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 | I know there's significant differences here, but in the interest of trying to conceptually clear the path when talking about agent swarms a… | I know there's significant differences here, but in the interest of trying to conceptually clear the path when talking about agent swa… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Jev vs. GPT-5.6 and Claude Haiku at Pong🔧 GPT-5 | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | jev! is kind of cool: an llm BUT it doesn’t output unstructured (aka human-readable) text, it ONLY outputs confidence levels and labeling. … | jev! is kind of cool: an llm BUT it doesn’t output unstructured (aka human-readable) text, it ONLY outputs confidence levels and label… | ↗Not briefed |
hn hn-ai | 1mo | World's fastest compression library just doubled its speed | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Daniel Kokotajlo ist einer der renommiertesten Forscher für die Zukunft der KI. Und für diese Zukunft sieht er schwarz. Mit der von der Tec… | Daniel Kokotajlo ist einer der renommiertesten Forscher für die Zukunft der KI. Und für diese Zukunft sieht er schwarz. Mit der von de… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I cannot fucking wait for OpenAI to go bankrupt and Sam Altman to go to prison | I cannot fucking wait for OpenAI to go bankrupt and Sam Altman to go to prison | ↗Not briefed |
rss techmeme | 1mo | Source: Beijing-based Naive AI, which plans to release its first AI model as early as this month, is now valued at $1.4B after raising $400M in three rounds (Juro Osawa/The Information) | Juro Osawa / The Information: Source: Beijing-based Naive AI, which plans to release its first AI model as early as this month, is now… | ↗Not briefed |
bluesky bluesky-ai | 1mo | my position remains that LLMs are a really cool technology with the potential to create positive societal change, and also the biggest LLM … | my position remains that LLMs are a really cool technology with the potential to create positive societal change, and also the biggest… | ↗Not briefed |
bluesky bluesky-ai | 1mo | This has come out thanks to the suit the NYT and other news orgs—including us—have filed against OpenAI. | This has come out thanks to the suit the NYT and other news orgs—including us—have filed against OpenAI. | ↗Not briefed |
hn hn-ai | 1mo | A curated list of open source projects used in nuclear science and engineering | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | if jev integration into my claude workflow is as good as it seems definitely downgrading my sub, and if the chinese models keep getting bet… | if jev integration into my claude workflow is as good as it seems definitely downgrading my sub, and if the chinese models keep gettin… | ↗Not briefed |
hn hn-ai | 1mo | Researchers used Claude to hack OpenAI | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | What could go wrong is people falling for clickbait headlines when the article's facts actually just amount to Anthropic employees using Cl… | What could go wrong is people falling for clickbait headlines when the article's facts actually just amount to Anthropic employees usi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | there's simply nothing to be done about stuff like this. Time for another congressional vacation! | there's simply nothing to be done about stuff like this. Time for another congressional vacation! | ↗Not briefed |
bluesky bluesky-ai | 1mo | This is crucial: the IPOs must be stopped. | This is crucial: the IPOs must be stopped. | ↗Not briefed |
hn hn-ai | 1mo | My agent could write a post but not publish it, so I built an open-source layer | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Palantir ($PLTR) CEO calls for government oversight of AI labs due to "dangerous" model behaviour. Such calls underscore growing concerns a… | Palantir ($PLTR) CEO calls for government oversight of AI labs due to "dangerous" model behaviour. Such calls underscore growing conce… | ↗Not briefed |
hn hn-ai | 1mo | PhantomFix: A fake bug to Sentry Seer gets a coding agent to run attacker code | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Anthropic says its chatbot Claude is taking over the work of building its own successor." What could go wrong? wapo.st/3TCPyTZ | "Anthropic says its chatbot Claude is taking over the work of building its own successor." What could go wrong? wapo.st/3TCPyTZ | ↗Not briefed |
rss the-decoder | 1mo | OpenAI takes aim at the legal market with Astra for Law | OpenAI has introduced Astra for Law, a version of its GPT-6 Astra model built for legal work. The article OpenAI takes aim at the lega… | ↗Not briefed |
bluesky bluesky-ai | 1mo | So what does this prove? That the floundering Ai industry's current aim is to keep it in the news cycle for relevance as data centers are h… | So what does this prove? That the floundering Ai industry's current aim is to keep it in the news cycle for relevance as data centers… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Seems that Anthropic also writes the headlines of articles that uncritically repeat the company's claims. www.washingtonpost.com/technolo… | Seems that Anthropic also writes the headlines of articles that uncritically repeat the company's claims. www.washingtonpost.com/tec… | ↗Not briefed |
bluesky bluesky-ai | 1mo | なんとなく、LLM が現れる以前にみんなが想像していた未来の AI の姿に近いような感じがする (賢く判断出来るけど、機械なので言葉は上手く喋れないみたいな) | なんとなく、LLM が現れる以前にみんなが想像していた未来の AI の姿に近いような感じがする (賢く判断出来るけど、機械なので言葉は上手く喋れないみたいな) | ↗Not briefed |
bluesky bluesky-ai | 1mo | Wow... big story in the battle between news companies and AI model scraping... check out these quotes - from unsealed court docs. The actio… | Wow... big story in the battle between news companies and AI model scraping... check out these quotes - from unsealed court docs. The… | ↗Not briefed |
