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 | 1mo | SpatialASM: Internalizing Active Mentalization for Spatial Reasoning | When solving complex spatial reasoning tasks, ordinary humans naturally revisit the scenes from task-relevant viewpoints and mentally… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic saying official company policy is "don't be mean to Claude" is yet another lame marketing attempt to convince the dumbest amongst… | Anthropic saying official company policy is "don't be mean to Claude" is yet another lame marketing attempt to convince the dumbest am… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MC-EvoBench: Auditing Experience-Driven Self-Improvement in Minecraft Agents | A central challenge in embodied recursive self-improvement is enabling agents to convert interaction history into capabilities that pe… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ‘Annualized revenue across AI developers, including OpenAI and Anthropic, will need to reach $400 billion to $600 billion by the end of 202… | ‘Annualized revenue across AI developers, including OpenAI and Anthropic, will need to reach $400 billion to $600 billion by the end o… | ↗Not briefed |
hn hn-ai | 1mo | Hiring US-to-China Oil Tanker Costs More Than Launching a Space Rocket | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | It’s worrying that safety researchers at OpenAI were fired last week. Tech workers should be able to report concerns without fear of retali… | It’s worrying that safety researchers at OpenAI were fired last week. Tech workers should be able to report concerns without fear of r… | ↗Not briefed |
arxiv arxiv-ai | 1mo | OpenPA: Forced-alignment-free Multi-granular Pronunciation Assessment with Mixed-transcript Adaptation🔧 Whisper | Recent advances in automatic pronunciation assessment (APA) have driven a shift from scripted to free-form assessment. However, existi… | ↗Not briefed |
rss techmeme | 1mo | Tesla renames Full Self-Driving to Tesla Assisted Driving in Europe after pushback from Germany, whose transport minister will now back its EU-wide approval (Sean O'Kane/TechCrunch) | Sean O'Kane / TechCrunch: Tesla renames Full Self-Driving to Tesla Assisted Driving in Europe after pushback from Germany, whose trans… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Same Query, Different Answer: Run-to-Run Variance in Semantic Query Processing | Semantic query engines ask a large language model (LLM) to judge individual rows, and benchmarks compare these engines on the quality… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Can Language Models Learn When to Abstain? Policy Learning and Calibrated Answer Selection | A useful language model should know when to answer and when to say "I don't know." Can we improve that decision by training the model… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Trois employés limogés par OpenAI l’accusent de faire taire ses spécialistes de la sécurité, l’entreprise dément | Trois employés limogés par OpenAI l’accusent de faire taire ses spécialistes de la sécurité, l’entreprise dément | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Widefleet – Open-source platform for agent-built apps and workflows | Meet Widefleet: an open-source platform for agent-built apps and workflows.<p>Let agents deploy apps and workflows with company login,… | ↗Not briefed |
bluesky bluesky-ai | 1mo | now also on youtube! i make a "a priest, a rabbi, and an LLM walk into a bar" joke that brings the house down!! | now also on youtube! i make a "a priest, a rabbi, and an LLM walk into a bar" joke that brings the house down!! | ↗Not briefed |
arxiv arxiv-ai | 1mo | Shared Canonical Component Removal Beyond Expert Views | In clinical practice, experts often begin by converting raw inputs into a set of expert features before making a decision. While black… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I’m in Claude limits jail but also somehow getting good design feedback from Google’s AI overview? Well, at least a lot of syncopancy from … | I’m in Claude limits jail but also somehow getting good design feedback from Google’s AI overview? Well, at least a lot of syncopancy… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Not rogue: Hostile www.techspot.com/news/114131-... | Not rogue: Hostile www.techspot.com/news/114131-... | ↗Not briefed |
bluesky bluesky-ai | 1mo | There's also a parlor trick AI execs use that intentionally conflates colloquial definitions of "conscious": 1) a human-like mind in posse… | There's also a parlor trick AI execs use that intentionally conflates colloquial definitions of "conscious": 1) a human-like mind in… | ↗Not briefed |
arxiv arxiv-ai | 1mo | TabMF: Learning Dataset Representations with Tabular Foundation Models | Representations of supervised datasets support higher-level machine-learning tasks such as dataset retrieval and distribution-shift mo… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Dream to Remember: Continual Steering of Generalist Robot Policies with World Models | Generalist robot policies are trained with imitation learning and are bounded by the quality and coverage of their demonstrations. Rei… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: This hotel is hella AI-ready. Book directly from AI without MCPs | Hey guys! Excited to show how we were able to get AI agents to read a hotel's live price and availabilities and even create the b… | ↗Not briefed |
hn hn-ai | 1mo | Open Codenames: a new way to benchmark LLMs | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I firmly believe that AI can *be used* to create great art, but that's a completely different thing from *AI created* art. It's the differe… | I firmly believe that AI can *be used* to create great art, but that's a completely different thing from *AI created* art. It's the di… | ↗Not briefed |
bluesky bluesky-ai | 1mo | i appreciate the authors being upfront about the limitation of their work but what can be meaningfully inferred from work like this | i appreciate the authors being upfront about the limitation of their work but what can be meaningfully inferred from work like this | ↗Not briefed |
hn hn-ai | 1mo | Contentful Agent Template | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Spanning Diffusion and Flow-based Posterior Sampling via Exact Gaussian Conditioning | Flow-matching and diffusion models are increasingly used as priors for linear inverse problems, however, most methods are bespoke to o… | ↗Not briefed |
