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 | VFold: Symmetry-Aware Cross-Layer Value Cache Compression | While caching key-value (KV) states accelerates Large Language Model (LLM) decoding, this cache can dominate memory usage at long cont… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Nicholas Carr likens LLM writing to JPGs becoming blurrier as they’re copied and shared on the web. They lose clarity with each iteration. … | Nicholas Carr likens LLM writing to JPGs becoming blurrier as they’re copied and shared on the web. They lose clarity with each iterat… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ContiLNN: Mitigating Slice Sampling Discontinuity with Liquid Neural Networks for Medical Image Restoration | Anatomical continuity provides complementary information for medical image restoration, but its use requires accounting for local anat… | ↗Not briefed |
arxiv arxiv-ai | 1mo | asdex: Automatic Sparse Differentiation in JAX | Many tasks in scientific computing and machine learning require the Jacobian or Hessian matrix of a function. Automatic differentiatio… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: AI SRE Arena, an Open Benchmark for AI SRE Agents on Kubernetes | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Wichtig: SynthID ist kein universeller KI-Detektor, sondern erkennt ausschließlich Inhalte, die ein SynthID-Wasserzeichen tragen." the-de… | "Wichtig: SynthID ist kein universeller KI-Detektor, sondern erkennt ausschließlich Inhalte, die ein SynthID-Wasserzeichen tragen." t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic wrote that the new policy update “is meant to apply only in extreme cases, where users repeatedly act cruelly toward our models, … | Anthropic wrote that the new policy update “is meant to apply only in extreme cases, where users repeatedly act cruelly toward our mod… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RiCo: Neural Simulation of Rigid-Body Interactions via Local Contact Reasoning | Accurate simulation of rigid-body interactions is essential for predictive physical world models. Despite recent progress in modeling… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Prediction-Powered Data Fusion for Treatment Effect Estimation | Randomized controlled trials (RCTs) identify treatment effects without confounding but are often small, whereas observational studies… | ↗Not briefed |
bluesky bluesky-ai | 1mo | So it's OpenAI going to do a full rebrand? | So it's OpenAI going to do a full rebrand? | ↗Not briefed |
bluesky bluesky-ai | 1mo | Ah, bad revenue news from OpenAI. Well that was easy to predict. To ever turn a profit these outfits will have to jack prices WAY up, and … | Ah, bad revenue news from OpenAI. Well that was easy to predict. To ever turn a profit these outfits will have to jack prices WAY up,… | ↗Not briefed |
rss techmeme | 1mo | Trump says the WH considers as "THE ENEMY" anyone who uses the term Artificial Intelligence, not "the highly accepted" and "more accurate" Super Intelligence (Donald J. Trump/@realDonaldTrump ...) | Donald J. Trump / @realDonaldTrump@truthsocial.com: Trump says the WH considers as “THE ENEMY” anyone who uses the term Artificial Int… | ↗Not briefed |
bluesky bluesky-ai | 1mo | In the past few weeks just by chance I’ve had to deal with multiple customer service LLM chatbots. Assessment: Abysmal. “Impressive” in t… | In the past few weeks just by chance I’ve had to deal with multiple customer service LLM chatbots. Assessment: Abysmal. “Impressive”… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Composite Online-to-Nonconvex Conversion with Optimal Oracle Complexity | We consider stochastic nonsmooth nonconvex composite optimization, which includes several important problems such as constrained optim… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SparseDecoding: Decoding-Aware Pruning for Accurate and Efficient LLM Inference | The memory-bound nature of the decoding stage of large language model (LLM) inference incurs significant latency. Layer-wise training-… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Prior or Feedback? What an LLM Uses When Adapting Neural Operators | Do LLM scientific agents rely only on their initial task context, or do they adapt their decisions in response to experimental feedbac… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI gonna be ClosedAI by the end of 2028. No way the make it to IPO without a major government bailout. | OpenAI gonna be ClosedAI by the end of 2028. No way the make it to IPO without a major government bailout. | ↗Not briefed |
