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 | 1mo | Show HN: Almostright.ai – a parody – AI that never asks why | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | The Anthropic one where they told it to upload malware to PyPI and forgot to turn on the fake internet or whatever is the funniest example … | The Anthropic one where they told it to upload malware to PyPI and forgot to turn on the fake internet or whatever is the funniest exa… | ↗Not briefed |
bluesky bluesky-ai | 1mo | اول حكايتين النهاردة من #حكايات_عمور هيكونوا من عالم Claude كلود هو برنامج ذكاء صناعي بس مميز جدا في الناحية الاكاديمية لدرجة انه بيتعمل بي… | اول حكايتين النهاردة من #حكايات_عمور هيكونوا من عالم Claude كلود هو برنامج ذكاء صناعي بس مميز جدا في الناحية الاكاديمية لدرجة انه بيتع… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It took nearly 1,000 Claude agents 21 hours and 210 million tokens to uncover the ‘novel enzyme system.’ www.theverge.com/ai-artificia... | It took nearly 1,000 Claude agents 21 hours and 210 million tokens to uncover the ‘novel enzyme system.’ www.theverge.com/ai-artificia… | ↗Not briefed |
rss wired-ai | 1mo | The Pope’s AI Guy Is Worried About ‘Cartel’ Behavior Among Big Labs | Paolo Benanti tells WIRED that hysteria over whether godlike AI could destroy humanity is distracting from the need for public debate… | ↗Not briefed |
hn hn-ai | 1mo | ChatGPT turns 4 soon. We moved on from GPT but never let go of Chat🔧 ChatGPT | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Training Object Permanence in World Models | Object permanence and solidity are hallmarks of human cognitive priors. Recent studies show that video generation models, a paradigmat… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Fellowship of the Query: Learning Retrieval Actions | Retrieval-augmented question answering requires control decisions about when to decompose a question, search, reformulate, extract evi… | ↗Not briefed |
hn hn-ai | 1mo | Squeezer – Use your idle Claude plan window to work through TODOs overnight | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Actually ... I can use Overlands as front-end for my agents this way. They simply get an atproto-account to log into #Overlands. And for pr… | Actually ... I can use Overlands as front-end for my agents this way. They simply get an atproto-account to log into #Overlands. And f… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RLVR landscapes for iterated multiplications can be benign: Insights from spin-glass theory | Despite the importance of reinforcement learning with verifiable rewards (RLVR), the extent to which it can learn new reasoning capabi… | ↗Not briefed |
rss theverge-ai | 1mo | Anthropic’s biolab made a discovery it’s comparing to Crispr | Anthropic says its AI Claude has "autonomously discovered" a new enzyme system similar to machinery behind the powerful gene-editing t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | On the Diffusibility of High-Dimensional Latents | Representation Autoencoders (RAEs) enable diffusion models to operate in the feature spaces of pretrained visual encoders. However, ma… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Contrastive Learning for Authorship Verification | Our results show that contrastive learning outperforms a classification-based approach to authorship verification under the tested set… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 7,5K followers maintenant. Et @radiofrance.fr qui lui a fait une pub avec un article qui reprend ses dires sans aucun contradictoire ni con… | 7,5K followers maintenant. Et @radiofrance.fr qui lui a fait une pub avec un article qui reprend ses dires sans aucun contradictoire n… | ↗Not briefed |
arxiv arxiv-ai | 1mo | StudentBench: AI and human tutoring yield equivalent GRE learning gains | Artificial intelligence offers an unprecedented opportunity to augment human capabilities, yet progress at the frontier has focused pr… | ↗Not briefed |
rss the-decoder | 1mo | ChatGPT Voice gets closer to "Her" with email, calendar, and Slack access🔧 ChatGPT | ChatGPT Voice now runs on OpenAI's new GPT-6 Astra, Sol, and Luna models and can tap into plugins like email, calendar, and Slack. Use… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Where Should I Join? Robot Group Joining via Language-Guided Goal Prediction | Social navigation typically assumes a specified goal and focuses on reaching it while respecting social conventions, whereas robot gro… | ↗Not briefed |
hn hn-ai | 1mo | Re: Skillsync | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Reward Hacking Challenges Oversight of Autonomous Research Agents | Autonomous research agents can design experiments, evaluate results, and write reports, giving them control over both a scientific res… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The code they write is dogshit. It's automated technical debt with built in landmines. You think people coding in C causes issues? Wait'll … | The code they write is dogshit. It's automated technical debt with built in landmines. You think people coding in C causes issues? Wai… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Even Sharper Bounds for Transductive Learning and Its Applications | We introduce Sharper Transductive Local Complexity (STLC), a localized complexity method for transductive learning under uniform sampl… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Reflex – a GGUF/CUDA inference engine tuned for cold-start latency | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Then it played with other kinds of AI before landing on claude because of why's suggestion. That was a leap forward. We formalized VDD and … | Then it played with other kinds of AI before landing on claude because of why's suggestion. That was a leap forward. We formalized VDD… | ↗Not briefed |
bluesky bluesky-ai | 1mo | new chatgpt voice is far more capable & agentic🔧 ChatGPT | new chatgpt voice is far more capable & agentic | ↗Not briefed |
