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 | MobiAgent: Dual-Loop Recursive Policy Self-Improvement for Long-Horizon Mobile Manipulation | Long-horizon mobile manipulation presents significant challenges due to compounding execution errors and capacity interference between… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic was lobbying the pope to declare that humanity had replaced God and recreated the divine miracle of breathing a soul into a vesse… | Anthropic was lobbying the pope to declare that humanity had replaced God and recreated the divine miracle of breathing a soul into a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Single or Multiple Policies for Phase-Structured Reinforcement Learning? | Many reinforcement-learning (RL) problems are non-stationary yet structured and can be decomposed into phases, each with its own trans… | ↗Not briefed |
bluesky bluesky-ai | 1mo | With all due respect I think the strategy of declaring by fiat that an LLM cannot do things that you can visibly and experientially confirm… | With all due respect I think the strategy of declaring by fiat that an LLM cannot do things that you can visibly and experientially co… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Can’t believe I find myself in agreement with the Catholic Church on something, but Pope Leo XIV is correct here. The AI models are not co… | Can’t believe I find myself in agreement with the Catholic Church on something, but Pope Leo XIV is correct here. The AI models are n… | ↗Not briefed |
bluesky bluesky-ai | 1mo | So, Open AI is angry because the things they've stolen are at risk of being stolen by "Moonshot" only... they aren't even the original thin… | So, Open AI is angry because the things they've stolen are at risk of being stolen by "Moonshot" only... they aren't even the original… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Ask LM Studio – Use local models directly in the Firefox sidebar🔧 LM Studio | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Preserving Anatomical Continuity: Three-Stage Pipeline for Colon Segmentation in 3D Abdominal CT Scans | Accurate colon segmentation from CT images is essential for colorectal disease analysis, yet deep learning based methods often produce… | ↗Not briefed |
arxiv arxiv-ai | 1mo | When Is Accuracy Evidence? A Unified Theory of Generalisation, Validation, and Information Fusion🔧 Gamma | K-fold cross-validation (CV) is widely used as evidence of out-of-sample performance, although folds are neither independent experimen… | ↗Not briefed |
rss techcrunch-ai | 1mo | Pope Leo XIV is not a fan of AI-generated art | "There is an ontological difference, even before an aesthetic one, between art and what a machine can generate through statistical cal… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Generalization of Transformer-Based Neural Quantum States via In-Context Learning | Neural quantum states based on modern deep learning architectures have emerged as powerful representations for quantum many-body syste… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I used to be annoyed at knowing the names of American right wingers against my will, but now I'm even more annyoed at knowing fucking AI mo… | I used to be annoyed at knowing the names of American right wingers against my will, but now I'm even more annyoed at knowing fucking… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Outis – Fight AI spam by sending fake "user unknown" bounce emails | Outis fights spam generating and sending a fake "user unknown" bounce for an email you received, so the sender believes your… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Folks should typically block "AI" enthusiasts on sight (the ones who bought into the propaganda / fad) because we can't trust them not to f… | Folks should typically block "AI" enthusiasts on sight (the ones who bought into the propaganda / fad) because we can't trust them not… | ↗Not briefed |
bluesky bluesky-ai | 1mo | don't know if I agree, on both ends. LLM long horizon task performance has improved so much that METR doesn't even know how to measure it a… | don't know if I agree, on both ends. LLM long horizon task performance has improved so much that METR doesn't even know how to measure… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Near-Zero Monitor Readout Is Not Evidence of Behavioral Control | Post-training with verifiable rewards can induce reward hacking, motivating the use of monitors within the training objective rather t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond Random Splits: Evaluating Drug-Target Affinity Models Under Chemically and Biologically Motivated Distribution Shifts Copy | Drug-target affinity (DTA) prediction is widely used to prioritize candidate compounds before costly experimental screening. DTA model… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Measure Less, Know More: Self-Supervised Test-Time Feature Acquisition🔧 Modal | Recent progress in multimodal, high-dimensional learning has enabled foundation models to process heterogeneous, large-scale data. How… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Causal Representation Learning with Instantaneous and Lagged Relations via Nonstationarity | Causal representation learning for time-series data aims to identify latent states and their causal relations from observations. In th… | ↗Not briefed |
hn hn-ai | 1mo | Best app stack of 2026, according to LLMs: Next.js, Supabase, Vercel, Stripe | — | ↗Not briefed |
hn hn-ai | 1mo | GitHub repository landing pages now show an accessibility tab, if provided | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Passing the Test You Trained On: Re-evaluating Prompt-Injection Detectors for LLM Agents | LLM agents increasingly screen tool outputs with small prompt-injection detectors, and teams choose among detectors by their scores on… | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLMs were always pointless because words themselves are empty. Language is just a vessel to carry parts of us from one to another. Words wi… | LLMs were always pointless because words themselves are empty. Language is just a vessel to carry parts of us from one to another. Wor… | ↗Not briefed |
