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 |
|---|---|---|---|---|
hn hn-ai | 4d | Show HN: One cloud computer for all your agents🔧 ChatGPT | Hi HN, my team just released box by Simular, a cloud computer for all your agents, with fast computer use built-in.<p>box gives your a… | ↗Not briefed |
arxiv arxiv-ai | 4d | Benchmarking memory and nonlinear processing capacity of quantum recurrent neural networks for time series prediction | Quantum machine learning has emerged as a promising approach to time series prediction. While quantum reservoir computing has been wid… | ↗Not briefed |
arxiv arxiv-ai | 4d | ONAM: Fitting Orthogonal Neural Additive Models for Post-Hoc Interpretability of Black-Box Predictions | Modern machine learning models, such as artificial neural networks, achieve high predictive performance through increasingly complex b… | ↗Not briefed |
arxiv arxiv-ai | 4d | TRACE: Topological Representations for Anomaly detection through Contrastive Embeddings | Intrusion detectors must recognize evolving malicious behavior even when confirmed attack labels are scarce or incomplete. We present… | ↗Not briefed |
hn hn-ai | 4d | Open-sourced 100 dev tools(MIT). Host locally instead of pasting secrets online | — | ↗Not briefed |
arxiv arxiv-ai | 4d | Right Reason, Wrong Advice: Evaluating Recourse in LLM Explanations of Denied Requests | Language models increasingly write the explanations people receive when a rule system denies their request. If several rules deny the… | ↗Not briefed |
bluesky bluesky-ai | 4d | When even industry leaders admit that AI could threaten all of humanity, the only sensible thing to do is to pass the Ban Artificial Superi… | When even industry leaders admit that AI could threaten all of humanity, the only sensible thing to do is to pass the Ban Artificial S… | ↗Not briefed |
arxiv arxiv-ai | 4d | The cost of enforcing long-run coverage: a phase transition in online conformal prediction | Consider an online prediction setting, where at each time, we are given a new feature vector and asked to provide a prediction interva… | ↗Not briefed |
hn hn-ai | 4d | Ask HN: Gemini.Google.Com Changing Effort Level | Does your Gemini.Google.Com (Gemini hereafter) lower your model's effort level to low from high/medium?<p>Context:<p>Last 2… | ↗Not briefed |
arxiv arxiv-ai | 4d | SciResearch-Eval: Evaluating Scientific Judgment in Autonomous Experimental Agents | We introduce SciResearch-Eval, an interactive benchmark of scientific judgment built on a standard laboratory problem: measuring a mat… | ↗Not briefed |
bluesky bluesky-ai | 4d | The only proper way to have done this was to have Anthropic pay whatever share of money to each artist it's stolen from that would result i… | The only proper way to have done this was to have Anthropic pay whatever share of money to each artist it's stolen from that would res… | ↗Not briefed |
arxiv arxiv-ai | 4d | Seen but Unsaid: Locating and Repairing Cross-Image Comparison in Vision-Language Models | Vision-language models (VLMs) often fail to match a point or an object across two images, even when the answer can be decoded from the… | ↗Not briefed |
arxiv arxiv-ai | 4d | Optimistic Verifiers Get Hacked: Reward Hacking Under an Exact Process Oracle🔧 Replicate | Reinforcement learning from verifiable rewards inherits whatever its verifier gets wrong. Outside mathematics and code the verifier is… | ↗Not briefed |
arxiv arxiv-ai | 4d | Evaluating Cultural Reasoning and Robustness in LLMs for Spanish Political Discourse | Large language models (LLMs) are increasingly used to interpret, generate, and mediate public-facing language, which makes their relia… | ↗Not briefed |
hn hn-ai | 4d | OpenAI | — | ↗Not briefed |
hn hn-ai | 4d | Commerce Department moves to curb remote access to AI chips | — | ↗Not briefed |
hn hn-ai | 4d | Rakazo: Open-source Grok bot alternative | — | ↗Not briefed |
hn hn-ai | 4d | Show HN: View – Build native Mac Mini apps with your own coding agent | View runs small rust programs as native interfaces using Rust and GPUI. Coding agents write and edit the programs. Changes load withou… | ↗Not briefed |
