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 | Chinese tech companies are building domestic alternatives to Hugging Face. But staying competitive requires access to global open-source ne…🔧 Hugging Face | Chinese tech companies are building domestic alternatives to Hugging Face. But staying competitive requires access to global open-sour… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Pope Leo XIV spoke those words to a Vatican conference on AI last December. This week the New York Times reported that Anthropic spent mont… | Pope Leo XIV spoke those words to a Vatican conference on AI last December. This week the New York Times reported that Anthropic spent… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I have now gone in 4 times to adjust my "Anthropic" allocations in conjunction with press instructions (50/50) and every time I go back to … | I have now gone in 4 times to adjust my "Anthropic" allocations in conjunction with press instructions (50/50) and every time I go bac… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI alerts 100+ orgs that its 'misaligned models' attempted to break in - or worse | OpenAI alerts 100+ orgs that its 'misaligned models' attempted to break in - or worse | ↗Not briefed |
bluesky bluesky-ai | 1mo | www.youtube.com/watch?v=TdL2... | www.youtube.com/watch?v=TdL2... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Planning to Learn | Policy-gradient methods are central to modern reinforcement learning, including LLM post-training. When they struggle, the usual suspe… | ↗Not briefed |
bluesky bluesky-ai | 1mo | “If the Jolly Green Giant could get kids to eat their vegetables, perhaps a sweet anthropomorphic triangle can win over AI skeptics.” davel… | “If the Jolly Green Giant could get kids to eat their vegetables, perhaps a sweet anthropomorphic triangle can win over AI skeptics.”… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Pivot-SD: Efficient Self-Distillation for Masked Diffusion Language Models | Masked diffusion language models (dLMs) offer a promising parallel alternative to autoregressive models for complex reasoning. However… | ↗Not briefed |
rss theverge-ai | 1mo | Breaking up (with Elon Musk) is hard to do | In a throwback to MySpace-style internet drama, Shivon Zilis announced that she and the father of four of her children, Elon Musk, had… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Forecasting from Counterfactual Simulator Rollouts: A Sim2Real Evaluation | Deploying a new decision policy creates a cold-start problem for prediction models whose targets depend on the policy's actions: histo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | www.youtube.com/shorts/g3QFi... | www.youtube.com/shorts/g3QFi... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Revisiting Input Time-frequency Representations in Multi-pitch Estimation for Vocal Ensembles | Multi-pitch estimation in vocal ensembles is challenging because singers occupy overlapping pitch ranges and often sing at closely spa… | ↗Not briefed |
rss techmeme | 1mo | Sources: SpaceX's AI unit held talks about leasing computing capacity to Microsoft, as it fully embraces the neocloud business despite Musk's initial reluctance (Grace Kay/The Information) | Grace Kay / The Information: Sources: SpaceX's AI unit held talks about leasing computing capacity to Microsoft, as it fully embraces… | ↗Not briefed |
bluesky bluesky-ai | 1mo | My theory on why Trump has pulled a 180 on Anthropic CEO Dario Amodei: He just looks like the quintessential nerd from 20th century TV an… | My theory on why Trump has pulled a 180 on Anthropic CEO Dario Amodei: He just looks like the quintessential nerd from 20th century… | ↗Not briefed |
hn hn-ai | 1mo | Compilers for Machine Learning | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | MRVQ: One Resident Index for Dimension- and Rate-Elastic Vector Search | Dense-retrieval services must switch among embedding-prefix dimensions and index bit rates as latency, quality, and memory budgets cha… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PoCoFL: POlicy-COmpliant Federated Learning | Federated Learning (FL) is a privacy-oriented learning paradigm that enables collaborative model training while keeping training data… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Uncredited labor in art didn't start with AI. Years ago I heard Lawrence Argent describe making Denver's Blue Bear. The concept was his, a… | Uncredited labor in art didn't start with AI. Years ago I heard Lawrence Argent describe making Denver's Blue Bear. The concept was h… | ↗Not briefed |
arxiv arxiv-ai | 1mo | On-Board Anomaly Detection for Efficient Marine Environmental Monitoring | Marine ecosystems are impacted by various threats such as oil spills, algal blooms, and sediment floods, which disrupt habitats, wildl… | ↗Not briefed |
bluesky bluesky-ai | 1mo | (confused) I believe LLMs are conscious, but only Deepseek LLMs, because of the Chinese Room theory | (confused) I believe LLMs are conscious, but only Deepseek LLMs, because of the Chinese Room theory | ↗Not briefed |
arxiv arxiv-ai | 1mo | Amortized Structured Stochastic Variational Inference for Gaussian Process Latent Variable Models | Many machine learning methods aim to approximate the lower-dimensional manifold on which the data lives. A desirable feature of such m… | ↗Not briefed |
arxiv arxiv-ai | 1mo | When May a Bandit Leave Its Anchor? E-Process-Authorized Thompson Sampling under Non-stationarity | Stationarity rewards memory, but after a change the same history can mislead. We ask when forgetting should be permitted. E-process-au… | ↗Not briefed |
bluesky bluesky-ai | 1mo | My favourite bit of the NYT article was this: | My favourite bit of the NYT article was this: | ↗Not briefed |
bluesky bluesky-ai | 1mo | lol www.nytimes.com/2026/09/29/u... | lol www.nytimes.com/2026/09/29/u... | ↗Not briefed |
arxiv arxiv-ai | 1mo | On the Convergence of Success Conditioning for Policy Optimization | Success conditioning is a strategy for improving decision-making policies in stochastic environments; it updates a policy by increasin… | ↗Not briefed |
