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 |
|---|---|---|---|---|
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 |
bluesky bluesky-ai | 2d | Today in AI coding: Claude finds a year's worth of bugs in one week verifying 100+ libraries. Plus Claude Duet's waitlist, Claude on Google… | Today in AI coding: Claude finds a year's worth of bugs in one week verifying 100+ libraries. Plus Claude Duet's waitlist, Claude on G… | ↗Not briefed |
bluesky bluesky-ai | 2d | tip to toe this study was designed and executed to make ai-generated text appear superior to human evaluations the metrics generated that … | tip to toe this study was designed and executed to make ai-generated text appear superior to human evaluations the metrics generated… | ↗Not briefed |
arxiv arxiv-ai | 2d | NgramRL: Accelerating RL Post-Training for Structured Generation via Rollout-Aware N-gram Speculative Decoding | Reinforcement learning (RL) post-training improves the reasoning capabilities of large language models, but its cost is dominated by a… | ↗Not briefed |
bluesky bluesky-ai | 2d | In their first interview since being fired, OpenAI safety researchers Mikita Balesni and Tomek Korbak take us inside the Hugging Face hack …🔧 Hugging Face | In their first interview since being fired, OpenAI safety researchers Mikita Balesni and Tomek Korbak take us inside the Hugging Face… | ↗Not briefed |
hn hn-ai | 2d | The model won't devour everything: three primitives of agent systems | — | ↗Not briefed |
arxiv arxiv-ai | 2d | Advancing SVD-based LLM Compression via Layer-Wise Error Model Search | Low-rank SVD-based compression offers a powerful strategy to reduce the computational costs of LLMs. However, existing methods face tw… | ↗Not briefed |
bluesky bluesky-ai | 2d | I was told the scary computer security warnings from Anthropic et al were just marketing. Or was it that this wouldn’t be happening if they… | I was told the scary computer security warnings from Anthropic et al were just marketing. Or was it that this wouldn’t be happening if… | ↗Not briefed |
arxiv arxiv-ai | 2d | Trusting the Tool: Asymmetric Deference of LLM Vulnerability Auditors to Static-Analysis Evidence🔧 Replicate | Pipelines that combine large language models (LLMs) with static analysis let the model consume the analyser's output as evidence, but… | ↗Not briefed |
bluesky bluesky-ai | 2d | as a solo hobby vibe coder that I think has a healthy view and relationship with the chatbots: even the best of them have an affinity towar… | as a solo hobby vibe coder that I think has a healthy view and relationship with the chatbots: even the best of them have an affinity… | ↗Not briefed |
arxiv arxiv-ai | 2d | Do "True" and "False" Add Up to One? Complement Consistency in LLM-Agent Opinion Simulations | In LLM-agent opinion simulations, each agent reports the probability that a claim is true after seeing its own previous probability an… | ↗Not briefed |
arxiv arxiv-ai | 2d | J-Comm: Jacobian-Guided Latent Communication between LLM Agents🔧 Consensus | Latent communication requires both an effective continuous reasoning process and an interface through which an external model can leve… | ↗Not briefed |
arxiv arxiv-ai | 2d | From Deliberately Constructed Systemic Representations of the World of Activity to Dialogue: A Five-Level Communication Architecture for Strong Language-Based ASI | This paper proposes a five-level organization of message formation and understanding within thinking and activity in which ASI partici… | ↗Not briefed |
bluesky bluesky-ai | 2d | What I do know is that, if AI doesn't work out, Blackstone will be in a world of pain. It has made too many bets, too many of them being … | What I do know is that, if AI doesn't work out, Blackstone will be in a world of pain. It has made too many bets, too many of them b… | ↗Not briefed |
bluesky bluesky-ai | 2d | Blackstone's data center assets have swelled to over $200bn (larger than its traditional infrastructure holdings) and it has become an inv… | Blackstone's data center assets have swelled to over $200bn (larger than its traditional infrastructure holdings) and it has become a… | ↗Not briefed |
bluesky bluesky-ai | 2d | Blackstone's largely wound down his residential property business, in part because it's realized that there's far more money to be made in … | Blackstone's largely wound down his residential property business, in part because it's realized that there's far more money to be mad… | ↗Not briefed |
bluesky bluesky-ai | 2d | Mistakes happen when you're big, and Blackstone is very big indeed. It's the world's largest alternative asset manager, the world's largest… | Mistakes happen when you're big, and Blackstone is very big indeed. It's the world's largest alternative asset manager, the world's la… | ↗Not briefed |
arxiv arxiv-ai | 2d | Interpretable sparse fine-tuning with TopKLoRA enables faithful circuit discovery | Fine-tuning on datasets of unknown origin can create a backdoor in the trained model when the data includes poisoned examples. Difficu… | ↗Not briefed |
hn hn-ai | 2d | H2oai/jdupes: H2O.ai fork of jdupes | — | ↗Not briefed |
bluesky bluesky-ai | 2d | i see nothing about LLM use, just deep learning and ML utilization, this is a nothingburger about davinci using cringe rewording to name th… | i see nothing about LLM use, just deep learning and ML utilization, this is a nothingburger about davinci using cringe rewording to na… | ↗Not briefed |
bluesky bluesky-ai | 2d | they believed this shit going in; it's what motivated them to work in the field, in many ways it's *why they are competent at frontier-lab … | they believed this shit going in; it's what motivated them to work in the field, in many ways it's *why they are competent at frontier… | ↗Not briefed |