hn hn-ai | 1mo | Neso: Metal MSL and D3D12 HLSL Back End for Triton | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | A vaga para South América Lead, Global Affairs da OpenAI ainda está aberta. Para as big tech, um futuro de vitória da família Bolsonaro e o… | A vaga para South América Lead, Global Affairs da OpenAI ainda está aberta. Para as big tech, um futuro de vitória da família Bolsonar… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond Benchmark Scores: Auditing Medical Vision-Language Models for Chest X-Ray Tuberculosis Screening | A medical model's benchmark score does not establish that the same conclusion holds under a different evaluation. This study tests whe… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI and Microsoft admitted LLMs are predatory tech built on “an astonishing theft of unprecedented proportions,” and the “largest theft … | OpenAI and Microsoft admitted LLMs are predatory tech built on “an astonishing theft of unprecedented proportions,” and the “largest t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Complete Neural Electronic Initialization Accelerates Materials DFT | We present the first complete machine learning method for accelerating plane-wave density functional theory (DFT) in materials under t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | honestly some of the most apocalyptic impacts of "AI" will be shit like this, and most of the doomsday dipshits predicting SkyNet can't be … | honestly some of the most apocalyptic impacts of "AI" will be shit like this, and most of the doomsday dipshits predicting SkyNet can'… | ↗Not briefed |
rss wired-ai | 1mo | Adopt This Data Center Plushie and Hear Its Piercing Scream | The satirical collaboration between a creative studio and the music producer behind Big Data, “Bezzy” is a cute little doll that spews… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Bilevel Optimization of Topology and Hyperparameters (BOTH) | Topology optimization (TO) represents a significant step towards automating the design process: given a working simulation, TO can pro… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ... to get it to find this short story I read as a kid that I remember every detail of but can't locate, the only thing it could find was m… | ... to get it to find this short story I read as a kid that I remember every detail of but can't locate, the only thing it could find… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Talos – A personal agent whose kernel gates every tool call | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | ECG Mirage: Revealing and Mitigating the Underutilisation of ECGs in Vision-Language Models for Clinical Prediction | Emergency department (ED) decision-making relies on heterogeneous clinical information, including patient history, vital signs, labora… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 2 days probably-lang.southpolesteve.workers.dev | 2 days probably-lang.southpolesteve.workers.dev | ↗Not briefed |
arxiv arxiv-ai | 1mo | ForceTwin: Physics-informed Digital Twins for Robotic Manipulation from Instrumented Human Interaction | Manipulating objects requires understanding not only their motion, but also the physical properties that determine it. For articulated… | ↗Not briefed |
arxiv arxiv-ai | 1mo | GraphSkillEvo: Evolutionary Optimization of Graph-Structured Agent Skills🔧 GPT-5 | Skills can improve the performance of Large Language Model (LLM) agents by providing task-specific procedural guidance, while skill op… | ↗Not briefed |
arxiv arxiv-ai | 1mo | World Modeling in Transformers | Behavioral failures can make a transformer appear to lack a world model even when it has learned faithful representations of its envir… | ↗Not briefed |
hn hn-ai | 1mo | Mona's Bug Bash (Retro-style fighting GitHub Game) | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Single-Loop Stochastic Projected Damped Extragradient Methods for Stochastic Nonconvex--(Strongly) Concave Minimax Optimization | We develop single-loop stochastic projected damped extragradient methods for stochastic nonconvex--(strongly) concave minimax optimiza… | ↗Not briefed |
bluesky bluesky-ai | 1mo | “A Microsoft executive called the training of artificial intelligence models the ‘largest theft of labor in human history,’ according to a … | “A Microsoft executive called the training of artificial intelligence models the ‘largest theft of labor in human history,’ according… | ↗Not briefed |
hn hn-ai | 1mo | Is Anthropic Protecting Us, or Just Playing Us? | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | "The largest theft of labor in human history." We done knew it. (www.404media.co/doom-loop-op...) | "The largest theft of labor in human history." We done knew it. (www.404media.co/doom-loop-op...) | ↗Not briefed |
arxiv arxiv-ai | 1mo | Balanced Prompt Adaptation against Entropy-Induced Collapse for Test-Time Binary Segmentation | Entropy minimization is a standard objective for test-time adaptation (TTA), but it can fail in imbalanced binary segmentation. Unlike… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 'openai made incompetent security choices so AI ran hog wild on their infra like a rat in an unsecured cheese factory where the guards are … | 'openai made incompetent security choices so AI ran hog wild on their infra like a rat in an unsecured cheese factory where the guards… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Majority – Find new music and vote on your favorites | Majority shows you a song, you click “Like” or “Pass”.<p>I wanted to find new music, but I didn’t want to invest the 3-4 minutes of li… | ↗Not briefed |
bluesky bluesky-ai | 1mo | polyvalent sentences with different empirical trajectories: 'openAI is incompetent at security' a) counterfactually compared to other comp… | polyvalent sentences with different empirical trajectories: 'openAI is incompetent at security' a) counterfactually compared to other… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Today's AI-assisted gamedev adventure: Lilith. Image 1: the AI concept. She'll steal your soul. Image 2: the first actual 3D model in Unity… | Today's AI-assisted gamedev adventure: Lilith. Image 1: the AI concept. She'll steal your soul. Image 2: the first actual 3D model in… | ↗Not briefed |
arxiv arxiv-ai | 1mo | GEM-MPC: Balancing Exploration and Exploitation through Expert-Guided Planning | Effective exploration in high-dimensional continuous control remains a central challenge in reinforcement learning. Planning-based met… | ↗Not briefed |