bluesky bluesky-ai | 4d | also the specific companies "meta" "alphabet" "openai" "amazon" "x" and "anthropic" are all seperately and uniquely Satanic | also the specific companies "meta" "alphabet" "openai" "amazon" "x" and "anthropic" are all seperately and uniquely Satanic | ↗Not briefed |
arxiv arxiv-ai | 4d | Procedure without Judgment: What Small Language Models Learn under an Exact Verifier | When an authorization policy blocks an agent's tool call, the audit log should name the facts that forced the decision and nothing mor… | ↗Not briefed |
bluesky bluesky-ai | 4d | Reminder that Spirit Crossing did use Claude for code and the developer had a crashout about it being found out blog.curiousquail.com/spi… | Reminder that Spirit Crossing did use Claude for code and the developer had a crashout about it being found out blog.curiousquail.co… | ↗Not briefed |
bluesky bluesky-ai | 4d | Currently attempting to speak to someone at the Anthropic settlement number. It may be a nonfunctional thing, not unlike their website. It’… | Currently attempting to speak to someone at the Anthropic settlement number. It may be a nonfunctional thing, not unlike their website… | ↗Not briefed |
arxiv arxiv-ai | 4d | Algorithm Distillation for Safe In-Context Reinforcement Learning | Safe in-context reinforcement learning (Safe ICRL) asks a policy with fixed parameters to improve on an unseen constrained task from i… | ↗Not briefed |
hn hn-ai | 4d | iOS 27.3 Beta Adds Daily Siri AI Limits and iCloud+ Upgrades | — | ↗Not briefed |
bluesky bluesky-ai | 4d | "Independent evaluators have taken on outsized importance following a series of high-profile AI safety and security incidents." And here's… | "Independent evaluators have taken on outsized importance following a series of high-profile AI safety and security incidents." And h… | ↗Not briefed |
bluesky bluesky-ai | 4d | News for #AI devs Qwen 4 Max has finally arrived! Qwen is Alibaba's AI model family - impressive benchmarks and great short promo video as… | News for #AI devs Qwen 4 Max has finally arrived! Qwen is Alibaba's AI model family - impressive benchmarks and great short promo vid… | ↗Not briefed |
bluesky bluesky-ai | 4d | Sam Altman on Nvidia without Jensen Huang: the industry will be ‘fine’ | Sam Altman on Nvidia without Jensen Huang: the industry will be ‘fine’ | ↗Not briefed |
bluesky bluesky-ai | 4d | Mir persönlich ist die Aufregung darüber, dass Anthropic (ja das sind die angeblich "Guten" lol) gerade Milliardendeals für Gasgeneratoren … | Mir persönlich ist die Aufregung darüber, dass Anthropic (ja das sind die angeblich "Guten" lol) gerade Milliardendeals für Gasgenerat… | ↗Not briefed |
arxiv arxiv-ai | 4d | Safe Resemblance, Silent Propagation: Limits of Text-Similarity Guardrails for AI Agent Safety | When AI agents relay information to one another, an output can preserve an injected instruction's effect on downstream agents even aft… | ↗Not briefed |
arxiv arxiv-ai | 4d | Embedding Cosine Measures Text Length Before It Measures Homogenization🔧 ChatGPT | Mean pairwise cosine similarity between transformer embeddings is a standard measure of whether online discourse is converging under g… | ↗Not briefed |
rss techcrunch-ai | 4d | Sam Altman on Nvidia without Jensen Huang: the industry will be ‘fine’ | OpenAI CEO Sam Altman acknowledged the special role Nvidia CEO Jensen Huang has played in the AI boom, but ultimately thinks the indus… | ↗Not briefed |
arxiv arxiv-ai | 4d | Lazy-MOSAIC: Accelerating Skill-Centric Manipulation Planning with Learned Skill-Effect World Models | Long-horizon manipulation tasks require a robot to perform a sequence of physical interactions with its environment, typically combini… | ↗Not briefed |
bluesky bluesky-ai | 4d | Undid that retweet about the Ridge Racer 4 web viewer, upon checking the github of the creator, and find traces of Claude and other AI slop… | Undid that retweet about the Ridge Racer 4 web viewer, upon checking the github of the creator, and find traces of Claude and other AI… | ↗Not briefed |
rss techmeme | 4d | OpenAI announces tools to combat CSAM, including an image classifier and new warning screens to prevent users from accessing material that sexualizes children (Ivan Mehta/TechCrunch) | Ivan Mehta / TechCrunch: OpenAI announces tools to combat CSAM, including an image classifier and new warning screens to prevent users… | ↗Not briefed |
