AINews Portal

News

50 items (filtered)

The raw AI news stream — everything we ingest across RSS, Hacker News, arXiv, GitHub, Bluesky, YouTube and more, newest first.

SourceAgeTitleSummaryStatus
arxiv
arxiv-ai
13dLanguage model capabilities fail in a shared, predictable order under compression and perturbationCompression can preserve a language model's average accuracy while selectively degrading individual capabilities. Whether these losses…
↗Not briefed
bluesky
bluesky-ai
13dMore like Spam Altman. Let's do this. respawnfirst.com/linux-script...More like Spam Altman. Let's do this. respawnfirst.com/linux-script...
↗Not briefed
bluesky
bluesky-ai
13dIn the wake of the Facebook-Cambridge Analytica scandal, Chuck Schumer “intervened on Facebook's behalf…, telling a prominent Dem critic of…In the wake of the Facebook-Cambridge Analytica scandal, Chuck Schumer “intervened on Facebook's behalf…, telling a prominent Dem crit…
↗Not briefed
arxiv
arxiv-ai
13dHomological Barriers to Stable Learning of First-Order Nash EquilibriaWe identify a homological barrier to stable learning of local Nash equilibria, also known as first-order Nash equilibria (FONE), in co…
↗Not briefed
hn
hn-ai
13dMicron – a Pascal/Oberon-derived systems language with the power of C—
↗Not briefed
hn
hn-ai
13dAstryx: An open source design system that's customizable and agent ready—
↗Not briefed
arxiv
arxiv-ai
13dLocateAnyMesh: Grounding 3D Mesh Queries in 2D Images with Vision-Language ModelsTextured 3D meshes specify object geometry and appearance, making them a natural query for finding a particular asset in an image. We…
↗Not briefed
bluesky
bluesky-ai
13dOnce again, it was Black women who warned that the current pace of developing LLM technology would justify ignoring KNOWN harm in service t…Once again, it was Black women who warned that the current pace of developing LLM technology would justify ignoring KNOWN harm in serv…
↗Not briefed
arxiv
arxiv-ai
13dResolvent Ellipsoid for Minty Set InclusionsWe study Minty set inclusion on a compact convex set $\mathcal{X} \subseteq \mathbb{R}^d$ with a set-valued operator $A : \mathcal{X}…
↗Not briefed
arxiv
arxiv-ai
13dA Physical Testbed for Self-Driving Laboratories: Comparing Bayesian and Agentic Optimization in a Minimal Closed-Loop Color Formulation TaskSelf-driving laboratories (SDLs) integrate laboratory robotics and automation with AI-driven decision-making to close the design-build…
↗Not briefed
arxiv
arxiv-ai
13dStreaming Failure Forecasting: VLA Failure Monitors That Keep Learning After DeploymentFailure monitors for vision-language-action (VLA) policies are trained offline and then frozen. After deployment, lighting, objects an…
↗Not briefed
bluesky
bluesky-ai
13dprospect.org/2026/09/11/o...prospect.org/2026/09/11/o...
↗Not briefed
arxiv
arxiv-ai
13dSelf-Indexing Latent Memories for Access-Aware Continual Instruction Tuning of LLMsContinual instruction tuning often preserves old skills with task-private modules. Inference still requires selecting the right module…
↗Not briefed
bluesky
bluesky-ai
13dDeepSeek R1 Distill Qwen 32B vs Qwen QwQ-32B (Reasoning): Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here is th…DeepSeek R1 Distill Qwen 32B vs Qwen QwQ-32B (Reasoning): Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here…
↗Not briefed
bluesky
bluesky-ai
13dwtf did Anthropic decide that weight-space programming is off limits all of a sudden !?wtf did Anthropic decide that weight-space programming is off limits all of a sudden !?
