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 | Foresight-over-Graph: Reasoning Beyond Local Horizons for Knowledge Base Question Answering | Large language models (LLMs) have demonstrated strong capabilities in question answering, yet they still frequently suffer from halluc… | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: Show me your agentic SDE semantics | I am considering describing a SDE role for my agent to load in at context opening.<p>Not in one file, but in many.<p>I have seen peopl… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The LLM alignment problem is mostly fake, but the cat alignment problem very much is not | The LLM alignment problem is mostly fake, but the cat alignment problem very much is not | ↗Not briefed |
bluesky bluesky-ai | 1mo | on the one hand its fuck the police every day and fuck anthropic every minute but on the other hand just GET AN ACTUAL DIARY, ARE YOU PROMP… | on the one hand its fuck the police every day and fuck anthropic every minute but on the other hand just GET AN ACTUAL DIARY, ARE YOU… | ↗Not briefed |
hn hn-ai | 1mo | The Internet Runs on Cats and Porn. It's About to Get Cattier and Pornier | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLM vendors are like an arsonist who lights up a couch and then writes a column in the local newspaper about how we should worry about the … | LLM vendors are like an arsonist who lights up a couch and then writes a column in the local newspaper about how we should worry about… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Decision-Focused Learning in MDPs: An Occupancy Measure Approach | In this work, we consider decision-focused learning (DFL) for a Markov decision process (MDP), where existing methods differentiate th… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: I turned Andrej Karpathy's tips on understanding LLM output into Skill | I turned Andrej Karpathy’s tips on understanding LLM output into an Open Source Agent Skill : [make-it-click](<a href="https:/&#x… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I mean lol like | I mean lol like | ↗Not briefed |
hn hn-ai | 1mo | OpenAI's B2B Marketplace: The Hyperscaler of AI Apps | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Actually one of the interesting problems with Talkie (which is that LLM. it already exists) is that it does get future bleed and knows who … | Actually one of the interesting problems with Talkie (which is that LLM. it already exists) is that it does get future bleed and knows… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Authrot – a writing studio that maps your characters, places and plot | Hi HN, I'm Joe.<p>I built Authrot, because Google Docs and Word slowed to a crawl on my phone once my story passed about 25 chapt… | ↗Not briefed |
hn hn-ai | 1mo | CharacterMemory: Memory System for AI Characters | — | ↗Not briefed |
rss techcrunch-ai | 1mo | Pinterest’s AI now turns beauty Pins into action plans | Pinterest’s new AI-powered Beauty Guides translate hair and nail Pins into salon terminology, with estimated costs, appointment times… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Wazzu partnering with an LLM to write for the team site; opposed to just using the writers you have on staff. Credit Greg Woods/Spokesman… | Wazzu partnering with an LLM to write for the team site; opposed to just using the writers you have on staff. Credit Greg Woods/Spok… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Tired: Some thoughts should be inside thoughts Wired: Some plots need to be inside plots | Tired: Some thoughts should be inside thoughts Wired: Some plots need to be inside plots | ↗Not briefed |
rss techcrunch-ai | 1mo | Get all your questions answered at TechCrunch Disrupt 2026: The full breakout session agenda revealed | Get all of your scaling and tech questions answered at TechCrunch Disrupt 2026. Here's the full breakout session agenda, happening Oct… | ↗Not briefed |
hn hn-ai | 1mo | Mistral to Release New AI Model to Better Compete with U.S. Rivals | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | It is possibly M-L's opportunity to shortcut right to good things. I've written about UBI, and UBI is now both more possible and more nece… | It is possibly M-L's opportunity to shortcut right to good things. I've written about UBI, and UBI is now both more possible and more… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: We benchmarked AST stubbing trade-offs on Airflow and Zulip (ctxfw) | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | lot of “THIS JOB WILL BE DISPLACED BY AI” stuff doesn’t answer the question of “who has legal liability if a mistake is made?” Google is g… | lot of “THIS JOB WILL BE DISPLACED BY AI” stuff doesn’t answer the question of “who has legal liability if a mistake is made?” Google… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Tested. Working. I am now running Qwen 3.8 Flash Next, a 120b model, locally at 45 tps thanks to Strata. This is certifiably insane. It's…🔧 Qwen 3 | Tested. Working. I am now running Qwen 3.8 Flash Next, a 120b model, locally at 45 tps thanks to Strata. This is certifiably insane.… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Accelerating the Development of PLGA In Situ Forming Depots Through AI-Driven Multi-Objective Optimization | Developing long-acting injectable formulations requires the simultaneous optimization of drug loading, release kinetics, viscosity, in… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evolutionary One-Step Generators: Fast and Diverse Sampling for Discrete Design | Several discrete design tasks, such as molecular discovery, require diverse collections of useful candidates at low computational cost… | ↗Not briefed |
bluesky bluesky-ai | 1mo | “The publisher of Wikipedia said Monday that OpenAI agents attempted to hack a note-taking tool it hosts, made unauthorized edits, and sent… | “The publisher of Wikipedia said Monday that OpenAI agents attempted to hack a note-taking tool it hosts, made unauthorized edits, and… | ↗Not briefed |
