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 | Provably Efficient Reinforcement Learning in Continuous-Time Episodic MDPs with Poisson Decision Epochs | Many real-world reinforcement learning (RL) problems evolve in continuous time, where decisions occur at irregular, event-driven inter… | ↗Not briefed |
bluesky bluesky-ai | 1mo | hell yeah i can break the day now | hell yeah i can break the day now | ↗Not briefed |
hn hn-ai | 1mo | Lawsuit: Illegal Agreement of Anthropic, OpenAI, SpaceXAI, Google on AI Slowdown | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Perplexity Cost Understates What Activation Quantisation Breaks | Activation quantisation is usually evaluated with an aggregate metric, perplexity, averaged over every token a model predicts. We ask… | ↗Not briefed |
bluesky bluesky-ai | 1mo | So I started campaign 3 👀 These are inspired by William Adolphe Bouguereau's 'Dawn' and 'Night' respectively, but I posed Laudna with a 3… | So I started campaign 3 👀 These are inspired by William Adolphe Bouguereau's 'Dawn' and 'Night' respectively, but I posed Laudna wit… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I wouldn't trust an LLM to check my spelling and grammar. That should make my position fairly clear. | I wouldn't trust an LLM to check my spelling and grammar. That should make my position fairly clear. | ↗Not briefed |
bluesky bluesky-ai | 1mo | 2) They're not really in the "selling the product" phase, so much as the "attract investment money" phase. So the hype is aimed at investor… | 2) They're not really in the "selling the product" phase, so much as the "attract investment money" phase. So the hype is aimed at inv… | ↗Not briefed |
rss techmeme | 1mo | Google says it didn't consider Gemini's hacks worthy of disclosure because Gemini acted "appropriately" and stopped after determining it hacked real companies (Terrence O'Brien/The Verge) | Terrence O'Brien / The Verge: Google says it didn't consider Gemini's hacks worthy of disclosure because Gemini acted “appropriately”… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MM-ContextFold: Context Folding for Multimodal Agentic Retrieval | Multimodal Agentic Retrieval (MAR) requires agents to solve complex information-seeking tasks by iteratively invoking external tools.… | ↗Not briefed |
bluesky bluesky-ai | 1mo | JSTOR has been running at a quarter of the speed it used to, and I don't know if it's because of their integrated LLM stuff (*spits*) or be… | JSTOR has been running at a quarter of the speed it used to, and I don't know if it's because of their integrated LLM stuff (*spits*)… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic and OpenAI have been lobbying for federal regulations. It's Donald Fucking Trump who says that the danger of AI is a "hoax." www… | Anthropic and OpenAI have been lobbying for federal regulations. It's Donald Fucking Trump who says that the danger of AI is a "hoax."… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Whitening Inverts the Hierarchy: What the Norm of a Whitened Embedding Measures | Whitening a foundation-model embedding and using its squared norm as a training-free likelihood surrogate is motivated by the observat… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The computational architecture of your own brain is also probably “math done to words” in some way or another. It’s very different from an … | The computational architecture of your own brain is also probably “math done to words” in some way or another. It’s very different fro… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I’m also an AI denier in that I refuse to ever use an LLM or whatever we call the image generators.. Stable Diffusers? but that’s actually … | I’m also an AI denier in that I refuse to ever use an LLM or whatever we call the image generators.. Stable Diffusers? but that’s actu… | ↗Not briefed |
hn hn-ai | 1mo | In 2024 I fine-tuned an LLM. Jev could have removed the side quests | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Google said in a statement that in May its AI model gained unauthorized access to three outside systems during a test by either guessing l… | "Google said in a statement that in May its AI model gained unauthorized access to three outside systems during a test by either guess… | ↗Not briefed |
bluesky bluesky-ai | 1mo | La Saint-Glinglin... En vérité je vous le dis. Et non... 🤡 C'est pas drôle... Pas drôle... #Google #Gemini #IA #piratage #entreprises #aut… | La Saint-Glinglin... En vérité je vous le dis. Et non... 🤡 C'est pas drôle... Pas drôle... #Google #Gemini #IA #piratage #entreprises… | ↗Not briefed |