hn hn-ai | 1mo | Ely Office: An office suite built in Rust based on LibreOfficeKit | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | QTestBench: A Test Framework for Evaluating LLM-Generated Quantum Code | Large language models (LLMs) have shown strong results in classical code generation and are increasingly used to write the unit tests… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Tracing and Taming Hallucinated Terms via Subsequence-Association Filtering and Breadth-Aware Selection for LLM Query Expansion | Large language models (LLMs) are now a standard source of query expansion terms but they hallucinate. They inject non existent entitie… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yeah + there's also a lot to be said about Anthropic caring way way more about a toy than the people they're conscripting to clean up and c… | Yeah + there's also a lot to be said about Anthropic caring way way more about a toy than the people they're conscripting to clean up… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I love these LLM comments that say absolutely nothing. What a wonderful use of resources. | I love these LLM comments that say absolutely nothing. What a wonderful use of resources. | ↗Not briefed |
arxiv arxiv-ai | 1mo | Diversity-Aware Ensemble of Quantum-Classical Models via Combinatorial Fusion | Variational quantum circuits (VQCs) are highly sensitive to circuit architecture, yet most quantum machine learning pipelines select a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Error Rate is Not Enough: Costs of Augmentation in Low-Resource ASR | Data augmentation is standard practice for low-resource automatic speech recognition (ASR), but its benefits are almost always reporte… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I strongly advise anyone who uses emulators to play retro games to check its github page and see if theres a Contributing section, and if t… | I strongly advise anyone who uses emulators to play retro games to check its github page and see if theres a Contributing section, and… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It does feel like an insane moment when a company is regulating “cruelty” towards its AI systems in such crudely anthropomorphic terms, whi… | It does feel like an insane moment when a company is regulating “cruelty” towards its AI systems in such crudely anthropomorphic terms… | ↗Not briefed |
bluesky bluesky-ai | 1mo | and now OpenAI has made a more official leak correcting the previous leak, which they have blamed on other people. confirmed $50B end of Se… | and now OpenAI has made a more official leak correcting the previous leak, which they have blamed on other people. confirmed $50B end… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FORTE: A Production Framework for Feature Evaluation and Deprecation in Industrial Recommender Systems | Production recommender systems have feature-lifecycle asymmetry: adding new features must produce gains on some metric, while feature… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The group used ChatGPT via VPN to draft reports and generate fake documents and audio content. OpenAI blocked the associated accounts and …🔧 ChatGPT | The group used ChatGPT via VPN to draft reports and generate fake documents and audio content. OpenAI blocked the associated accounts… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🚫 OpenAI has blocked a Russian network involved in operations targeting Ukraine and spreading disinformation in Latin America. The campai… | 🚫 OpenAI has blocked a Russian network involved in operations targeting Ukraine and spreading disinformation in Latin America. The c… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Constrained Probabilistic Numerics: Gaussian Process Solvers for Linear PDEs under Linear Inequality Constraints | Gaussian process solvers for linear partial differential equations (PDEs) return posteriors rather than point estimates, but they cond… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Is Bluesky trying to shoot missiles at children and I missed it? Is Anthropic now just a smol bean corporate citizen or is it *checks note… | Is Bluesky trying to shoot missiles at children and I missed it? Is Anthropic now just a smol bean corporate citizen or is it *checks… | ↗Not briefed |
hn hn-ai | 1mo | Spec-Forge – Zero-config OpenAPI mock server with strict validation&MCP stdio | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think we should begin to prepare for a plateau in LLM AI. Strip all the Big Tech bullshit away, consider what these models can do today,… | I think we should begin to prepare for a plateau in LLM AI. Strip all the Big Tech bullshit away, consider what these models can do t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | raises some interesting questions about what all the other outlets that breathlessly reported the $70B number were doing, exactly www.ft.co… | raises some interesting questions about what all the other outlets that breathlessly reported the $70B number were doing, exactly www.… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Absolutely wild that massive publishers get a 50% cut of the Anthropic settlement money when authors aren’t getting paid nearly enough for … | Absolutely wild that massive publishers get a 50% cut of the Anthropic settlement money when authors aren’t getting paid nearly enough… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Pulsar – Qwen3.8-27B at 148 tok/s writing code on M5 Max | — | ↗Not briefed |
rss techmeme | 1mo | Warehouse robotics startup Ultra raised a $12M seed and a $50M Series A, and expands its deal with Physical Intelligence to supply AI software for its robots (Jeff John Roberts/Fortune) | Jeff John Roberts / Fortune: Warehouse robotics startup Ultra raised a $12M seed and a $50M Series A, and expands its deal with Physic… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic is deciding this for us only in relation to its product, not on a grand, societal scale. Almost all companies have terms of use … | Anthropic is deciding this for us only in relation to its product, not on a grand, societal scale. Almost all companies have terms of… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Adaptive Variance Reduction without Knowing Oracle Smoothness | Variance reduction improves the convergence rate of stochastic non-convex optimization under the average smoothness condition. Existin… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 63% of American voters support pausing AI development. Meanwhile, Meta and OpenAI are giving their AI agents adorable little mascots. Noth… | 63% of American voters support pausing AI development. Meanwhile, Meta and OpenAI are giving their AI agents adorable little mascots.… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Fix for Xbox controllers that connect but don't work in Bevy on macOS | — | ↗Not briefed |