hn hn-ai | 1mo | Neon//Ping Latency Tool | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Transect: Retaining Observability for Long-Horizon LLM Agent Evaluations | Frontier AI evaluations increasingly use open-ended, agentic, long-horizon tasks whose transcripts can span hundreds of pages of outpu… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Explainable Failure Prediction and Prevention in Maritime | Maritime systems operate in highly dynamic environments where unexpected equipment failures can compromise safety, reliability, and op… | ↗Not briefed |
hn hn-ai | 1mo | Voice assistant that runs on your own Claude or ChatGPT plan🔧 ChatGPT | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | DeepSeek lève plus de 10 milliards d’euros lors d’un nouveau tour de table #DeepSeek #levéeDeFonds #investissement #entrepreneuriat #startu… | DeepSeek lève plus de 10 milliards d’euros lors d’un nouveau tour de table #DeepSeek #levéeDeFonds #investissement #entrepreneuriat #s… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Test-Time Adaptation of Quantized ViTs via Single-Pass Quantizer-Aligned Recalibration | Post-training quantization is a standard route to fitting vision transformers (ViTs) into edge compute and memory budgets, yet quantiz… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Поигрался с новой ии моделью Jev и переписал тулзу для оценки качества перевода с ней вместо llm as a judge. И короче выяснилось что эта … | Поигрался с новой ии моделью Jev и переписал тулзу для оценки качества перевода с ней вместо llm as a judge. И короче выяснилось что… | ↗Not briefed |
hn hn-ai | 1mo | GPT 6 Astra as an Embodied Policy | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Sensor Geometry as a Flow-Matching Prior for Multi-Channel Brain Signals | Flow-matching models start from an isotropic Gaussian source, the standard choice when the correlation structure of the data is unknow… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The em dashes also kinda make it seem like he queried an LLM to determine the severity of injury one would suffer from getting stone cold s… | The em dashes also kinda make it seem like he queried an LLM to determine the severity of injury one would suffer from getting stone c… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ai psychosis? not me. claude says im special. different than the rest of you. more than mortal. immune to the trappings of humanity. a god?… | ai psychosis? not me. claude says im special. different than the rest of you. more than mortal. immune to the trappings of humanity. a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Uncertainty Quantification Is Indispensable for Reliable Connectome-Based Graph Learning: A Narrative Review and Case Study | While graph neural networks (GNNs) have shown substantial promise in connectome-based diagnostic classification, deterministic models… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Love the Le Chonk Memes, but its (list most mistral models) still horrible value... The french are getting mogged, both on intelligence an… | Love the Le Chonk Memes, but its (list most mistral models) still horrible value... The french are getting mogged, both on intelligen… | ↗Not briefed |
bluesky bluesky-ai | 1mo | One of those UAE funds is MGX, which— 💰️ is chaired by the country's National Security Adviser, Sheikh Tahnoon bin Zayed Al Nahyan 💰️ … | One of those UAE funds is MGX, which— 💰️ is chaired by the country's National Security Adviser, Sheikh Tahnoon bin Zayed Al Nahyan… | ↗Not briefed |
arxiv arxiv-ai | 1mo | How Much Planning Is Enough? Reducing Search and Computation in World-Model Planning | Visual world models enable goal-directed control through decision-time action search, but their deployment efficiency is often limited… | ↗Not briefed |
arxiv arxiv-ai | 1mo | High-Dimensional Statistical Inference for Sparse Support Vector Machines | Using a replica-symmetric high-dimensional characterization, we develop an inferential framework for sparse support vector machines wh… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Mistral said Le Chonk - officially named Mistral Large 4 - will be made fully public on October 27." How is this real text from a real Reu…🔧 Mistral Large | "Mistral said Le Chonk - officially named Mistral Large 4 - will be made fully public on October 27." How is this real text from a rea… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Last Reskeet: I think it's important to be able to differentiate what is Machine Learning and what is an LLM and I'll never forgive the AI… | Last Reskeet: I think it's important to be able to differentiate what is Machine Learning and what is an LLM and I'll never forgive t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DIPrune: Task-Aware Token Pruning with Dual Importance for Efficient Multimodal Language Models | Recent training-free pruning approaches for Multimodal Large Language Models (MLLMs) effectively cut computational overhead by exploit… | ↗Not briefed |
hn hn-ai | 1mo | How HN: Duck Out – webcam duck hunt you play with your hands | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Machine Learning for German Redispatch Forecasting under Data Delays and Temporal Distribution Shift | Public redispatch records provide empirical data for grid congestion forecasting, but delayed reporting, zero-inflated distributions,… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Vev – Ask questions about a screenshot, get a probability per answer | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | The end goal of all of this for AI guys is to have an LLM that generates a wife to which they can then say "LLM, please fuck my wife good,"… | The end goal of all of this for AI guys is to have an LLM that generates a wife to which they can then say "LLM, please fuck my wife g… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ”Pursuing a tech path that can never support the guarantees of safety that we will need when systems become sufficiently capable to be of r… | ”Pursuing a tech path that can never support the guarantees of safety that we will need when systems become sufficiently capable to be… | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: What models and harnesses are you using that are not Claude or Codex? | — | ↗Not briefed |