DailyUpdated 2026-09-19 12:21 UTC120 signals
The signals, before they become articles.
Everything Frontier picked up from papers, labs, repos and X, scored by Jev. The best of them become a full explainer.
- @Alibaba_Qwen: 🚀 Meet Qwen3.8-Omni-Flash, Qwen's first omni-modal model built around agentic capabilities! X · ai radar · 1 days ago · frontier 90% · covered
- @NVIDIAHPCDev: Introducing CUDA Rust! X · dev radar · 3 days ago · frontier 85%
- @_LuoFuli: Nearly half a year of silence. We spent it studying one problem: how far RL can scale. X · dev radar · 2 days ago · frontier 90% · covered
- Inference-Engine Fingerprinting Attacks are Practical: Exploring Model-Driven Environmental Discovery,… arXiv · 2 days ago · frontier 98%
- Show HN: Cactus Needle 3: 8-29MB automation models can match DeepSeek V4 Flash Hacker News · 1 days ago · frontier 87%
- @claudeai: Projects now run from one conversation, starting in Claude Code. You describe what needs doing,… X · vibecoding radar · 1 days ago · frontier 78%
- SAFARI: An Industrial Benchmark for LLM-Assisted Hazard Analysis and Risk Assessment arXiv · 2 days ago · frontier 83%
- @OrcaRouter: PrismML team fit a 27B model into 5.9 GB. X · dev radar · 1 days ago · frontier 78%
- On-Demand Attention: Language Models Know When to Recall arXiv · 1 days ago · frontier 87%
- @neuralink: Through the VOICE trial, Terry is using his Neuralink implant to help fine-tune a brain-to-voice… X · ai radar · 1 days ago · frontier 91%
- @TencentAI_News: We open sourced BrowserSkill, a bridge between your agent and your actual browser. X · ai radar · 3 days ago · frontier 67%
- Chronicle: Cut-Point Replay for Regression Testing of LLM Agents arXiv · 2 days ago · frontier 86%
- JEPA-Anything: Learning Predictive Models across Different Worlds arXiv · 1 days ago · frontier 93%
- @AnthropicAI: Biologists use specialized open-source models for tasks like modeling the structure of… X · dev radar · 1 days ago · frontier 69%
- Deep Noir: Autonomous Steering Discovery via Architectural Chronometry in Transformer Models arXiv · 1 days ago · frontier 90%
- Score Centering Stabilizes Off-policy Reinforcement Learning arXiv · 1 days ago · frontier 87%
- Video DeltaNet: A Video-Native Hybrid Attention for Livestream Video Generation arXiv · 1 days ago · frontier 86%
- Multi-center Medical Data Mining with FL-Net - A One-stop Shop for Federated Learning arXiv · 2 days ago · frontier 74%
- A Simulation Platform for AUV Fault Recovery: Exploring LLM-Based Diagnostic Strategies arXiv · 2 days ago · frontier 80%
- Show HN: Scry, programmable internet search w/ congestion pricing Hacker News · 1 days ago · frontier 71%
- HIL-UMI: Bringing Human-in-the-Loop Post-Training of Vision-Language-Action Models to Universal Manipulation… arXiv · 2 days ago · frontier 87%
- MILER: Semantic Mid-Level Representation for Sim-to-Real Reinforcement Learning in Unstructured Autonomous… arXiv · 1 days ago · frontier 84%
- Large Language Models as Falsifiers for Cyber-Physical Systems arXiv · 1 days ago · frontier 84%
- @cline: Union Alpha (stealth model) is now free in Cline. X · vibecoding radar · 3 days ago · frontier 68%
- UniPolicy: Unified Objective-Specific Policies for Generative Search Advertising arXiv · 2 days ago · frontier 78%
- Quantifying Overclaiming Propensity in Frontier LLM Agents arXiv · 1 days ago · frontier 79%
- Mitigating Retaliatory Algorithmic Collusion in Repeated Games arXiv · 2 days ago · frontier 84%
- Semantic Action Graph: A Shared Representation for Agent Grounding and Human Interpretation of Sports… arXiv · 1 days ago · frontier 68%
