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 | Nobody Truly Agrees on Sentiment: Humans, Bespoke Tools, and LLMs Struggle with Social Media Texts | Social media is a rich source of real-time public sentiment, but widely used sentiment analysis tools are often applied without unders… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Bleibt die Frage, wie unsere Arbeit in Zukunft aussehen soll. Haben wir noch Kompetenzen, die keine KI aufbringt? Und was bringen wir unser… | Bleibt die Frage, wie unsere Arbeit in Zukunft aussehen soll. Haben wir noch Kompetenzen, die keine KI aufbringt? Und was bringen wir… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Thinking in Depth: Retrospective Inference for Tabular Foundation Models | Tabular foundation models (TFMs) are pretrained across diverse tabular tasks and make predictions on a new table at inference time usi… | ↗Not briefed |
hn hn-ai | 1mo | Mac Do It Menu bar todo list powered by Shia LaBeouf's meme | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | PoreML: A Data-Driven Framework for Learning Multiphase Flow in Porous Media | Multiphase flow in porous microstructures is central to CO$_2$ storage, fuel-cell operation, and flip-chip packaging. Predicting these… | ↗Not briefed |
bluesky bluesky-ai | 1mo | BECAUSE PUTTING AI IN THE HANDS OF YOUNG PEOPLE GOING TO SCHOOL WORKS OUT SUPER DUPER WELL. www.bbc.com/news/article... | BECAUSE PUTTING AI IN THE HANDS OF YOUNG PEOPLE GOING TO SCHOOL WORKS OUT SUPER DUPER WELL. www.bbc.com/news/article... | ↗Not briefed |
hn hn-ai | 1mo | Study: Claude, ChatGPT Offer Different Shopping Prices Based on Wealth🔧 ChatGPT | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | It's about #innovating. I just asked Claude to cook my kids breakfast. The time savings for this would be immeasurable -- time is money, it… | It's about #innovating. I just asked Claude to cook my kids breakfast. The time savings for this would be immeasurable -- time is mone… | ↗Not briefed |
hn hn-ai | 1mo | LHS – one language for ease and native code | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Fault-tolerant foundation models | Emerging computer hardware often trades reliability for energy efficiency; here we show that large-language models (LLMs) can be train… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Poetry is the new AI security threat as PoeLLM malware infects 3K+ servers | Poetry is the new AI security threat as PoeLLM malware infects 3K+ servers | ↗Not briefed |
bluesky bluesky-ai | 1mo | lol and "agentic ai" which are just llm prompts in a marketing trenchcoat. the perfect name for something with zero agency | lol and "agentic ai" which are just llm prompts in a marketing trenchcoat. the perfect name for something with zero agency | ↗Not briefed |
arxiv arxiv-ai | 1mo | RSIGym: A Flexible Environment for Recursive Self-Improvement | Recursive self-improvement requires carrying accepted changes into later improvement cycles, while studying agent-proposed changes als… | ↗Not briefed |
rss techmeme | 1mo | Game studio Gardens Interactive raised a $35M+ Series B led by Lightspeed at a pre-money valuation of $190M, taking its total funding to $71M (Cecilia D'Anastasio/Bloomberg) | Cecilia D'Anastasio / Bloomberg: Game studio Gardens Interactive raised a $35M+ Series B led by Lightspeed at a pre-money valuation of… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Close your eyes, and let the AI ransack the inner sanctum of your gray matter. | Close your eyes, and let the AI ransack the inner sanctum of your gray matter. | ↗Not briefed |
bluesky bluesky-ai | 1mo | I’ll admit that I don’t think any positive of LLMs is worth having to read LLM writing or the internet being awash with shitty genAI imager… | I’ll admit that I don’t think any positive of LLMs is worth having to read LLM writing or the internet being awash with shitty genAI i… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Individuals with early onset blindness can perceive and report what they experience.They trust their senses and integrate them fully. Beli… | Individuals with early onset blindness can perceive and report what they experience.They trust their senses and integrate them fully.… | ↗Not briefed |
bluesky bluesky-ai | 1mo | In a conversation with Vanity Fair on Sunday, Altman shared his thoughts on Luca Guadagnino’s upcoming film about his brief OpenAI ousting … | In a conversation with Vanity Fair on Sunday, Altman shared his thoughts on Luca Guadagnino’s upcoming film about his brief OpenAI ous… | ↗Not briefed |
arxiv arxiv-ai | 1mo | How Do Transformers Learn to Represent Symmetries? | Training Transformer-based architectures with finite data augmentation has become an increasingly popular approach in geometric machin… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SemanticFold: Latent Sequence Compression SeparatesLanguage Modeling, Decodability, and Reasoning | We study whether latent sequence compression of prompt prefixes preserves the capabilities that large language models rely on during i… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: CrowDB – an iceberg table storage single Docker container | I built a new storage platform that will provide modern storage interfaces such as Iceberg, datasets, and S3.<p>The single-node Docker… | ↗Not briefed |
rss microsoft-research | 1mo | Agent Lightning v1.0: A 3,500-Line Lightweight Agentic RL Framework for Training Agents with Real Harnesses | Training AI agents with reinforcement learning can be challenging because their tools, context, and decision-making are managed by com… | ↗Not briefed |
rss theverge-ai | 1mo | ChatGPT is getting college planning tools🔧 ChatGPT | A preview of OpenAI’s College Planner. | Image: OpenAI OpenAI is bringing new tools to ChatGPT for Teens, a mode for teens introduce… | ↗Not briefed |
bluesky bluesky-ai | 1mo | hasn't she been using it this whole time? i vaguely remember some wack llm pseudostatistics project she was involved in | hasn't she been using it this whole time? i vaguely remember some wack llm pseudostatistics project she was involved in | ↗Not briefed |