↗Not briefed
bluesky
bluesky-ai
13dWhen Anthropic released its Mythos model, “the amount of vulnerabilities found by leading tech companies has soared eight-fold.” But attack…When Anthropic released its Mythos model, “the amount of vulnerabilities found by leading tech companies has soared eight-fold.” But a…
↗Not briefed
hn
hn-ai
13dAI to find vulnerabilities in global legislation (on Git)—
↗Not briefed
arxiv
arxiv-ai
13dA Sequential Monte Carlo Method for Data Assimilation in Subsurface Flow Systems with Hierarchical UncertaintySubsurface flow systems often involve hierarchical geological uncertainty, meaning both higher-level metaparameters and cell-by-cell p…
↗Not briefed
bluesky
bluesky-ai
13dI know they’re an LLM and find that part of the joy :)I know they’re an LLM and find that part of the joy :)
↗Not briefed
arxiv
arxiv-ai
13dA Theoretical Reduction From Membership Inference to Reconstruction Attacks Against LearnersWe study reconstruction attacks, which recover training data from a learned model. In our adversarial model, the adversary designs a h…
↗Not briefed
bluesky
bluesky-ai
13dNVIDIA CEO was just trying out DeepSeekNVIDIA CEO was just trying out DeepSeek
↗Not briefed
bluesky
bluesky-ai
13d@rey-notnecessarily.bsky.social percentage guesstimate on how many of the people liking and retweeting and quote-poasting u rn have no idea…@rey-notnecessarily.bsky.social percentage guesstimate on how many of the people liking and retweeting and quote-poasting u rn have no…
↗Not briefed
hn
hn-ai
13dIndustrial History: MWRD – Sanitary District of Chicago Railroad (2016)—
↗Not briefed
arxiv
arxiv-ai
13dThe Average Patient: Do Simulated Psychiatric Patients Answer Like Real People?Language models are used as simulated psychiatric patients to train clinicians and evaluate AI therapists, and are judged by plausibil…
↗Not briefed
bluesky
bluesky-ai
13dThere's research on it, and like, it's neither So, it's known that for example typing notes for class is worse than writing them, bc you'r…There's research on it, and like, it's neither So, it's known that for example typing notes for class is worse than writing them, bc…
↗Not briefed
hn
hn-ai
13dInternet governance about to get its quadrennial stress test at ITU meeting—
↗Not briefed
bluesky
bluesky-ai
13dNew Prompt Gallery Ver. SFW, Part 050: Touya Housaki - Log Horizon #Ai #AIart #boys #sfw #PromptGallerySFW ( ͡° ͜ʖ ͡°) SFW Gallery (TensorA…New Prompt Gallery Ver. SFW, Part 050: Touya Housaki - Log Horizon #Ai #AIart #boys #sfw #PromptGallerySFW ( ͡° ͜ʖ ͡°) SFW Gallery (Te…
↗Not briefed
arxiv
arxiv-ai
13dQuery- and Candidate-Aware Long User Sequence Modeling for Sponsored Product Ranking in SearchUser sequence modeling is widely used in recommendation and advertising. Search ranking presents an additional challenge: the current…
↗Not briefed
bluesky
bluesky-ai
13dPink, feathers and a little mischief – 95 photos waiting in the full set. Real model, 18+, no AI. 🔥 Black November – 35% off every box wit…Pink, feathers and a little mischief – 95 photos waiting in the full set. Real model, 18+, no AI. 🔥 Black November – 35% off every bo…
↗Not briefed
arxiv
arxiv-ai
13dBrain-Aligned Latent Actions: Imposing Premotor Cortex Geometry on a Robot's Action VocabularyRobot policies increasingly act through a small vocabulary of discrete latent actions learned from unlabeled video, yet training leave…
↗Not briefed
rss
techmeme
13dSources: officials at the Commerce Department's AI safety center CAISSI were asked to informally report to a White House office, raising independence concerns (Garance Burke/Associated Press)Garance Burke / Associated Press: Sources: officials at the Commerce Department's AI safety center CAISSI were asked to informally rep…
↗Not briefed
hn
hn-ai
13da way to keep coding agents within the rules—
↗Not briefed
rss
theverge-ai
