieve one-step action generation in Robotic Foundation Models (RFMs), aiming to overcome the high inf… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Stochastic parrot as a term actually tells you very little about what a normal LLM does, as it implies that there is some context-free rand… | Stochastic parrot as a term actually tells you very little about what a normal LLM does, as it implies that there is some context-free… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Don't Waste the Noise: Importance-Guided Perturbation Allocation under Joint Global and Local Constraints | Adversarial optimization under a shared $\ell_1$ budget requires deciding not only how much perturbation to use, but also where that l… | ↗Not briefed |
bluesky bluesky-ai | 1mo | FTC is investigating OpenAI and Anthropic over possible risks to consumers - The Associated Press | FTC is investigating OpenAI and Anthropic over possible risks to consumers - The Associated Press | ↗Not briefed |
rss theverge-ai | 1mo | Elon Musk’s Grokipedia has a ‘newly refreshed’ design | Grokipedia, SpaceXAI's AI-powered competitor to Wikipedia, recently started incorporating edits again, and today, it got some design t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Fixing a Model That Learned Worse Cancer Means Lower Risk: Monotonic Constraints in Bladder Cancer Recurrence Prediction | Background and Objective: Clinicians expect recurrence risk to climb with cancer severity. In a UK multicentre trial, an unconstrained… | ↗Not briefed |
bluesky bluesky-ai | 1mo | No you dont understand they need to use the orphan crushing machine to be able to get your jingling keys in 1 day rather than put the effor… | No you dont understand they need to use the orphan crushing machine to be able to get your jingling keys in 1 day rather than put the… | ↗Not briefed |
bluesky bluesky-ai | 1mo | #TellCarney #AI is unfit for Canadian businesses, government, children or education. #LLM #RegulateAI #NoGasForAI #CdnPoli | #TellCarney #AI is unfit for Canadian businesses, government, children or education. #LLM #RegulateAI #NoGasForAI #CdnPoli | ↗Not briefed |
bluesky bluesky-ai | 1mo | Trying so hard to stay out of The Discourse, but if your evidence someone used an LLM is “Their writing once contained a factual error” or … | Trying so hard to stay out of The Discourse, but if your evidence someone used an LLM is “Their writing once contained a factual error… | ↗Not briefed |
bluesky bluesky-ai | 1mo | A midweek primal scream alert for every author caught in the malicious squeeze of the Anthropic “settlement.” Deep breath in….. And 🙀😱… | A midweek primal scream alert for every author caught in the malicious squeeze of the Anthropic “settlement.” Deep breath in….. And… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI co-founder and President Greg Brockman told employees he won’t make a pledged multimillion-dollar donation to a political group work… | OpenAI co-founder and President Greg Brockman told employees he won’t make a pledged multimillion-dollar donation to a political group… | ↗Not briefed |
bluesky bluesky-ai | 1mo | anthropic upset at moonshot for acquiring claude sessions to train on meanwhile im upset at moonshot bc i think the model i am served shoul… | anthropic upset at moonshot for acquiring claude sessions to train on meanwhile im upset at moonshot bc i think the model i am served… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Fuck academic publishers who demand 50% of Anthropic settlements. Their sales weren't affected--authors' intellectual property was. Some p… | Fuck academic publishers who demand 50% of Anthropic settlements. Their sales weren't affected--authors' intellectual property was. S… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Child-Adapted Structured Phonological Representations for Interpretable Speech Sound Analysis | Structured phonological representations provide an interpretable alternative to generic speech embeddings, but existing models are lar… | ↗Not briefed |
rss techmeme | 1mo | California Governor Gavin Newsom signs the No Robo Bosses Act, which prevents employers in the state from relying solely on AI to fire or discipline workers (Paxton Honerkamp/CNBC) | Paxton Honerkamp / CNBC: California Governor Gavin Newsom signs the No Robo Bosses Act, which prevents employers in the state from rel… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SmoothOperator: Enhancing Representations for Fine-grained Open-set Recognition via Modulated Label Smoothing | Open Set Recognition (OSR) aims to enable models to accurately classify known classes while rejecting samples from unseen classes. A k… | ↗Not briefed |
bluesky bluesky-ai | 1mo | #Amazon uses “Atlas” an AI-driven prediction model to process & interpret info on workers — their thoughts, behaviors, the workplace envi t… | #Amazon uses “Atlas” an AI-driven prediction model to process & interpret info on workers — their thoughts, behaviors, the workplace e… | ↗Not briefed |
hn hn-ai | 1mo | Gemini 4 Argon: Google is back among the top three labs in intelligence achieved | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | AuraForge: Scaling Security Supervision for Training Coding Agents | Coding agents are now proficient enough to generate complex software applications from a single prompt. As their capabilities have gro… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning Multiple Timescales for Goal-Conditioned Reinforcement Learning | Existing approaches to offline goal-conditioned reinforcement learning (GCRL) struggle with long-horizon tasks. Discounting shrinks va… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The LLM majors in America are often bad faith actors. The social consequences of mass compute, prediction, and function in many professiona… | The LLM majors in America are often bad faith actors. The social consequences of mass compute, prediction, and function in many profes… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Geometric Similarity in VLM Low-Level Vision Representations | Vision-language models (VLMs) have emerged as powerful candidates for universal vision backbones, with representative architectures in… | ↗Not briefed |
bluesky bluesky-ai | 1mo | A 3D shape that can tile 3D space, leaving no gaps and never forming a repeating pattern, has been discovered using an AI model. The resear… | A 3D shape that can tile 3D space, leaving no gaps and never forming a repeating pattern, has been discovered using an AI model. The r… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I heard there was a secret code that openai ran and so did claude but you don't really care for bignum addition, do ya? | I heard there was a secret code that openai ran and so did claude but you don't really care for bignum addition, do ya? | ↗Not briefed |
bluesky bluesky-ai | 1mo | Open AI says it will delay its IPO. Where's its funding coming from in the meantime? Plus: agritourism, full employment, and too-spicy mus… | Open AI says it will delay its IPO. Where's its funding coming from in the meantime? Plus: agritourism, full employment, and too-spic… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yeah, the LLM can‘t access the insurance company's in-network list, but TBH I suspect that being able to comb providers’ websites is more l… | Yeah, the LLM can‘t access the insurance company's in-network list, but TBH I suspect that being able to comb providers’ websites is m… | ↗Not briefed |