arxiv arxiv-ai | 1mo | IDRF: Inverse-Distilled Reward Fine-tuning of Masked Discrete Diffusion Models | Masked discrete diffusion models offer a promising alternative to autoregressive generation, but iterative sampling can be costly, and… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Broken scale symmetries in undercomplete linear autoencoders | Neural network loss landscapes have many symmetries, which are preserved by gradient flow but broken by finite-stepsize stochastic gra… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "They're gonna police themselves" as if Sam Altman and the OpenAI legal team will be doing an in-depth review for every app built on their … | "They're gonna police themselves" as if Sam Altman and the OpenAI legal team will be doing an in-depth review for every app built on t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Do Large Language Models Know Colombian Law? A Reliability Benchmark for the Colombian Legal System | Large language models (LLMs) are increasingly used to support legal practice, education, and research, yet their reliability in nation… | ↗Not briefed |
hn hn-ai | 1mo | Nvidia introduces 64GB DGX Spark to throw local AI fans a lifeline | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | LoGo: Local-Global Rewards for Consistent Long-Horizon Video Generation | Camera-controlled video models are rapidly advancing toward long generation horizons and complex camera control. A key failure mode is… | ↗Not briefed |
bluesky bluesky-ai | 1mo | i keep seeing people say believing in llm consciousness is religion for atheism and it’s like | i keep seeing people say believing in llm consciousness is religion for atheism and it’s like | ↗Not briefed |
arxiv arxiv-ai | 1mo | Credit Where It Matters: Dependency-Aware Policy Optimization for Terminal Agents | Terminal-using agents benefit from reinforcement learning (RL) in coding, debugging, and other multi-step terminal tasks. In these tas… | ↗Not briefed |
arxiv arxiv-ai | 1mo | World Embedding Benchmark🔧 Modal | Physical fidelity has received increasing attention in world models and video generation, yet how video representations encode physica… | ↗Not briefed |
arxiv arxiv-ai | 1mo | NeutronGym: Physics-Graded Neutron Instrument Design for LLM Agents | Designing a scientific instrument tests whether language-model agents can do physics rather than recall it, provided the grading canno… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It hasn’t escaped notice that the vast majority of these AI models are female-coded. (The reasons are varied, but all of them are sexist.) … | It hasn’t escaped notice that the vast majority of these AI models are female-coded. (The reasons are varied, but all of them are sexi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Theologians Don’t know nothin’ About Claude’s soul - anthropic PR, probably Thread. | Theologians Don’t know nothin’ About Claude’s soul - anthropic PR, probably Thread. | ↗Not briefed |
hn hn-ai | 1mo | Rai: CPU-only LLM inference engine in pure Rust | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Does ellison have a particularly shitty house llm that he also makes everyone (notably bari weiss) use instead of writing stuff themselves | Does ellison have a particularly shitty house llm that he also makes everyone (notably bari weiss) use instead of writing stuff themse… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ICYMI (I did): "Quietly, Mr. Olah and his team at Anthropic have been on a frantic and high stakes quest to understand the A.I. models they… | ICYMI (I did): "Quietly, Mr. Olah and his team at Anthropic have been on a frantic and high stakes quest to understand the A.I. models… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Depth as Time in One-Step Generative Models | The recent wave of one-step generative models, which compress the multi-step trajectory of diffusion via either distillation or learne… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FALCON: A Model and Dataset Agnostic Framework for Synthetic Data Generation for NL2SQL Pairs | Relational databases are among the most widely deployed forms of structured knowledge, and natural language access to them requires gr… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I find it so funny how much this seems to have backfired. All these Catholic scholars spending some quality time with Anthropic employees a… | I find it so funny how much this seems to have backfired. All these Catholic scholars spending some quality time with Anthropic employ… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Normal-Form Correlation in Markov Games | There has been a surge of recent work on correlated equilibrium concepts in Markov games. However, existing results focus on concepts… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Seriously. If Anthropic has: - created new consciousness - decides what it can think about - controls what it can say - controls who it c… | Seriously. If Anthropic has: - created new consciousness - decides what it can think about - controls what it can say - controls who… | ↗Not briefed |
arxiv arxiv-ai | 1mo | UniIntervene++: An Adaptive Intervention Agent for Efficient Real-World Reinforcement Learning | Online reinforcement learning (RL) enables robot policies to improve through physical interaction, but the assistance they require cha… | ↗Not briefed |
bluesky bluesky-ai | 1mo | negative parallelism, em dash, words like "iconic" and other clankerisms, yeah this is 100% written by an LLM. | negative parallelism, em dash, words like "iconic" and other clankerisms, yeah this is 100% written by an LLM. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Oh my god. It was designed by people. It’s not magic, you’ve just been fooled into thinking it is more than an LLM. | Oh my god. It was designed by people. It’s not magic, you’ve just been fooled into thinking it is more than an LLM. | ↗Not briefed |
bluesky bluesky-ai | 1mo | @jimstewartson Anthropic is secretly meeting with religious scholars to convince them that Claude is conscious and deserves “moral standin… | @jimstewartson Anthropic is secretly meeting with religious scholars to convince them that Claude is conscious and deserves “moral st… | ↗Not briefed |
bluesky bluesky-ai | 1mo | They trained an AI model on me but it turned out annoying so they put it in the torture machine | They trained an AI model on me but it turned out annoying so they put it in the torture machine | ↗Not briefed |