hn hn-ai | 1mo | Not reading the code cannot become the new norm | i get that a lot of these people out here using llms, especially if you go to places like twitter dot com, want to live in this fantas… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Crest – Answer Claude Code approvals from your MacBook's notch | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Of course, IKEA was always such a natural fit for bench marking. | Of course, IKEA was always such a natural fit for bench marking. | ↗Not briefed |
arxiv arxiv-ai | 1mo | Decision Hijacking: Prompt Injection Attacks on Jev's Typed Probabilistic Decisions | Most studies of prompt injection focus on generative agents, leaving their effects on models with schema-defined outputs unclear. We e… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Can LLMs Reason About Runtime Behavior? A Repository-Level Dynamic Benchmark🔧 FLUX | Large language models (LLMs) are increasingly used in coding tasks, but their ability to reason about code execution remains unclear.… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Nonequilibrium Phases of Repulsive Self-Attention: Chaos, Attention Condensation, and Emergent Locality🔧 Consensus | We study the nonequilibrium dynamics of a minimal recurrent transformer with $N$ normalized tokens, $Q=K=I$, and a negative value map… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The old cybersecurity model is breaking | The old cybersecurity model is breaking | ↗Not briefed |
bluesky bluesky-ai | 1mo | Our podcast on science and technology. This week, the strange new science of model psychology could make AI safer | Our podcast on science and technology. This week, the strange new science of model psychology could make AI safer | ↗Not briefed |
bluesky bluesky-ai | 1mo | This has been happening at the archives I work with too! We’ve had researchers ask for super specific things, then when they come in for an… | This has been happening at the archives I work with too! We’ve had researchers ask for super specific things, then when they come in f… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic is making a wet lab. If they want to maintain the permanent Mandate of Heaven and also pare back AI backlash, have your models m… | Anthropic is making a wet lab. If they want to maintain the permanent Mandate of Heaven and also pare back AI backlash, have your mod… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Order-Invariant Answers, Order-Sensitive Representations in Mathematical Reasoning | Reordering a set of mathematical rules without changing its meaning should preserve the correct answer, but must a model's internal re… | ↗Not briefed |
rss the-decoder | 1mo | Google's new Flash TTS models let you design AI voices from scratch using text descriptions | Google is introducing two new text-to-speech models, Gemini 3.8 Flash TTS and Flash-Lite TTS, which support more than 100 languages. F… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I don't know if this gets talked about too much, but we are starting to get a lot of archive requests that feel like LLM-generated research… | I don't know if this gets talked about too much, but we are starting to get a lot of archive requests that feel like LLM-generated res… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think people expect this to be bad in a different way because we've learned how to spot the problems with AI videos, but this is a js ani… | I think people expect this to be bad in a different way because we've learned how to spot the problems with AI videos, but this is a j… | ↗Not briefed |
hn hn-ai | 1mo | Big dumb moron explains Jev in March 2025 – and then does nothing about it | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Minimal-Norm Univariate Two-Layer ReLU Classification: Exact Solutions and Global Optimality with Skip Connections | We study minimal-norm interpolation and $\ell_2$-regularized logistic-loss minimization for binary classification by univariate two-la… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I'm a masochist so I clicked on a Hacker News thread about OpenAI doing self-driving that predictably also attracted Elon fans. About as ba… | I'm a masochist so I clicked on a Hacker News thread about OpenAI doing self-driving that predictably also attracted Elon fans. About… | ↗Not briefed |
hn hn-ai | 1mo | GPT-6 Sol might acutally be GPT-6 Terra | — | ↗Not briefed |
hn hn-ai | 1mo | Cortera Framework:agentic framework with containment for rogue agents and action | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Hntui – Another Hacker News TUI (Built on OpenTUI) | Please check it out! I use it a lot and have a lot of fun using it.<p>I recently effect-ified it to learn more about the Effect librar… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Cross-Scale Transfer Learning for Depression Severity Prediction: From PHQ-8 to HAMD-17 Across Languages and Clinical Paradigms | This work addresses continuous depression-severity score prediction from clinical interview transcripts under data scarcity. We propos… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I love this headline. www.404media.co/people-train... | I love this headline. www.404media.co/people-train... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Context-Continuous Preference Learning for Exoskeleton Personalization | Personalizing exoskeleton assistance across operating conditions is constrained by the time and physical effort required to collect us… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Repairability of Inexact Solvers in Recursive State Estimation with Machine Learning | Recursive state estimation often executes approximate numerical solutions inside a feedback loop, where highly accurate local steps do… | ↗Not briefed |
hn hn-ai | 1mo | Ontological Language Models (OLMs) | — | ↗Not briefed |