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 | Why doesn't she try to reproduce it with an open model? Or at least ask if someone did something like that? I think she just got into a co… | Why doesn't she try to reproduce it with an open model? Or at least ask if someone did something like that? I think she just got into… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RiboUnmix: Learning Shared Translational Dynamics from Biased and Noisy Ribo-seq Measurements | Ribosome profiling (Ribo-seq) measures ribosome distributions along mRNAs, but observed occupancy profiles also contain experiment-spe… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Jesus christ what a fuckin tool | Jesus christ what a fuckin tool | ↗Not briefed |
hn hn-ai | 1mo | Show HN: A neural network that trains in the browser to play Clash Royale | — | ↗Not briefed |
hn hn-ai | 1mo | OpenAI Is Paying Me to Move My Business to Anthropic | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Zero-Compute Cross-Lingual Transferability Estimation Using Typological Feature Proxies | Cross-lingual transfer describes how knowledge in a source language benefits a target language. Measuring it quantitatively requires b… | ↗Not briefed |
bluesky bluesky-ai | 1mo | (Besides which, do you want some drone in sector 7-G at Anthropic reading about your "Barney cucks Fred with Wilma" fantasy? Especially whe… | (Besides which, do you want some drone in sector 7-G at Anthropic reading about your "Barney cucks Fred with Wilma" fantasy? Especiall… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Wn – find the files a bug touches, trained on 1.1M real fixes | — | ↗Not briefed |
hn hn-ai | 1mo | Value Noise Derivatives | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Marginal Response Surface Elicitation for Zero-Label Tabular Learning🔧 Elicit | Tabular learning uses structured data to predict target outcomes. Traditionally, this process has relied on labeled data. However, lar… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Now the humans are hallucinating. (Claude has no such mode.) | Now the humans are hallucinating. (Claude has no such mode.) | ↗Not briefed |
hn hn-ai | 1mo | I Turned 17 Hours of Video into a Second Brain with Local LLM | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | This Claude AI vinyl mode thing got me thinking about the use of records and turntables in commercials etc as short hand for cool…record co… | This Claude AI vinyl mode thing got me thinking about the use of records and turntables in commercials etc as short hand for cool…reco… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SAGE: Salient Factor Discovery and Generation with Visual Foundation Representations | Given a target dataset, such as faces with eyeglasses, and a background dataset, such as faces without, contrastive analysis separates… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Free Everywhere, Exact on Trees: PPO's Dropped Correction Buys Sample Efficiency Under Aggressive Reuse | Common policy improvement methods, including TRPO, PPO, and GRPO, estimate policy improvement under the behavioral policy's state-visi… | ↗Not briefed |
rss techmeme | 1mo | Tinder rolls out Group Hangouts in select US markets, letting a user form a group of three to 15 friends around shared activities, as it courts younger users (Samantha Kelly/Bloomberg) | Samantha Kelly / Bloomberg: Tinder rolls out Group Hangouts in select US markets, letting a user form a group of three to 15 friends a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "I was really struggling to do _____ and then I found [AI model]—" the proper diagnosis of the situation is, unfortunately, that you're laz… | "I was really struggling to do _____ and then I found [AI model]—" the proper diagnosis of the situation is, unfortunately, that you'r… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Towards Better Exploration in Sequential Test-Time Scaling | Test-time scaling improves language model reasoning by spending additional compute at inference. However, both classes of existing met… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ai this, gpt that, wait till an LLM can chase a house lizard out of my house without me lifting a finger, and then we'll talk. i lost 4… | ai this, gpt that, wait till an LLM can chase a house lizard out of my house without me lifting a finger, and then we'll talk. i l… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PEG-Tab: Sampling-Time Record Repair and Release Control for Tabular Synthesis | Pretrained tabular generators can reproduce training records even when aggregate utility remains high. When retraining is unavailable… | ↗Not briefed |
bluesky bluesky-ai | 1mo | since anthropic posted this taking this with a huge grain of salt but their glm-5.3 investigation is prima facie pretty fuckin terrifying! | since anthropic posted this taking this with a huge grain of salt but their glm-5.3 investigation is prima facie pretty fuckin terrify… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Certification-Based Differentially Private Learning | Differential privacy (DP) in machine learning is typically achieved by adding noise to model parameters (private learning) or to model… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MIND: Marginal-Invariant Neural Dependency Diffusion for Mixed-Type Tabular Generation | This paper proposes MIND, a marginal-invariant neural dependency diffusion model for mixed-type tabular data. MIND does not directly l… | ↗Not briefed |
hn hn-ai | 1mo | For the love of god I just want the LLM | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Same day Fuckface signed his one page AI pact calling it a constitution, the Times reported OpenAI ignored internal warnings before its mod… | Same day Fuckface signed his one page AI pact calling it a constitution, the Times reported OpenAI ignored internal warnings before it… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Parameterization method of reservoir properties for ensemble-based data assimilation using intermediate latent space of StyleGAN | Ensemble smoothers are the most successful and efficient techniques currently available for history matching. However, because these m… | ↗Not briefed |
