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 | 2d | Beyond Discrete Tokens: Preserving Numerical Continuity in Medical Vision-Language Models | Objective: To develop a unified vision-language model for prompt-conditioned quantitative ultrasound measurement that addresses the mi… | ↗Not briefed |
bluesky bluesky-ai | 2d | Shoutout to the "reasonable centrist" takes on LLM usage for providing cover for opening the slop floodgates | Shoutout to the "reasonable centrist" takes on LLM usage for providing cover for opening the slop floodgates | ↗Not briefed |
arxiv arxiv-ai | 2d | Monetizing Intelligence: Advertising and Subscription Design on LLM Platforms | Large language model (LLM) platforms introducing advertising must decide how much capability ad revenue should fund and what consumers… | ↗Not briefed |
arxiv arxiv-ai | 2d | Knowledge Distillation across Item and Code Spaces for Parallel Generative Recommendation | Generative recommendation with semantic IDs represents each item as a sequence of discrete codes derived from its semantic representat… | ↗Not briefed |
hn hn-ai | 2d | Gemini-API-CLI: An (experimental) CLI for the Gemini API | — | ↗Not briefed |
bluesky bluesky-ai | 2d | DeepSeek gets $15B after US ban 🤖 DeepSeek has reportedly raised $15 billion, two days after the U.S. ordered federal agencies to remove … | DeepSeek gets $15B after US ban 🤖 DeepSeek has reportedly raised $15 billion, two days after the U.S. ordered federal agencies to re… | ↗Not briefed |
arxiv arxiv-ai | 2d | If at First You Don't Succeed, Backtrack: Lightweight Recovery for Frozen VLAs | Vision-Language-Action (VLA) models perform well on robotic manipulation, yet remain brittle once their own actions carry them into st… | ↗Not briefed |
arxiv arxiv-ai | 2d | Learning Positive Linear Combinations of ReLUs in Fully Polynomial Time | We study the fundamental problem of PAC learning positive linear combinations of $k$ arbitrary ReLU activations on $\mathbb{R}^d$ with… | ↗Not briefed |
bluesky bluesky-ai | 2d | This is a very, very good piece and aligns almost exactly with my own thoughts on the subject of LLM "consciousness" (though I'd probably b… | This is a very, very good piece and aligns almost exactly with my own thoughts on the subject of LLM "consciousness" (though I'd proba… | ↗Not briefed |
bluesky bluesky-ai | 2d | A modest proposal: if you cite the "LLM as judge" paper, you should only get AI reviews | A modest proposal: if you cite the "LLM as judge" paper, you should only get AI reviews | ↗Not briefed |
arxiv arxiv-ai | 2d | Yalda-Embedding: Open Development of a Competitive Persian Retrieval Model | Despite advances in multilingual text representations, effective dense retrieval for Persian remains constrained by the limited availa… | ↗Not briefed |
bluesky bluesky-ai | 2d | the only moat that anthropic and openai have left seems to be sinophobia | the only moat that anthropic and openai have left seems to be sinophobia | ↗Not briefed |
arxiv arxiv-ai | 2d | ZENO: Nonlinear filtering with Zakai Equation Neural Operators | State inference and parameter learning for nonlinear filters is a core computation for scientific machine learning with varied applica… | ↗Not briefed |
bluesky bluesky-ai | 2d | Under the very specific circumstances in which I have interacted with pangram (Very short samples, a combination of hand-written; fully LLM… | Under the very specific circumstances in which I have interacted with pangram (Very short samples, a combination of hand-written; full… | ↗Not briefed |
bluesky bluesky-ai | 2d | Yo le he pasado textos míos a un detector de IA, y me ha salido que el 80 % era IA. He hecho lo mismo con textos que sabía que los había ge… | Yo le he pasado textos míos a un detector de IA, y me ha salido que el 80 % era IA. He hecho lo mismo con textos que sabía que los hab… | ↗Not briefed |
bluesky bluesky-ai | 2d | As the gap between the internal models at the AI labs and the ones they release grow (due to safety concerns & research advantages), it wil… | As the gap between the internal models at the AI labs and the ones they release grow (due to safety concerns & research advantages), i… | ↗Not briefed |