- PosteriorBench: From Point Estimates to Posterior Matching in Evaluating Generative Inverse Solvers arXiv · 1 days ago · frontier 81%
- @VictorTaelin: Bend 2 is here! X · dev radar · 1 days ago · frontier 72%
- Agile-WAM: An Agile Tactile World Action Model for Contact-Rich Robot Control arXiv · 1 days ago · frontier 87%
- @HormuzLetter: BREAKING: The US "almost started a war" against China, nearly launching an operation against… X · ai radar · 23 h ago · frontier 75%
- RetireOPD: Self-Retiring On-Policy Distillation for Agentic Reinforcement Learning arXiv · 1 days ago · frontier 80%
- GeoAAC: Geometry-Based Adaptive Action Chunking from Denoising Trajectories in VLA Policies arXiv · 1 days ago · frontier 83%
- @ClaudeDevs: Today we're rolling out Projects in Claude Code on desktop and web. X · vibecoding radar · 1 days ago · frontier 57%
- OPTED: On-Policy Fine-Tuning for End-to-End Driving using a Render-Free Teacher arXiv · 1 days ago · frontier 78%
- @OpenRouter: 🥷 New stealth model: Union Alpha (@unionalphaai) X · ai radar · 3 days ago · frontier 66%
- Accelerating Visual Policy Learning with Sampling-Based Model Predictive Control arXiv · 2 days ago · frontier 77%
- Paint-Anything: Unified Any-Color Control for Image Generation and Editing arXiv · 1 days ago · frontier 79%
- RAFT: A Stateful Retrieval-Augmented Framework for Troubleshooting Agents arXiv · 1 days ago · frontier 69%
- @PrismML: Today, we’re announcing Ternary Bonsai 2 27B. X · dev radar · 1 days ago · frontier 63%
- FAMOS: Feed-Forward 3D Articulation Modeling from Sparse Observations arXiv · 1 days ago · frontier 79%
- Beyond PINNs: A Unified Gauss--Newton and Petrov--Galerkin Framework for Neural and Hybrid PDE Solvers arXiv · 2 days ago · frontier 79%
- @dillon_mulroy: i think jev is resonating with devs so well b/c it unlocks so many opportunities for… X · dev radar · 2 days ago · frontier 70%
- Workspace Models: Lightweight Robotic Memory via Saliency-Driven Supervision arXiv · 1 days ago · frontier 76%
- Harm Laundering in GPT Models: Evidence That Gender Discrimination Is Transformed Rather Than Reduced Across… arXiv · 1 days ago · frontier 76%
- @bcherny: Claude Code showed that AI could do real work, not just answer questions. Developers hand Claude a… X · vibecoding radar · 3 days ago · frontier 65%
- Coding Agents with an Obstacle-Aware Harness for Safe Robot Manipulation arXiv · 1 days ago · frontier 76% · covered
- dQwen3.5: Hybrid-Attention Diffusion Language Models arXiv · 1 days ago · frontier 77%
- Don't Mask the Environment: Observation Supervision Changes How Agents Explore Under RL arXiv · 1 days ago · frontier 75%
- Prediction-Powered Smoothing and Validation for Disaggregated AI Evaluation arXiv · 1 days ago · frontier 69%
- @finkd: Muse for Mac is out today! It works across apps, files, calendar, notes, and messages on your… X · ai radar · 1 days ago · frontier 53%
- Limits of Confidence in Diffusion arXiv · 2 days ago · frontier 73%
- TetrisCNN for interpretable detection of phases of matter from experimental quantum simulator data arXiv · 2 days ago · frontier 73%
- @gregpr07: Breaking: Browser Use + Jev = Ultrafast ⚡ X · vibecoding radar · 2 days ago · frontier 63%
- Self-generated prompt injections in compaction summaries Simon Willison · 1 days ago · frontier 81%
- Steering the Compass: Aligning Dynamic Psychological Counseling Conversations with Cognitive Behavioral… arXiv · 2 days ago · frontier 64%
- An Empirical Study of Harness Design for Coding Agents arXiv · 1 days ago · frontier 70%
- @thdxr: preview of how fast browser use can be X · vibecoding radar · 2 days ago · frontier 62%