hn hn-ai | 1mo | App's front end UI is now optional | — | ↗Not briefed |
hn hn-ai | 1mo | One month coding with GLM 5.3 Flash | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | What if an LLM had as much intelligence and self-awareness and capacity for pain as, say, a chicken? What would that require of us, morally… | What if an LLM had as much intelligence and self-awareness and capacity for pain as, say, a chicken? What would that require of us, mo… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI's wandering AI agents earn it a California subpoena | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Status lines in the Claude desktop app | The new modding functionality made it possible to add our beloved status line into the Claude desktop app, so I whipped this up.<p>Rep… | ↗Not briefed |
hn hn-ai | 1mo | PewDiePie uncensors Qwen3.5-9B with Heretic | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Corrupted but Correct: Why Vision-Language Models Lie to Themselves Internally | A targeted adversarial perturbation can drive a vision-language model's (VLM's) teacher-forced training loss for a fixed target captio… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Electronic Density versus Geometry for Machine-Learned Molecular Absorption Spectra | Molecular optical absorption spectroscopy provides a direct probe of electronic structure and is widely used for molecular identificat… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Pope Leo XIV presented his first encyclical at the Vatican on May 25 and set the fight over AI beside the Church’s long campaign against nu… | Pope Leo XIV presented his first encyclical at the Vatican on May 25 and set the fight over AI beside the Church’s long campaign again… | ↗Not briefed |
rss marktechpost | 1mo | Datalab Introduces OmniExtractBench to Fix Bias and Opacity in Extraction Benchmarks | Content-based row matching, 6 per-value verdicts and a null rule make OmniExtractBench an extraction benchmark anyone can audit. The p… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic posted a loss of eight billion in 2025 Vindicated! | Anthropic posted a loss of eight billion in 2025 Vindicated! | ↗Not briefed |
bluesky bluesky-ai | 1mo | nypost.com/2026/09/30/b... | nypost.com/2026/09/30/b... | ↗Not briefed |
rss huggingface-blog | 1mo | Open-sourcing AstaBrief, the fast report-generation model in Asta | — | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: Why do all LLM interfaces look the same? | They all seem to have the same exact layout (conversations on the left, primary chat exchange on the right, resources and windows on t… | ↗Not briefed |
hn hn-ai | 1mo | Knowledge platform powered by agents demonstration | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | bah. they're hosting some inhouse LLM models. but at least they do not tolerate vibecoding (un-checked pure AI code PRs) there is no corpo… | bah. they're hosting some inhouse LLM models. but at least they do not tolerate vibecoding (un-checked pure AI code PRs) there is no… | ↗Not briefed |
hn hn-ai | 1mo | What Dots Choose: OpenAI's Agent Has a Stack | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | At last, someone has managed to fuse interfaith dialogue with tech lobbying. www.nytimes.com/2026/09/29/u... | At last, someone has managed to fuse interfaith dialogue with tech lobbying. www.nytimes.com/2026/09/29/u... | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI has not and will never walk down a rainy street at night. They could hook it up to cameras on a big dumb walking robot that walks down t… | AI has not and will never walk down a rainy street at night. They could hook it up to cameras on a big dumb walking robot that walks d… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The guys who read this and think "Wow, OpenAI is killing it!!!!" also think they can make betting on prediction markets their new career | The guys who read this and think "Wow, OpenAI is killing it!!!!" also think they can make betting on prediction markets their new care… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It's so comforting that OpenAI and Anthropic leadership are insane in completely different ways. | It's so comforting that OpenAI and Anthropic leadership are insane in completely different ways. | ↗Not briefed |
arxiv arxiv-ai | 1mo | OptiSelect: How does the Optimizer Shape Data Curriculum? | Online data selection has demonstrated substantial efficiency gains for LLM pretraining by training on the most valuable candidates wi… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Jumping the Line: Exploiting Length Predictions in LLM Scheduling | Efficient request scheduling is increasingly important for reducing completion time in large language model (LLM) serving. Size-based… | ↗Not briefed |
hn hn-ai | 1mo | Interview with Liam Quin | — | ↗Not briefed |
hn hn-ai | 1mo | Transition Off of Digital Ocean – Node.js | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | > hoping they would bless the idea that Claude has moral standing I know some AI boosters believe the LLMs are conscious or reasoning, for… | > hoping they would bless the idea that Claude has moral standing I know some AI boosters believe the LLMs are conscious or reasoning… | ↗Not briefed |
bluesky bluesky-ai | 1mo | the ai race in four charts… usa winning on computing power, capital investment china leading w/ global model usage, ai research | the ai race in four charts… usa winning on computing power, capital investment china leading w/ global model usage, ai research | ↗Not briefed |