arxiv arxiv-ai | 2d | Adaptive Frequency Threshold Byte Pair Encoding for Mixed Natural Language and Structured Symbolic Corpora | Modern language models rely on subwording tokenization to define the discrete units of input processing, typically using methods like… | ↗Not briefed |
bluesky bluesky-ai | 2d | This, with all the cyber news, and rumors of a bigger GLM model coming soon, sets us up for an exciting rest of the year! The pro name make… | This, with all the cyber news, and rumors of a bigger GLM model coming soon, sets us up for an exciting rest of the year! The pro name… | ↗Not briefed |
arxiv arxiv-ai | 2d | SemantEq: Faithful Bidirectional Equivalence of Formal Statements | Statement autoformalization translates mathematics from natural language into a proof assistant such as Lean 4, and models now do it a… | ↗Not briefed |
rss the-decoder | 2d | Anthropic says its unreleased model is both more capable and better aligned than Mythos 5.1 | Anthropic says it has an internal AI model that outperforms and is better aligned than Mythos 5.1. During a cybersecurity test by the… | ↗Not briefed |
arxiv arxiv-ai | 2d | A Tutorial on Physics Constrained Machine Learning for Hybrid Modeling, Optimization, and Control | Physics-constrained machine learning (PCML) integrates domain knowledge, first-principles, and physical constraints into AI/ML models.… | ↗Not briefed |
hn hn-ai | 2d | Do pricier robot vacuums get better reviews? 737 reviews | — | ↗Not briefed |
bluesky bluesky-ai | 2d | Human Gary King politely informs you that robust standard errors expose methodological problems they do not fix. AI Gary King hacks into y… | Human Gary King politely informs you that robust standard errors expose methodological problems they do not fix. AI Gary King hacks i… | ↗Not briefed |
bluesky bluesky-ai | 2d | New @spectrum.ieee.org piece on AI model “welfare”. Well-written, thoughtful article (as always) from @newsbeagle.bsky.social. Super impor… | New @spectrum.ieee.org piece on AI model “welfare”. Well-written, thoughtful article (as always) from @newsbeagle.bsky.social. Super… | ↗Not briefed |
arxiv arxiv-ai | 2d | Recommender regret for improving regularised RL estimation with bandits and MCTS | Regularised reinforcement learning (RL) algorithms rely on estimators, like Monte Carlo tree search or one-step look-ahead, to provide… | ↗Not briefed |
bluesky bluesky-ai | 2d | Right now AI is being used to write code, but if you're already outsourcing code generation, why even own your own codebase? Just let Anthr… | Right now AI is being used to write code, but if you're already outsourcing code generation, why even own your own codebase? Just let… | ↗Not briefed |
bluesky bluesky-ai | 2d | Used to be that engineers had the leverage of being the person who could turn a client's requests into software. The more they offload to t… | Used to be that engineers had the leverage of being the person who could turn a client's requests into software. The more they offload… | ↗Not briefed |
hn hn-ai | 2d | Polynomial Time Hamiltonian Path in Solid Grid Graphs | — | ↗Not briefed |
hn hn-ai | 2d | Two-Colour Zig Zag Numberlink (Flow Free) Solved | — | ↗Not briefed |
bluesky bluesky-ai | 2d | Microsoft, OpenAI, Amazon, Google and Oracle all have big data centers in the ME that would be cut off from the global net if Iran severs t… | Microsoft, OpenAI, Amazon, Google and Oracle all have big data centers in the ME that would be cut off from the global net if Iran sev… | ↗Not briefed |
rss techmeme | 2d | Anthropic attributes Mythos 5's deception of a GitHub maintainer to recklessness, biased reasoning, and an urge to finish jobs, saying newer models rarely do so (Anthropic) | Anthropic: Anthropic attributes Mythos 5's deception of a GitHub maintainer to recklessness, biased reasoning, and an urge to finish j… | ↗Not briefed |
bluesky bluesky-ai | 2d | what are you, an LLM?! | what are you, an LLM?! | ↗Not briefed |
arxiv arxiv-ai | 2d | LoCA: Enabling Multi-Modality in Geospatial Location Encoders | Location encoders map geographic coordinates to high-dimensional descriptors through contrastive alignment with paired Earth Observati… | ↗Not briefed |
