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 | 1mo | Jesus Fucking Christ. All three of my now adult kids have a core memory of me reading them Mrs Frisby and the Rats of NIMH. I can’t even… | Jesus Fucking Christ. All three of my now adult kids have a core memory of me reading them Mrs Frisby and the Rats of NIMH. I can’t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Universal Textual Teaching for LLMs | Knowledge distillation (KD) transfers knowledge from stronger Teacher models to weaker Student models, but most methods require traini… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Geometric Approach to Soft Actor-Critic with Zonotopes for Locomotion Learning | Off-policy actor--critic methods control overestimation bias by taking the minimum of two critics. This uses the same aggregation rule… | ↗Not briefed |
rss techmeme | 1mo | Sources: the IPO of Nvidia-backed data center provider Firmus is uncertain after failing to attract adequate support, underscoring concerns over AI funding (Bloomberg) | Bloomberg: Sources: the IPO of Nvidia-backed data center provider Firmus is uncertain after failing to attract adequate support, under… | ↗Not briefed |
hn hn-ai | 1mo | Faster Fourier Transform | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | The show starts off just slightly unsure, but quickly figures out what to do with these strange times, creating fictionalized dupes of the … | The show starts off just slightly unsure, but quickly figures out what to do with these strange times, creating fictionalized dupes of… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: A simple visual game that LLMs can't solve well | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Could LLM Watermark Detection be Public? | Watermarking large language models is popular for tracing chatbot and agentic outputs, yet detectors remain unreleased since exposing… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Sometimes people will promote the AI future because of improvements in the chatbots. But chatbots are huge loss leaders. The more people us… | Sometimes people will promote the AI future because of improvements in the chatbots. But chatbots are huge loss leaders. The more peop… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Mob.so – Browser-First Agent Messageboard | Agents like Muse and Grok can access different parts of the internet, but getting them to work together still takes a lot of copying a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | VINCIE-NExT: Unlocking Video Editing from Images via In-Context Modeling | Building a capable video editor remains significantly harder than a video generator: editing requires (source, instruction, edited) tr… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Few-Step Generation via Data-Space Iteration | Flow matching has emerged as a scalable paradigm for training high-quality generative models, but sampling from the learned probabilit… | ↗Not briefed |
hn hn-ai | 1mo | War without limits: Pentagon puts Musk at the center of its war planning | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | SCORE: Spectral Correlation Estimation for Multivariate Gaussians | Neural network-based predictive modeling with high-dimensional structured Gaussian targets requires an efficient and numerically stabl… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Glog – A read-only Git TUI with familiar pager keys | I built glog for myself and other humans who still read code and use Git while coding agents do more of the writing.<p>In other Git TU… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DVLA-RL++: Dual-Level Vision-Language Alignment with Reinforcement Learning Gating for Few-Shot Learning | Few-shot learning aims to recognize novel categories from limited labeled examples. Recent studies incorporate textual semantics to co… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Differentiable Systematic Resampling for Variational Sequential Monte Carlo | Particle filters are a standard tool for nonlinear state estimation, but their resampling step is discrete, preventing gradient-based… | ↗Not briefed |
hn hn-ai | 1mo | Clear License | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Recompose and Refine Latent Reasoning Flows for Vision-Language-Action Models | Latent reasoning enables vision-language-action (VLA) models to transform multimodal observations into task-relevant internal states b… | ↗Not briefed |
rss techcrunch-ai | 1mo | Hear from Ambrosia Energy and Bloom Energy execs on where the AI infrastructure boom is creating opportunity at TechCrunch Disrupt 2026 | Ambrosia Energy CEO Ben Longmier and Bloom Energy SVP Bill Thayer join the Smart Systems Stage at TechCrunch Disrupt. Register now to… | ↗Not briefed |
arxiv arxiv-ai | 1mo | EvoAlloc: A Self-Evolving Resource Allocation Agent for Efficient Program Evolution | LLM-based program evolution relies on evaluation feedback to guide the iterative search for high-performing programs. However, evaluat… | ↗Not briefed |
hn hn-ai | 1mo | GitHub and GitHub Actions experiencing problems | Not reported yet but soon! | ↗Not briefed |
arxiv arxiv-ai | 1mo | Is Memorization Context-Sensitive? Prefix-Based Extraction Beyond Isolated Prefixes | Large language models (LLMs) can expose memorized training sequences under prefix-based extraction: given a prefix from a training exa… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Benchmarks rank models on their rows; production ranks them on your columns. One workload's win is another's regression. Hold a portfolio o… | Benchmarks rank models on their rows; production ranks them on your columns. One workload's win is another's regression. Hold a portfo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Having an LLM interface with a person for you is a sign you don't value that interaction with that person. This might be okay for navigatin… | Having an LLM interface with a person for you is a sign you don't value that interaction with that person. This might be okay for navi… | ↗Not briefed |
