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 | 29d | Does Pretraining Teach Chemical Language Models Chemistry? | Chemical language models (CLMs) are pretrained on large corpora of molecules with the aim of learning representations that transfer to… | ↗Not briefed |
bluesky bluesky-ai | 29d | I was on hold with anthropic again trying to figure out which of the 3 phone calls and 2 emails I got has the correct information and the p… | I was on hold with anthropic again trying to figure out which of the 3 phone calls and 2 emails I got has the correct information and… | ↗Not briefed |
arxiv arxiv-ai | 29d | Tiny Recursive Message Passing | Message passing neural networks have shown remarkable performance in algorithmic and constraint reasoning, due to their ability to enc… | ↗Not briefed |
arxiv arxiv-ai | 29d | ST-GRU PredRNN-V2: A GRU-Based Reformulation of PredRNN-V2 for Video Prediction | Video prediction requires models to capture complex spatiotemporal dependencies and to retain relevant information over long sequences… | ↗Not briefed |
bluesky bluesky-ai | 29d | Computer Scientist here! This entire fear is alarmist BS to try and push for more data centers. Only way this happens is if Anthropic tell… | Computer Scientist here! This entire fear is alarmist BS to try and push for more data centers. Only way this happens is if Anthropic… | ↗Not briefed |
arxiv arxiv-ai | 29d | MultiAlloc: Budget-Aware Precision Allocation for KV Cache Quantization | The Key-Value (KV) cache constitutes a dominant source of memory consumption during large language model (LLM) inference, particularly… | ↗Not briefed |
arxiv arxiv-ai | 29d | To Align or Not To Align: Check Your COMPASS Before You Train🔧 Modal | Combining pretrained models is a common strategy for building AI systems, whether by aligning a vision encoder with a language model o… | ↗Not briefed |
arxiv arxiv-ai | 29d | Grid or Nodes? Probing Object Memory in World Models Under Partial Observability | A robot's world model must remember objects that leave the camera's view. FloWM, a learned world model, keeps them in a spatial grid o… | ↗Not briefed |
bluesky bluesky-ai | 29d | github.com/pfrazee/pdoo... | github.com/pfrazee/pdoo... | ↗Not briefed |
hn hn-ai | 29d | Xara Xtreme is back. After 20 years. With a vengeance | — | ↗Not briefed |
rss techmeme | 29d | Google rolls out Gemini 4 Argon globally via the Gemini API and to Google AI Pro subscribers in the Gemini app and AI Mode, and in Antigravity for Ultra users (Robert Hart/The Verge) | Robert Hart / The Verge: Google rolls out Gemini 4 Argon globally via the Gemini API and to Google AI Pro subscribers in the Gemini ap… | ↗Not briefed |
bluesky bluesky-ai | 29d | LLM discourse is like sex discourse: the loudest voices have the lowest body count | LLM discourse is like sex discourse: the loudest voices have the lowest body count | ↗Not briefed |
arxiv arxiv-ai | 29d | DTScribe: Generating Models for Digital Twin Reporting using Local Large Language Models🔧 Consensus | The development of a digital twin brings together many stakeholders and technical components. We have previously investigated the repo… | ↗Not briefed |
arxiv arxiv-ai | 29d | Geometric Calibration of Supervised Fine-Tuning | Recent studies investigate the weight geometry through its spectral structure, which characterizes how the weights transform different… | ↗Not briefed |
hn hn-ai | 29d | AI causes uproar in mathematics after OpenAI shares 722 manuscripts and proofs | — | ↗Not briefed |
hn hn-ai | 29d | Today we are rolling out Gemini 4 Argon globally (API, App, Antigravity) | — | ↗Not briefed |
hn hn-ai | 29d | Architecture clarity that moves teams forward | — | ↗Not briefed |
bluesky bluesky-ai | 29d | "we cannot understand the massive capital outlay on AI/LLM without the critical context that the tech platforms were super hungry for somet… | "we cannot understand the massive capital outlay on AI/LLM without the critical context that the tech platforms were super hungry for… | ↗Not briefed |
arxiv arxiv-ai | 29d | Breakdown of Conversational Involvement: A Joint Multi-Modal Analysis of Conversation Floors and F-formation Detection with Multiple Features🔧 Modal | Interactions happen universally and spontaneously in daily lives. Studying involvement in interactions helps us analyze who can influe… | ↗Not briefed |
hn hn-ai | 29d | Gemini 4 Argon Model Card | — | ↗Not briefed |
arxiv arxiv-ai | 29d | Falsification-proof non-market allocation mechanisms | Relying on manipulable characteristics to target the allocation of resources is fraught with difficulties. Naively optimal rules lead… | ↗Not briefed |
hn hn-ai | 29d | Gemini 4 Argon | — | ↗Not briefed |
arxiv arxiv-ai | 29d | MajorTOM:Elliot Pretrain, A Multi-Sensor, Multi-Temporal Pretraining Dataset for Earth Observation | Pretraining generalist Earth observation foundation models requires data that combine several sensors, varied temporal sampling and, i… | ↗Not briefed |
arxiv arxiv-ai | 29d | Minimax rates for pointwise CATE estimation: the role of design smoothness | We study the role of design smoothness in the minimax estimation of conditional average treatment effects at a fixed covariate point.… | ↗Not briefed |
bluesky bluesky-ai | 29d | just found a game on roblox that is DEFINITELY training some AI model on how to fly missiles | just found a game on roblox that is DEFINITELY training some AI model on how to fly missiles | ↗Not briefed |