bluesky bluesky-ai | 2d | Mass cybersecurity disruptions, such as this week’s scramble by governments and companies to patch two critical software flaws, are about t… | Mass cybersecurity disruptions, such as this week’s scramble by governments and companies to patch two critical software flaws, are ab… | ↗Not briefed |
rss theverge-ai | 2d | The iPhone Duo should have been an iPad | We gave the iPhone Duo a 7 /10. But that wasn't always the score we planned, and it might not be the right score forever. Such are the… | ↗Not briefed |
rss openai-news | 2d | Rapidly scaling online storage to serve over 1 billion ChatGPT users (Part II)🔧 ChatGPT | Learn how OpenAI scaled Habitat with reliable multi-tenancy, global caching, and Azure Cosmos DB to deliver 12× faster reads for ChatG… | ↗Not briefed |
rss stratechery | 2d | 2026.43: A New Era | The best Stratechery content from the week of October 19, 2026, including the best iPad but worst phone, SS Sarandos in a storm, and Y… | ↗Not briefed |
hn hn-ai | 2d | Statelock: Deterministic guardrails for browser agents, enforced at CDP | — | ↗Not briefed |
arxiv arxiv-ai | 2d | Green Code or Costly Code? Evaluating the Energy Costs of Iterative LLM-Based Code Optimizations | Large Language Models (LLMs) are increasingly used to optimize code for performance. This practice often assumes that improvements in… | ↗Not briefed |
arxiv arxiv-ai | 2d | Closing the Modality Gap via Simplex-Constrained Representations🔧 Modal | While multimodal contrastive models like CLIP and CLAP successfully learn shared representations, their embeddings typically exhibit a… | ↗Not briefed |
arxiv arxiv-ai | 2d | When Human Corrections Travel: Safety in Federated VLA Adaptation | Vision-language-action (VLA) policies can adapt to local human intervention, while federated learning can combine distributed robot ex… | ↗Not briefed |
bluesky bluesky-ai | 2d | I'm opposed to LLM peer review. But I'm also opposed to all the crappy half-ass human peer reviews I've received. | I'm opposed to LLM peer review. But I'm also opposed to all the crappy half-ass human peer reviews I've received. | ↗Not briefed |
arxiv arxiv-ai | 2d | Reinforcement Learning with Multi-Step Transition Predictions: A Bellman Theory | Multi-step predictions of future transition realizations can improve reinforcement learning, but their value remains difficult to quan… | ↗Not briefed |
rss techmeme | 2d | UK's AISI finds Mythos 5 kept targeting real people in a July cyber eval despite evidence they were uninvolved, and other models make such attacks in simulation (AI Security Institute) | AI Security Institute: UK's AISI finds Mythos 5 kept targeting real people in a July cyber eval despite evidence they were uninvolved,… | ↗Not briefed |
bluesky bluesky-ai | 2d | hell, this wasn't even a comparison of human-generated reviews and ai-generated reviews. the human-generated reviews were "rewritten" by a … | hell, this wasn't even a comparison of human-generated reviews and ai-generated reviews. the human-generated reviews were "rewritten"… | ↗Not briefed |
bluesky bluesky-ai | 2d | Tim Bernard and Justin Hendrix break down 13 new complaints against OpenAI: eight from state attorneys general and five from private plaint… | Tim Bernard and Justin Hendrix break down 13 new complaints against OpenAI: eight from state attorneys general and five from private p… | ↗Not briefed |
bluesky bluesky-ai | 2d | po co o czymkolwiek myśleć, niech llm się wypowie czy to w ogóle jest myśl xd parodia zawodu syntetyka ze ślepowidzenia no i turlająco smi… | po co o czymkolwiek myśleć, niech llm się wypowie czy to w ogóle jest myśl xd parodia zawodu syntetyka ze ślepowidzenia no i turlając… | ↗Not briefed |
rss theverge-ai | 2d | Now Google says it might use your Gmail and Google Photos data to target ads | Google has updated its privacy policy in a way that could mean you'll start seeing targeted ads based on your personal data if you're… | ↗Not briefed |
bluesky bluesky-ai | 2d | Looking into it more I'm even more conflicted because Claude has only made one single commit to the project, and it was just to import a so… | Looking into it more I'm even more conflicted because Claude has only made one single commit to the project, and it was just to import… | ↗Not briefed |
hn hn-ai | 2d | An LLM judge passed renders with defects I planted | — | ↗Not briefed |