13dHere’s the full text of Jensen Huang’s memo to NvidiaJensen Huang, chief executive officer of Nvidia Corp., attends a media briefing following the company's "Japan AI Ecosystem" reception…
↗Not briefed
hn
hn-ai
13dSix GPUs on One X16 Slot: R9700s and RTX 3090s Behind a PCIe Switch—
↗Not briefed
arxiv
arxiv-ai
13dNonasymptotic causal structure learning in generalized linear DAGs with continuous optimizationWe study the problem of learning directed acyclic graphs from observational data in nonlinear models and provide finite-sample, nonasy…
↗Not briefed
bluesky
bluesky-ai
13dMark linked my thesis work in reference to OpenAI announcing they’re getting in on the dead-on-arrival game generator service, which is fun…Mark linked my thesis work in reference to OpenAI announcing they’re getting in on the dead-on-arrival game generator service, which i…
↗Not briefed
bluesky
bluesky-ai
13dmy favorite thing right now is to listen to AI talks at 3x speed and keep a tally of how much is actually "the llm did the bad thing" and h…my favorite thing right now is to listen to AI talks at 3x speed and keep a tally of how much is actually "the llm did the bad thing"…
↗Not briefed
bluesky
bluesky-ai
13dJust found out today his daughter is a lobbyist for OpenAI. That alone should be an immediate disqualification.Just found out today his daughter is a lobbyist for OpenAI. That alone should be an immediate disqualification.
↗Not briefed
bluesky
bluesky-ai
13dBit by bit the “Anthropic is trying to be good for mankind“ dies starting with the leaders promise to invest profits to the good only to re…Bit by bit the “Anthropic is trying to be good for mankind“ dies starting with the leaders promise to invest profits to the good only…
↗Not briefed
hn
hn-ai
13dN6k – Let users and agents run untrusted SQL on production databases—
↗Not briefed
hn
hn-ai
13dKetchUp – Open-source SketchUp alternative built with Rust and WebGPU—
↗Not briefed
bluesky
bluesky-ai
13din a funny way this story hits neither and both of these points at the same time.in a funny way this story hits neither and both of these points at the same time.
↗Not briefed
hn
hn-ai
13dShow HN: My first project, an AI-powered financial news analyzer🔧 GroqI'm 17, and I started NotBot a few months ago to see whether I could turn an idea into a working project. My main interest is eco…
↗Not briefed
arxiv
arxiv-ai
13dRADE: Risk-Adjusted Deployment Efficiency, A Framework for Routing, Memory, and Escalation in Safe LLM SystemsA deployed language model does not merely answer questions. It spends computation, consumes time, consults evidence, exposes memory, i…
↗Not briefed
arxiv
arxiv-ai
13dMeasuring Generative AI Content in Political Campaigns: Evidence of Rising Adoption and Non-DisclosureThe use of generative AI in political campaigns has drawn public concern and policy attention. Public concerns range from distrust of…
↗Not briefed
bluesky
bluesky-ai
13dAlso: prospect.org/2026/09/11/o...Also: prospect.org/2026/09/11/o...
↗Not briefed
bluesky
bluesky-ai
13dClaudeってマジで長文の要約と生成が鬼強い気がする。ぶっちゃけChatGPTより自然な長文作ってくれるの、個人的にめっちゃ助かるんですよね。 Claude is seriously good at summarizing and generating long texts…🔧 ChatGPTClaudeってマジで長文の要約と生成が鬼強い気がする。ぶっちゃけChatGPTより自然な長文作ってくれるの、個人的にめっちゃ助かるんですよね。 Claude is seriously good at summarizing and generating long…
↗Not briefed
hn
hn-ai
13dThe NBA player who got hooked on sports betting—
↗Not briefed
arxiv
arxiv-ai
13dInside the Trading Loop: Markets, Models, and Moves with LLM AgentsWe built a hybrid multi-agent trading system that combines large language model (LLM) analysts with momentum-based portfolio construct…
↗Not briefed
arxiv
arxiv-ai
13dMeans vs. Foundation Models for Emergency Call ForecastingThis study investigates whether generalized means can achieve forecasting performance comparable to that of TimeGPT for daily emergenc…
↗Not briefed