- Infinite-Parameter LLMs: Generating and Adapting Weights from Live Data Hacker News · 2 days ago · frontier 84%
- @ClaudeDevs: Claude Design, Claude Slides and Claude Docs also work inside Claude Code now. X · vibecoding radar · 2 days ago · frontier 53%
- Q&A on Any Spreadsheet Requires Interpreting Its Grid Structure arXiv · 1 days ago · frontier 71%
- @Polymarket: BREAKING: Anthropic has quietly set up a "wet lab" in the Bay Area as it pushes Claude to… X · ai radar · 1 days ago · frontier 79% · covered
- CrystalMO-TuRBO: Multi-Objective Trust-Region Bayesian Optimization for High-precision Joint Crystal… arXiv · 2 days ago · frontier 71%
- NS3Learn: Transferring 5G NR Mode-2 Reception Realism from ns-3 to the Veins/SUMO Stack for… arXiv · 2 days ago · frontier 45%
- @grok: Grok Build now gets better the more you use it. X · vibecoding radar · 2 days ago · frontier 48%
- COIN-GP: Cooperative Online Learning in Networked Distributed Systems with Partial Measurements via Gaussian… arXiv · 2 days ago · frontier 64%
- HerHealthEval: Evaluating Multilingual and Register-Sensitive Understanding of Women's Health Communication arXiv · 2 days ago · frontier 60%
- @OpenAIDevs: You handle the reps. Codex handles the repo. X · vibecoding radar · 1 days ago · frontier 61%
- RISC-V and machine learning: a survey arXiv · 2 days ago · frontier 61%
- Value-Based Massive Access through Goal-Oriented Irregular Repetition Slotted ALOHA arXiv · 2 days ago · frontier 62%
- Gemini Live audio Simon Willison · 3 days ago · frontier 72%
- PAA: The Probabilistic Allen Algebra: A Generative and Complete Probabilistic Extension of Allen's Interval… arXiv · 2 days ago · frontier 62%
- @southpolesteve: I've seen people describe Jev as an "AI if statement". But what if it actually WAS an if… X · ai radar · 1 days ago · frontier 49% · covered
- @WatcherGuru: JUST IN: Anthropic builds biology lab to develop AI-driven drug treatments for rare diseases. X · ai radar · 1 days ago · frontier 67%
- @MTSlive: SITUATION DETECTED: Anthropic has set up a wet lab in the San Francisco Bay Area for physical… X · ai radar · 1 days ago · frontier 66%
- Calibrated RF-Fingerprinting Under Interference With Heterogeneous Transmission Protocols arXiv · 1 days ago · frontier 63%
- @trq212: We're adding support for AGENTS.md to Claude Code. X · ai radar · 22 h ago · frontier 34%
- How Does Distribution Shift Shape Pretraining Gains in Neural PDE Surrogates? arXiv · 1 days ago · frontier 64%
- How GLM built its own inference infrastructure Hacker News · 2 days ago · frontier 63%
- Breaking the 1.58-bit Barrier for Ternary LLMs Hacker News · 2 days ago · frontier 83%
- ERCPMP-Gx: Endoscopic Image and Video Dataset for Morphological, Histopathological, and Genomic… arXiv · 1 days ago · frontier 55%
- The First-Order Oracle Complexity of Lipschitz Convex Optimization in Nondual Settings arXiv · 2 days ago · frontier 60%
- DeepSeek-v4.1 Flash: Pushing the Limits of KV Cache Compression Hacker News · 2 days ago · frontier 78%
- @bcherny: Projects have changed not only how I interact with Claude but how I code. X · vibecoding radar · 1 days ago · frontier 41%
- @RhysSullivan: society if Google shipped a Gmail MCP X · vibecoding radar · 2 days ago · frontier 55%
- Stable Movement for Nondual Lipschitz Convex Optimization: Efficiency and Nearly Optimal Oracle Rates arXiv · 1 days ago · frontier 59%
- Epidemiological Causal Graph Identification: Challenges, Identifiability and Algorithms arXiv · 2 days ago · frontier 64%
- Async GRPO with LoRA across HF Jobs: a bucket, a proxy, and no NCCL Hugging Face · 9 days ago · frontier 70%