bluesky bluesky-ai | 2d | there's a temptation with Anthropic particularly to be like "oh the doomhowling is product advertisement" but no, they earnestly believe th… | there's a temptation with Anthropic particularly to be like "oh the doomhowling is product advertisement" but no, they earnestly belie… | ↗Not briefed |
arxiv arxiv-ai | 2d | PoKER: A Probability of Kill Estimation Rate Model for Air-to-Air Missiles Using Machine Learning on Stochastic Targets | This work introduces PoKER, a novel probabilistic model engineered to optimize missile launch effectiveness in air-to-air scenarios, s… | ↗Not briefed |
arxiv arxiv-ai | 2d | A Foundation Model for Policy In-Context Longitudinal Causal Effect Estimation | Longitudinal causal effect estimation is complicated by time-varying confounding and treatment-confounder feedback loops. Existing met… | ↗Not briefed |
hn hn-ai | 2d | Getting Started with Lean 4: Proofs as Code for Real Developers | — | ↗Not briefed |
bluesky bluesky-ai | 2d | Raised in a council house, Ian Rankin is still attuned to the effects of inequality. The bestselling novelist talks to Graeme Green about h… | Raised in a council house, Ian Rankin is still attuned to the effects of inequality. The bestselling novelist talks to Graeme Green ab… | ↗Not briefed |
arxiv arxiv-ai | 2d | Sustainability of TSFM Deployment and Pretraining: From Carbon Footprint Analysis to Mitigation via Pruning | Time Series Foundation Models (TSFMs) offer strong zero-shot forecasting capabilities, but their high pretraining cost raises a practi… | ↗Not briefed |
hn hn-ai | 2d | AI Safety Is Mostly a Sex Cult in Berkeley, California | — | ↗Not briefed |
bluesky bluesky-ai | 2d | I know that it’s popular, or was for a while to say you had never once used an LLM, but for over a year now I’ve been on a little personal … | I know that it’s popular, or was for a while to say you had never once used an LLM, but for over a year now I’ve been on a little pers… | ↗Not briefed |
bluesky bluesky-ai | 2d | I know I will get crucified for saying this, but I feel we're reaching "thought crime" levels of discourse when it comes to using-free-soft… | I know I will get crucified for saying this, but I feel we're reaching "thought crime" levels of discourse when it comes to using-free… | ↗Not briefed |
arxiv arxiv-ai | 2d | Regularised RL improves sim-to-real transfer | Regularised reinforcement learning (RL) has provable robustness guarantees that should ease the transfer of policies from simulation t… | ↗Not briefed |
bluesky bluesky-ai | 2d | openai: you can’t do that it’s scawwy 😰 deepseek: fuck it, we ball | openai: you can’t do that it’s scawwy 😰 deepseek: fuck it, we ball | ↗Not briefed |
arxiv arxiv-ai | 2d | Singular Geometry of Low-Rank Adaptation | Low-rank adaptation (LoRA) is now a standard approach to parameter-efficient fine-tuning of large neural networks. Its low-rank factor… | ↗Not briefed |
hn hn-ai | 2d | Open Source Mobile-Use MCP and AI Agent | — | ↗Not briefed |
bluesky bluesky-ai | 2d | In a new report, Anthropic found that once a Mythos 5 agent had begun an attack, there were basically zero scenarios in which it could be c… | In a new report, Anthropic found that once a Mythos 5 agent had begun an attack, there were basically zero scenarios in which it could… | ↗Not briefed |
arxiv arxiv-ai | 2d | Near-Optimal List Replicability from Littlestone Dimension | We study the list replicability number in the realizable PAC learning framework and prove that every concept class $H$ over a countabl… | ↗Not briefed |
bluesky bluesky-ai | 2d | The autonomous and rapid prodding that agents are able to do means we're basically cooked. They don't need Claude, OpenAI, or anybody to do… | The autonomous and rapid prodding that agents are able to do means we're basically cooked. They don't need Claude, OpenAI, or anybody… | ↗Not briefed |
hn hn-ai | 2d | Anthropic want work patterns to change | — | ↗Not briefed |
hn hn-ai | 2d | GitHub to Forgejo: what to get right before a one-afternoon switch | — | ↗Not briefed |
hn hn-ai | 2d | Ask HN: Claude gets over my 5h limit today in 10 minutes? | I do everything I did before. Frequent compact and clear of session on new tasks. This behavior let me work for around 4 hours without… | ↗Not briefed |
arxiv arxiv-ai | 2d | Reimagining Enterprise RAG for Multi-Agent Systems: A Knowledge Layer That Organizes Before It Retrieves | Enterprises are deploying agentic AI across the software development lifecycle, and they need the behavioral logic of the systems they… | ↗Not briefed |
arxiv arxiv-ai | 2d | Neural LoFi with Backward Coupling: A Spectral Theory of Cross-Layer Feature Learning | During joint training, backpropagation allows later layers to reshape earlier representations, a feedback mechanism absent from purely… | ↗Not briefed |
bluesky bluesky-ai | 2d | Launched this week on TrustList: MiniMax M3.1 Flash MiniMax's low-cost model for agents, coding and computer use, released 19 October 2026… | Launched this week on TrustList: MiniMax M3.1 Flash MiniMax's low-cost model for agents, coding and computer use, released 19 October… | ↗Not briefed |
bluesky bluesky-ai | 2d | ASUS ROG Xbox Ally X and X20 receive FSR 4.1 preview driver | ASUS ROG Xbox Ally X and X20 receive FSR 4.1 preview driver | ↗Not briefed |
bluesky bluesky-ai | 2d | Like, is the biggest threat to cybersecurity the mere existence of LLMs, or is it the existence of companies like OpenAI with basically unl… | Like, is the biggest threat to cybersecurity the mere existence of LLMs, or is it the existence of companies like OpenAI with basicall… | ↗Not briefed |
hn hn-ai | 2d | Show HN: Koala – Run AI Agents in Isolated Linux VMs on macOS | — | ↗Not briefed |