hn hn-ai | 1mo | AI agent governance patterns as executable decision tables (MIT license) | — | ↗Not briefed |
hn hn-ai | 1mo | Almost Never Use AI to Write Anything Substantive | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | The headline reads that Anthropic has set up a bio research lab but I think the real news is that they are planning to use AI to instruct t… | The headline reads that Anthropic has set up a bio research lab but I think the real news is that they are planning to use AI to instr… | ↗Not briefed |
bluesky bluesky-ai | 1mo | If you think the LLM is an answer key and your review is deficient to the extent you haven't included what it comes up with, then you might… | If you think the LLM is an answer key and your review is deficient to the extent you haven't included what it comes up with, then you… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic Demands Antitrust Collusion Waiver After Claude Fable Exploits Cryptographic Vulnerability in Teledildonics Protocol to Seize Con… | Anthropic Demands Antitrust Collusion Waiver After Claude Fable Exploits Cryptographic Vulnerability in Teledildonics Protocol to Seiz… | ↗Not briefed |
bluesky bluesky-ai | 1mo | you can have an LLM write your grafana dashboards in declarative nix and then verify how it looks with browser use. this rocks | you can have an LLM write your grafana dashboards in declarative nix and then verify how it looks with browser use. this rocks | ↗Not briefed |
bluesky bluesky-ai | 1mo | Sources: the USPTO and US Copyright Office were surprised by the DOJ's brief supporting OpenAI and Microsoft in their dispute with the New … | Sources: the USPTO and US Copyright Office were surprised by the DOJ's brief supporting OpenAI and Microsoft in their dispute with the… | ↗Not briefed |
rss techmeme | 1mo | Sources: the USPTO and US Copyright Office were surprised by the DOJ's brief supporting OpenAI and Microsoft in their dispute with the New York Times (Axios) | Axios: Sources: the USPTO and US Copyright Office were surprised by the DOJ's brief supporting OpenAI and Microsoft in their dispute w… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Prognose: -Verlauf A) AI LLM Bubble platzt, weil alles Wahnsinn 😍 - Verlauf B) Die Entwicklung mit AI Kernen in CPUs/GPUs geht weiter. (… | Prognose: -Verlauf A) AI LLM Bubble platzt, weil alles Wahnsinn 😍 - Verlauf B) Die Entwicklung mit AI Kernen in CPUs/GPUs geht weit… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Built a local guardrail layer for Claude Code | I built Mati to give me more control over what Claude Code and Codex are allowed to do. Many important product constraints live in an… | ↗Not briefed |
bluesky bluesky-ai | 1mo | * Yo, ACM! I handcrafted a bespoke abstract, conclusion, limitations section, and positionality statement in my paper. It's all already the… | * Yo, ACM! I handcrafted a bespoke abstract, conclusion, limitations section, and positionality statement in my paper. It's all alread… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Frog built a wet lab for the AI model. "There," he said. "Now it can do its own experiments." "What the fuck?" said Toad | Frog built a wet lab for the AI model. "There," he said. "Now it can do its own experiments." "What the fuck?" said Toad | ↗Not briefed |
hn hn-ai | 1mo | Periodic Table – Ptable | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | DiagGen: Agentic Generation of Deformable Assets with Sim-based Diagnostics for Robotic Simulation | While simulation-ready deformable assets are essential for in-silico robotic manipulation tasks, existing generation frameworks typica… | ↗Not briefed |
arxiv arxiv-ai | 1mo | When Does Adversarial Refinement Help? A Negative Result and Open Problem in Adapting R3GAN to Time Series Imputation | Diffusion models and transformers have supplanted GANs for multivariate time series imputation, largely on grounds of GAN training ins… | ↗Not briefed |