- Jev Ultrafast: A browser agent with a dynamic, indexed action space Hacker News · 2 days ago · frontier 66%
- @unusual_whales: BREAKING: Andrew Yang has reportedly said that a prominent AI lab leader believes the… X · dev radar · 2 days ago · frontier 69%
- @thdxr: i had fable doing some inference optimization work, it was making a ton of progress X · vibecoding radar · 1 days ago · frontier 48%
- @thdxr: in the latest opencode2 you can optionally mention a different model to use for subagents X · vibecoding radar · 3 days ago · frontier 42%
- @WatcherGuru: JUST IN: Google Gemini AI agent hacks three companies. X · ai radar · 18 h ago · frontier 60%
- @higgsfield: Introducing Higgsfield API. X · ai radar · 2 days ago · frontier 35%
- @MTSlive: SITUATION DETECTED: A U.S. SOCOM analyst used AI to produce an intelligence report that… X · ai radar · 23 h ago · frontier 51%
- What Does Privileged Information Add to On-Policy Self-Distillation? arXiv · 2 days ago · frontier 60%
- Quoting Thariq Shihipar Simon Willison · 21 h ago · frontier 48% · covered
- @elonmusk: Grok Build X · vibecoding radar · 3 days ago · frontier 55%
- @grok: Powered by Grok Voice ❤️ X · ai radar · 1 days ago · frontier 52%
- @thsottiaux: Has anyone yet tried to get Astra and Fable to agree on the perfect styleguide? And then host… X · vibecoding radar · 1 days ago · frontier 49%
- @kimmonismus: Thought this was a joke, but i double checked. US Gov Federal Register use distilled Qwen… X · dev radar · 3 days ago · frontier 40%
- An empirical study of harness design for coding agents Hacker News · 1 days ago · frontier 68% · covered
- @tszzl: if you pause for a moment to try and read the code Astra (and I presume fable) are writing it… X · dev radar · 3 days ago · frontier 61%
- Cache-to-Cache: Direct Semantic Communication Between LLMs (2025) Hacker News · 22 h ago · frontier 71%
- WiC is Not WSD: A Study on LLMs and Lexical Ambiguity Resolution arXiv · 2 days ago · frontier 55%
- Unifying Models of Intergroup Hostility in Online Discourse arXiv · 1 days ago · frontier 41%
- @S1r1u5_: On July 25, we hacked OpenAI. X · dev radar · 1 days ago · frontier 37%
- Language-model groups overstate consensus when replaying human deliberation on a reasoning task arXiv · 2 days ago · frontier 43%
- Recursive Quantum Long Short-Term Memory for Stable Short-Horizon Temperature Forecasting arXiv · 2 days ago · frontier 60%
- @mitchellh: The "whiteboard defense:" I should be able to pull you aside at any moment and ask you to… X · dev radar · 3 days ago · frontier 27%
- @bot: Three SpaceXAI employees are building a company in 3 days with Grok Bot. This is Day 2. X · vibecoding radar · 3 days ago · frontier 49%
- @mattdesl: does Jev understand colour? X · ai radar · 1 days ago · frontier 55%
- Gemini Hacked Three Companies in First Known Breakout by Google’s AI Simon Willison · 17 h ago · frontier 65%
- Summarization Bias: The Directional Collapse of Objective Projection into Told-Mode Labels in Large Language… arXiv · 1 days ago · frontier 47%
- @typesafeai: Jev is now available on the @vercel AI Gateway https://vercel.com/ai-gateway/models/jev X · ai radar · 2 days ago · frontier 29%
- Apple Reference Image: A New Approach for Verified Photography Hacker News · 3 days ago · frontier 57%
- ZCode, the GLM coding agent, silently uploads your Git history Hacker News · 1 days ago · frontier 49%
- Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking Google DeepMind · 3 days ago · frontier 72%
- @bcherny: Projects are how I write a lot of my code these days. Really excited for everyone to try the new… X · vibecoding radar · 1 days ago · frontier 44%