arxiv arxiv-ai | 1mo | D-Scope: Decomposing and Steering Diffusion Transformers with Sparse Autoencoders | Sparse autoencoders (SAEs) reveal visual structure in diffusion transformers (DiTs), but interpreting a feature does not establish whe… | ↗Not briefed |
hn hn-ai | 1mo | Transcoding UTF-8 to UTF-16 with replacement at gigabytes per second | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Heads up for book authors: I’ve stumbled into a LLM-powered scam that invites authors for seemingly journalistic interviews, then asks for… | Heads up for book authors: I’ve stumbled into a LLM-powered scam that invites authors for seemingly journalistic interviews, then ask… | ↗Not briefed |
hn hn-ai | 1mo | CaveCode – why input many code token when few do trick | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | This kind of thinking does not square with how the harnessing that is typical of AI agents these days is not considered part of the LLM | This kind of thinking does not square with how the harnessing that is typical of AI agents these days is not considered part of the LLM | ↗Not briefed |
arxiv arxiv-ai | 1mo | Hybrid Methods for Robust Tabular Data Imputation | Missing data are a fundamental challenge in statistical analysis and machine learning, as the choice of imputation method substantiall… | ↗Not briefed |
hn hn-ai | 1mo | Voynich Project – A Python suite for manuscript structural validation | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Generative LLMs aren't an intelligence, so I have stopped referring to them as "AI". Further, generative LLMs are designed to create text t… | Generative LLMs aren't an intelligence, so I have stopped referring to them as "AI". Further, generative LLMs are designed to create t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Is This Evidence Decision-Critical? Learning to Verify Rule-Governed Decisions | Rule-based reasoning, as in eligibility checks and contract reviews, requires language models to assess evidence against individual co… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Pretext: Defeating Malicious Skill Detection Frameworks for AI Agents | Skills extend an agent's capabilities by injecting instructions and information into the context, and are widely used by agents such a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | anthropic sounds like a furry convention but its actually some other inconceivable bullshit | anthropic sounds like a furry convention but its actually some other inconceivable bullshit | ↗Not briefed |
bluesky bluesky-ai | 1mo | SORKIN: I'm half joking -- does OpenAI need to change its name? DNI JAY CLAYTON: Ha ha ha. You know what? Why don't you ask them, Andrew | SORKIN: I'm half joking -- does OpenAI need to change its name? DNI JAY CLAYTON: Ha ha ha. You know what? Why don't you ask them, And… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Why Do Conventional World Models Fail to Learn Cellular Automata? | Although conventional world models - auto-regressive or diffusion models based on transformers or convolutional networks - may learn s… | ↗Not briefed |
rss theverge-ai | 1mo | Here’s how tech leaders will self-police AI safety under Trump’s deal | Anyone else getting “The Last Supper” vibes from this photo taken at the White House tech gathering? | Photographer: Tierney L. Cross/… | ↗Not briefed |
arxiv arxiv-ai | 1mo | GroundingPI: A Grounding Foundation Model towards Physical Intelligence with Visual Primitives | Precise grounding matters. It specifies which object is the target and where that object is, even in clutter and for tiny objects, and… | ↗Not briefed |
arxiv arxiv-ai | 1mo | GroundAnything: Reconciling Parallel Decoding with Precise Visual Grounding at Flash Speed | Autoregressive (AR) grounding models serialize spatial predictions, introducing sequential latency and imposing a causal order on outp… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Beware LLM-generated writing that's used in human communication. It's always excessively verbose and typically useless. Nothing good comes … | Beware LLM-generated writing that's used in human communication. It's always excessively verbose and typically useless. Nothing good c… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Text-to-3D Policy: Fine-Grained Language-Behavior Alignment for Unseen Specification Generalization | 3D visuomotor policies provide a strong foundation for spatially precise manipulation, yet current text-to-3D policies struggle to fol… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Convergence of Practical Muon with Finite Newton-Schulz Iterations and Nesterov Momentum | Practical Muon maintains momentum and performs a small, fixed number of Newton--Schulz iterations separately for each parameter matrix… | ↗Not briefed |
bluesky bluesky-ai | 1mo | When you tell an AI that the technician tasked to shut it off is stuck dying in a room and an alarm has been set to go save them, the AI wi… | When you tell an AI that the technician tasked to shut it off is stuck dying in a room and an alarm has been set to go save them, the… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yes. For example, you have actually BEEN an executive assistant, you know there’s much more to it than robotically scheduling appointments… | Yes. For example, you have actually BEEN an executive assistant, you know there’s much more to it than robotically scheduling appoint… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: FinTrace – point-in-time guardrails for financial LLM agents | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: QuatEngine 4^8 – A Base-4^8 Quaternary OS and hardware simulator | — | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: Anyone noticed that Gemini is not giving credit to brands | Hello HN, I noticed that Gemini is not giving cerdit to brand ownwers, it uses their data on the page but does not drive the searchers… | ↗Not briefed |