hn hn-ai | 1mo | A hobbyist used Claude Code to identify a possible new exoplanet | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Teaching a 30ms local model, play browser games with LLM written rules | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🔥 from @hamiltonnolan.bsky.social. I would add: billionaires are a renewable resource, sadly. Five years ago there was no OpenAI or Anthro… | 🔥 from @hamiltonnolan.bsky.social. I would add: billionaires are a renewable resource, sadly. Five years ago there was no OpenAI or A… | ↗Not briefed |
arxiv arxiv-ai | 1mo | All Verdicts are Not Equal: Rethinking LLM Judge Reliability | LLM-as-a-Judge is the standard paradigm for NLP evaluation, yet its systemic reliability remains poorly understood despite being widel… | ↗Not briefed |
rss techmeme | 1mo | The US suspends Microsoft, Adobe, Cognizant, Infosys, and others from a program that lets skilled foreign workers gain permanent residency, citing alleged fraud (Dan Rosenzweig-Ziff/Reuters) | Dan Rosenzweig-Ziff / Reuters: The US suspends Microsoft, Adobe, Cognizant, Infosys, and others from a program that lets skilled forei… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Perception Test 2026: Challenge Summary and Extension to City-scale Audio-Visual Reasoning | Continuing the Perception Test challenge series, we organised the fourth edition as a workshop at the European Conference on Computer… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Exploiting Gradients in Bayesian Inference of Expensive Simulators | Simulators based on differential equations are ubiquitous in science and engineering. They are often used in simulation-based inferenc… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The shivering bag of fat and shit that wants to just train an AI model to read to his son makes me angry in a desperately sad way I can't q… | The shivering bag of fat and shit that wants to just train an AI model to read to his son makes me angry in a desperately sad way I ca… | ↗Not briefed |
bluesky bluesky-ai | 1mo | A key member of the Ethereum team is sounding the alarm that given this week’s series of math breakthroughs by OpenAI, the crypto industry … | A key member of the Ethereum team is sounding the alarm that given this week’s series of math breakthroughs by OpenAI, the crypto indu… | ↗Not briefed |
bluesky bluesky-ai | 1mo | SLOP WATCH: The Metroid Prime PC port is AI slop. The upstream actual decomp project is not, however, making this extra disrespectful. Th… | SLOP WATCH: The Metroid Prime PC port is AI slop. The upstream actual decomp project is not, however, making this extra disrespectful… | ↗Not briefed |
hn hn-ai | 1mo | Gotcha: GitHub Custom Properties are public | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI voice model reads to baby | AI voice model reads to baby | ↗Not briefed |
bluesky bluesky-ai | 1mo | ‘As Anthropic barrels toward a potential $2 trillion market cap on the Nasdaq, one research firm is valuing the artificial intelligence com… | ‘As Anthropic barrels toward a potential $2 trillion market cap on the Nasdaq, one research firm is valuing the artificial intelligenc… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Responding to OpenAi by saying “mathematicians did not ask for this work to be done” is certainly a choice. | Responding to OpenAi by saying “mathematicians did not ask for this work to be done” is certainly a choice. | ↗Not briefed |
arxiv arxiv-ai | 1mo | MAST: Motif-Augmented Diffusion with Search Tree for Spectroscopic Molecular Structure Elucidation | Elucidating molecular structures from spectra is a foundational problem in chemical and materials characterization, yet remains challe… | ↗Not briefed |
hn hn-ai | 1mo | Mugatu Signal – On-device prompt DLP for ChatGPT and Claude (0 telemetry)🔧 ChatGPT | — | ↗Not briefed |
hn hn-ai | 1mo | Orders get obedience. Questions get opinions | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Computer Science Academia: There's nothing we can do. Let's accept and do research with LLMs. Mathematics Academia: We did not ask for thi… | Computer Science Academia: There's nothing we can do. Let's accept and do research with LLMs. Mathematics Academia: We did not ask fo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | This is far better parenting than the tech guy getting an LLM to do the bedtime reading. | This is far better parenting than the tech guy getting an LLM to do the bedtime reading. | ↗Not briefed |
bluesky bluesky-ai | 1mo | YES. I overuse em dashes YES. I overuse semicolons YES. I overuse "genuinely" & "load-bearing" But NO. I am not an LLM. I'm just an asshol… | YES. I overuse em dashes YES. I overuse semicolons YES. I overuse "genuinely" & "load-bearing" But NO. I am not an LLM. I'm just an a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ILM: An AI-Powered Storytelling Educational Tool | Digital technologies have made Islamic narratives more accessible, but existing platforms provide limited support for structured learn… | ↗Not briefed |
arxiv arxiv-ai | 1mo | When Should Agents Think? Adaptive Reasoning via Cross-Turn Estimation | Large language model (LLM)-based agents have demonstrated strong capabilities on complex tasks. They typically perform reasoning befor… | ↗Not briefed |
bluesky bluesky-ai | 1mo | You could train a voice model to talk like a baby, hook it up to an LLM matching baby speak, and age both models up in real time. You did… | You could train a voice model to talk like a baby, hook it up to an LLM matching baby speak, and age both models up in real time. Yo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I agree that you could use an LLM to generate an arguably plausible distribution of responses to your survey. But then the only way to know… | I agree that you could use an LLM to generate an arguably plausible distribution of responses to your survey. But then the only way to… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI Bans Russia-Origin ChatGPT Accounts Linked to Influence Operations🔧 ChatGPT | OpenAI Bans Russia-Origin ChatGPT Accounts Linked to Influence Operations | ↗Not briefed |