bluesky bluesky-ai | 1mo | this is how #anthropic tries to lure startups into their #AI trap #ecosystem www.techradar.com/pro/anthropi... #Europe stay away of this #M… | this is how #anthropic tries to lure startups into their #AI trap #ecosystem www.techradar.com/pro/anthropi... #Europe stay away of th… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Watching this video really comforts me in the idea that LLM vibe Coded games are not going make good game any time soon this shit is just… | Watching this video really comforts me in the idea that LLM vibe Coded games are not going make good game any time soon this shit is… | ↗Not briefed |
bluesky bluesky-ai | 1mo | If post-training makes it no longer an LLM then, ok, fine, no one is using LLMs in that case so questions about LLMs are kind of academic a… | If post-training makes it no longer an LLM then, ok, fine, no one is using LLMs in that case so questions about LLMs are kind of acade… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Continual Graph Multi-Agent Reinforcement Learning | In Continual Multi-Agent Reinforcement Learning (CMARL), agents learn cooperative policies across sequences of tasks, aiming to adapt… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Revisiting Explainable AI through Model-Independent Concept Dictionaries | Modern applications of AI rely on increasingly complex models. Explainable AI (XAI) has emerged as a set of techniques aimed at improv… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI partners for autonomous chip design 🦾 OpenAI and Synopsys are teaming up to develop "GPT-Synopsys," an AI model that will operate … | OpenAI partners for autonomous chip design 🦾 OpenAI and Synopsys are teaming up to develop "GPT-Synopsys," an AI model that will ope… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Not a lot of investment in using euphemisms for the eugenic fantasies anymore. | Not a lot of investment in using euphemisms for the eugenic fantasies anymore. | ↗Not briefed |
arxiv arxiv-ai | 1mo | Shared Gaussianization: What Gaussian Regularizers Certify About Contrastive Learning, and What They Miss | What can a distribution-matching regularizer such as SIGReg in LeJEPA certify about contrastive learning? We study shared Gaussianizat… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Physics-Aligned Electronic Ground-State Learning Improves Generalization | Machine-learned interatomic potentials (MLIPs) excel at in-distribution tasks, accelerating drug and material development, yet they st… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Energy-Efficient Gait Adaptation via Hierarchical Reinforcement Learning for Quadrupedal Locomotion Across Diverse Terrains | While energy efficiency is a critical objective for legged-robot locomotion control, achieving low energy consumption while maintainin… | ↗Not briefed |
hn hn-ai | 1mo | Live config for Spring Boot via MongoDB change streams | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think if you’re pivoting to being an AI doomer from a trans doomer you should at least learn what a model harness is | I think if you’re pivoting to being an AI doomer from a trans doomer you should at least learn what a model harness is | ↗Not briefed |
hn hn-ai | 1mo | An OpenAI Insider Explains Why He Quit [video] | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | AI Safety Considerations for Agents With Limited Time to Act | In the wake of the increasingly public discussion about AI alignment, recent work has tried to propose specific AI architectures that… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Temporal Visuo-Tactile Learning for Dexterous Grasp Stability | Humans can grasp everyday objects with almost perfect success rates using fingertip tactile feedback, yet much of the robotic grasping… | ↗Not briefed |
bluesky bluesky-ai | 1mo | deepseek gave up on steering and just shuts everything down the moment ANYTHING about china comes from user User: Could you please find m… | deepseek gave up on steering and just shuts everything down the moment ANYTHING about china comes from user User: Could you please f… | ↗Not briefed |
bluesky bluesky-ai | 1mo | What LLM wrote the script for this because "Space Corridor" and "Of course" followed by "...?" is frying the hell out of me | What LLM wrote the script for this because "Space Corridor" and "Of course" followed by "...?" is frying the hell out of me | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Trigora – durable execution without history replay | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Neural Sampling with Reweighted Normalizing Flows via the Wasserstein--Fisher--Rao JKO Scheme | We propose a neural algorithm for sampling from distributions specified by unnormalized Boltzmann densities. Our approach is based on… | ↗Not briefed |
bluesky bluesky-ai | 1mo | In retrospect, the whole efficient markets thing is fucking nonsense, because if you'd bought NVDA in like 2022 when ChatGPT came out, you'…🔧 ChatGPT | In retrospect, the whole efficient markets thing is fucking nonsense, because if you'd bought NVDA in like 2022 when ChatGPT came out,… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PatchBench: Measuring Collateral Damage in Activation Patching | An LLM safety patch can pass a benchmark while still being a poor repair. This risk is especially acute for jailbreak repairs, where t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | irrespective of how much AI you shove into the system you can’t really get rid of these pressures in an existing org. is it stupid to spend… | irrespective of how much AI you shove into the system you can’t really get rid of these pressures in an existing org. is it stupid to… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Srl – Angular-style web components with no build step | — | ↗Not briefed |
hn hn-ai | 1mo | Redundancy Is the Engine of Progress | — | ↗Not briefed |
hn hn-ai | 1mo | Boredpolymath/Truth-Beacon | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | and anyway, we'll know what the "frontier" for model censorship is via the Chinese models long before OpenAI or Musk get into it. like, do … | and anyway, we'll know what the "frontier" for model censorship is via the Chinese models long before OpenAI or Musk get into it. like… | ↗Not briefed |