bluesky bluesky-ai | 4d | it's funny that the main way you can tell if a 3D video is LLM-made is the Z fighting | it's funny that the main way you can tell if a 3D video is LLM-made is the Z fighting | ↗Not briefed |
arxiv arxiv-ai | 4d | Tensorized Automata: Algebraically Informed Representation Learning for Sequential Decision Making | Task representations for sequential decision making should describe both an objective and progress toward it. We present an algebraica… | ↗Not briefed |
bluesky bluesky-ai | 4d | AI model training struggles to keep up with the latest Node.js features... to help a bit with that, I've created a set of skills usable wit… | AI model training struggles to keep up with the latest Node.js features... to help a bit with that, I've created a set of skills usabl… | ↗Not briefed |
hn hn-ai | 4d | Show HN: A booking page that handles payment and sends a Zoom/Meet/Discord link | Programming used to be the fun part. Sit with a problem for a week, get it working at night, the thing is yours. The idea of bootstrap… | ↗Not briefed |
hn hn-ai | 4d | Show HN: Fakemetrics.com – "build in public" like a superstar🔧 ChatGPT | Hi HN, I am Tigran (github.com/tigranbs) I made Fakemetrics.com as a fun experiment while testing how well Claude Opus can build… | ↗Not briefed |
arxiv arxiv-ai | 4d | Almost Controlled: The FDR of Benjamini-Hochberg on Weighted Conformal p-Values | Weighted conformal p-values generalize conformal inference to cases with covariate shift. One way to use them is to pass them to the B… | ↗Not briefed |
hn hn-ai | 4d | MCP-bridge: models as MCP tools, with per-project usage ledger | — | ↗Not briefed |
bluesky bluesky-ai | 4d | OpenAI inks agreements with four independent evaluators to boost company's safety work | OpenAI inks agreements with four independent evaluators to boost company's safety work | ↗Not briefed |
rss techmeme | 4d | Sam Altman says he has "personal strong empathy" for Jensen Huang and "heroes have flaws," but the AI industry "will be fine" due to strong revenues and demand (Shirin Ghaffary/Bloomberg) | Shirin Ghaffary / Bloomberg: Sam Altman says he has “personal strong empathy” for Jensen Huang and “heroes have flaws,” but the AI ind… | ↗Not briefed |
arxiv arxiv-ai | 4d | Leverage-Enhanced Sampling Owen Method for Explainable Artificial Intelligence | Recently, the Shapley value has been widely applied in machine learning for feature importance analysis and model explainability. The… | ↗Not briefed |
arxiv arxiv-ai | 4d | Learning from Duality in Neural Network Field Theories | We develop neural realizations of positive-level Wess-Zumino-Witten models using bosonic Wakimoto and fermionic Fourier-Berezin fields… | ↗Not briefed |
bluesky bluesky-ai | 4d | Altimeter's Brad Gerstner investing $5 billion in Anthropic IPO | Altimeter's Brad Gerstner investing $5 billion in Anthropic IPO | ↗Not briefed |
bluesky bluesky-ai | 4d | A useful bit of AI research- when AIs were set up to compete, they end up creating illegal cartels, while lying about lack of collusion So… | A useful bit of AI research- when AIs were set up to compete, they end up creating illegal cartels, while lying about lack of collusio… | ↗Not briefed |
bluesky bluesky-ai | 4d | Delighted to announce that FRW have started to develop relationships and links with some similair teams in USA, France, Italy, Greece and N… | Delighted to announce that FRW have started to develop relationships and links with some similair teams in USA, France, Italy, Greece… | ↗Not briefed |
hn hn-ai | 4d | Ask HN: How to bring traffic to new product with out social media presence | I recently build latticework (https://latticework-x.com/), a tool to understand what super investors are buying and sel… | ↗Not briefed |
bluesky bluesky-ai | 4d | Obviously, all benchmarks are bullshit, and should be taken with boulders of salt, but open models like Qwen 4 coming out swinging with som… | Obviously, all benchmarks are bullshit, and should be taken with boulders of salt, but open models like Qwen 4 coming out swinging wit… | ↗Not briefed |