arxiv arxiv-ai | 29d | Deterministic Prediction as Global Velocity: Flow Matching for Probabilistic Time-Series Forecast | Accurate uncertainty quantification is critical for many time series forecasting applications, and recent studies have increasingly ad… | ↗Not briefed |
rss deepmind-blog | 29d | Gemini 4 Argon for Robotics | — | ↗Not briefed |
rss deepmind-blog | 29d | Gemini 4 Argon: Built to solve your hardest problems | — | ↗Not briefed |
arxiv arxiv-ai | 29d | Two Routes to Interoceptive Prioritisation: Dissociating Likelihood Precision and Preference Reweighting in Active Inference | An embodied agent must prioritise among competing bodily needs. In active inference, it can respond to an urgent need by increasing th… | ↗Not briefed |
arxiv arxiv-ai | 29d | Knowledge Graphs as Files: An Agent Harness Where Tool Creation and Tool Use Co-Evolve | Searching text-rich knowledge graphs requires following relations while examining textual evidence. Predefined graph interfaces constr… | ↗Not briefed |
hn hn-ai | 29d | Lift: High-fidelity 4D surface and volume physics transformer in latent space | — | ↗Not briefed |
arxiv arxiv-ai | 29d | Can an Agent Learn to Predict Its Own Failure? Outcome-Grounded Confidence Calibration in LLM Agents | Controlling an acting LLM agent means deciding, at each step of a long-horizon task, whether to trust its next action, abstain, or han… | ↗Not briefed |
rss theverge-ai | 29d | Google finally releases Gemini 4 Argon | After months sitting on the sidelines of the frontier AI race, Google has finally launched its long-awaited Gemini 4 Argon model, clai… | ↗Not briefed |
arxiv arxiv-ai | 29d | Does self-distillation improve reasoning? | Self-distillation is an emerging approach that trains a model to imitate its own predictions when conditioned on privileged informatio… | ↗Not briefed |
arxiv arxiv-ai | 29d | Nudge the Judge: Uncovering Program Decision Hijacking in Real-World Jev Applications | System One models such as Jev change how software makes decisions by mapping open-ended semantic contexts into structured judgments th… | ↗Not briefed |
bluesky bluesky-ai | 29d | so while we're all fighting to get our money out of anthropic, publishers have been filing to claim their share of the settlement and then … | so while we're all fighting to get our money out of anthropic, publishers have been filing to claim their share of the settlement and… | ↗Not briefed |
hn hn-ai | 29d | Caire: Content aware image resize library | — | ↗Not briefed |
arxiv arxiv-ai | 29d | Legal-R1: Agentic Retrieval for Evidence-Based Legal Reasoning | Legal question answering requires answers that are not only plausible, but also grounded in legally valid and citable evidence. Standa… | ↗Not briefed |
arxiv arxiv-ai | 29d | Jev-ToM: Evaluating Theory of Mind in Typed Decision Models | Dialogue and collaborative assistants repeatedly judge what a partner knows, believes, or intends. Typed decision models offer a poten… | ↗Not briefed |
arxiv arxiv-ai | 29d | DPOison-VL: Clean Preference Poisoning in Vision-Language Direct Preference Optimization | Direct Preference Optimization (DPO) has become a widely adopted approach for aligning vision-language models (VLMs), yet the security… | ↗Not briefed |
bluesky bluesky-ai | 29d | Which AI models does the administration’s framework cover? Who decides on “trusted partners”? Can developers share models with allied gover… | Which AI models does the administration’s framework cover? Who decides on “trusted partners”? Can developers share models with allied… | ↗Not briefed |
arxiv arxiv-ai | 29d | Beyond Answer Voting: Fine-Grained Correction of LLM Reasoning through Graph Alignment across Sampled Traces | Large language models increasingly tackle tasks that require multi-step reasoning, where errors in intermediate claims can affect the… | ↗Not briefed |
arxiv arxiv-ai | 29d | From Hindsight to Foresight: Dissecting Clinical Evidence Acquisition in Large Language Models | Large language models (LLMs) often diagnose full sight, static clinical cases well yet struggle in foresight, dynamic diagnosis. Prior… | ↗Not briefed |
hn hn-ai | 29d | Gemini 4 Argon in Google Antigravity | — | ↗Not briefed |
arxiv arxiv-ai | 29d | LLM2R: From Large Language Models to 3D Reassemblers | 3D part assembly aims to reconstruct a complete shape from a set of unassembled parts by predicting a rigid SE(3) transformation for e… | ↗Not briefed |
arxiv arxiv-ai | 29d | AgentStripes: An Evidence-Graded Maturity Model for AI Agents | Benchmarks report what an Artificial Intelligence (AI) agent can do on a task suite, but not whether one released version has been eng… | ↗Not briefed |
arxiv arxiv-ai | 29d | Auditing Future-Memory Use in Point-in-Time Language Agents | Point-in-time (PIT) agents can recall post-cutoff facts even when their external evidence respects a historical cutoff. We examine whe… | ↗Not briefed |
bluesky bluesky-ai | 29d | Anthropic co-founder Jack Clark says "I now think it's more likely we'll see full RSI become possible in 2027", earlier than his previous p… | Anthropic co-founder Jack Clark says "I now think it's more likely we'll see full RSI become possible in 2027", earlier than his previ… | ↗Not briefed |
hn hn-ai | 29d | Show HN: Sayso, free open source dictation for Mac, Windows, and Linux | — | ↗Not briefed |
bluesky bluesky-ai | 29d | uh yeah so has anyone have a template for an anti AI waiver workers can add to contracts? ie i don't want my work fed into LLM/genAI and i … | uh yeah so has anyone have a template for an anti AI waiver workers can add to contracts? ie i don't want my work fed into LLM/genAI a… | ↗Not briefed |