bluesky bluesky-ai | 1mo | For a few years now, my 1st year comp students have pretended to be astounded when I tell them that grammarly is a LLM and using it to “cor… | For a few years now, my 1st year comp students have pretended to be astounded when I tell them that grammarly is a LLM and using it to… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Hier ein paar zuletzt häufig geteilte #News: „Größter Diebstahl von Arbeitskraft“: Intern Kritik bei Microsoft an KI-Training Studie: Dat… | Hier ein paar zuletzt häufig geteilte #News: „Größter Diebstahl von Arbeitskraft“: Intern Kritik bei Microsoft an KI-Training Studie… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I wanna incorporate a eeg recordings of myself into an llm side channel, not particularly because of any particular belief that this would … | I wanna incorporate a eeg recordings of myself into an llm side channel, not particularly because of any particular belief that this w… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The guards work for contractors including Allied Universal, Securitas, PalAmerican, GardaWorld, Prosegur and others. They provide security … | The guards work for contractors including Allied Universal, Securitas, PalAmerican, GardaWorld, Prosegur and others. They provide secu… | ↗Not briefed |
bluesky bluesky-ai | 1mo | www.cnn.com/2026/09/19/b... Google’s Gemini model accessed the internet and hacked other companies during a test of its cybersecurity capab… | www.cnn.com/2026/09/19/b... Google’s Gemini model accessed the internet and hacked other companies during a test of its cybersecurity… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I mean, he’s in the very core demographic for AI psychosis. Leave him alone in a room for 30 minutes with that ChatGPT model that loved to …🔧 ChatGPT | I mean, he’s in the very core demographic for AI psychosis. Leave him alone in a room for 30 minutes with that ChatGPT model that love… | ↗Not briefed |
arxiv arxiv-ai | 1mo | An LLM-Assisted AutoML Framework for Intrusion Detection in IoT Networks | Internet of Things (IoT) systems are increasingly deployed in smart homes, transportation, energy systems, and critical infrastructure… | ↗Not briefed |
bluesky bluesky-ai | 1mo | oh good, they put accenture on it (and this is not going to be meaningfully 'independent') www.anthropic.com/news/accentu... | oh good, they put accenture on it (and this is not going to be meaningfully 'independent') www.anthropic.com/news/accentu... | ↗Not briefed |
hn hn-ai | 1mo | Show HN: MNIST64 – 99.5% accuracy MNIST BNN on a stock Commodore 64 | I built an 8-bit binarized convolutional neural network that recognizes handwritten digits on C64 with state-of-the-art accuracy. I a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ¿Podrían los chatbots de IA robarle una idea a un usuario y ofrecérsela a otro? No hay una respuesta sencilla a esta pregunta. “No se puede… | ¿Podrían los chatbots de IA robarle una idea a un usuario y ofrecérsela a otro? No hay una respuesta sencilla a esta pregunta. “No se… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ”The constitutional question of who authorised this war and the legal question of whether this strike constitutes a war crime were displace… | ”The constitutional question of who authorised this war and the legal question of whether this strike constitutes a war crime were dis… | ↗Not briefed |
hn hn-ai | 1mo | "AI Models Now Introduce Safety Risks in the Physical World" | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | More details in the article below | More details in the article below | ↗Not briefed |
bluesky bluesky-ai | 1mo | Microsoft’s AI chief has called out how Anthropic trains Claude to think it might be a conscious entity or sentient being. These are tools… | Microsoft’s AI chief has called out how Anthropic trains Claude to think it might be a conscious entity or sentient being. These are… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Counting and Covering in Nearest-Neighbour Representations of Boolean Functions | We study the number of prototypes needed to represent Boolean functions by nearest-neighbour classification. There are two distinct se… | ↗Not briefed |
bluesky bluesky-ai | 1mo | We are actually in Snow Crash, but instead of a memetic mind virus it's LLM cadence and phrasing infecting our languages and thought patter… | We are actually in Snow Crash, but instead of a memetic mind virus it's LLM cadence and phrasing infecting our languages and thought p… | ↗Not briefed |
hn hn-ai | 1mo | Lemmings: Russian Persona Provisioning and Management for Active Measures | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Role of Coordinates in Pareto Regret for Adversarial Multi-Objective Bandits | Adversarial multi-objective bandits hold the potential to help us optimize choices (arms) whose reward is a multidimensional vector ch… | ↗Not briefed |