bluesky bluesky-ai | 1mo | "OpenAI annualised revenues $20bn less than previously signalled" (FT) It "recently told investors (it) was approaching $50bn in September… | "OpenAI annualised revenues $20bn less than previously signalled" (FT) It "recently told investors (it) was approaching $50bn in Sept… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "According to a person with knowledge of the matter, the discrepancy arose from attempts by OpenAI’s own investors to produce a direct comp… | "According to a person with knowledge of the matter, the discrepancy arose from attempts by OpenAI’s own investors to produce a direct… | ↗Not briefed |
hn hn-ai | 1mo | XPCH Publisher Danielle Lambert | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Spatial Pattern Formation from Multi-Agent Learning in Public Goods Dilemmas | Spatial public goods models show that prescribed movement toward richer locations can generate spatial patterns. We ask how such patte… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OOPSIE-AI strikes again I blame the agents OpenAI annualised revenues $20bn less than previously signalled - as.ft.com/r/fcd72bb4-1... | OOPSIE-AI strikes again I blame the agents OpenAI annualised revenues $20bn less than previously signalled - as.ft.com/r/fcd72bb4-1… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It's so WEIRD being a person who is like "I am not reflexively anti-AI but holy shit OpenAI was literally the worst set of people to bring … | It's so WEIRD being a person who is like "I am not reflexively anti-AI but holy shit OpenAI was literally the worst set of people to b… | ↗Not briefed |
rss theverge-ai | 1mo | Anthropic bans ‘abusive or cruel behavior’ towards Claude | Anthropic is making changes to its usage policy for the first time in over a year to reflect new and high-risk cases of misuse - inclu… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Verdict Without the Rule: Diagnosing and Auditing Regulatory Rule Sensitivity in LLM Compliance Systems | Large language model compliance systems are deployed on the assumption that a verdict depends on the regulatory rule it is given. We t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Tech bros are out here absolutely determined to prove me right about how all societal ills stem from how we treat children youtu.be/EnhUo… | Tech bros are out here absolutely determined to prove me right about how all societal ills stem from how we treat children youtu.be/… | ↗Not briefed |
bluesky bluesky-ai | 1mo | just the implications of your rental application being run thru an AI model that (accurately or in error) flags your income verification as… | just the implications of your rental application being run thru an AI model that (accurately or in error) flags your income verificati… | ↗Not briefed |
hn hn-ai | 1mo | OpemAI mathematics papers have no correspondance address | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think at this point all we’ve really learned about AI is that about 40% of Americans are dumber than an LLM | I think at this point all we’ve really learned about AI is that about 40% of Americans are dumber than an LLM | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think a big aspect of this is there's always been a group of people who view human interaction as an onerous obligation (also: LLM fantas… | I think a big aspect of this is there's always been a group of people who view human interaction as an onerous obligation (also: LLM f… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It might. One thing that *could* happen is that as people use LLMs more, personal LLMs that do news functions could effectively reduce tx c… | It might. One thing that *could* happen is that as people use LLMs more, personal LLMs that do news functions could effectively reduce… | ↗Not briefed |
bluesky bluesky-ai | 1mo | their lack of ability to connect on a human level is constantly on display. the best they can do is replicate what they observe empathy to …🔧 Replicate | their lack of ability to connect on a human level is constantly on display. the best they can do is replicate what they observe empath… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Structural Theory of Cognitive Representation and Problem Solving,Contexts, Invariance, and the Knowledge Space | Learning and problem solving depend critically on the structure of internal representations. While many modern data-driven artificial… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Looking Inside LLMs: Small-World Connectivity as a Signature of Reasoning Performance | Understanding large language model (LLM) reasoning requires looking beyond behavioral performance to examine how reasoning ability is… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Nasdaq 100 tanking on George and Stephen's scoop www.ft.com/content/b66a... | Nasdaq 100 tanking on George and Stephen's scoop www.ft.com/content/b66a... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning Probabilistic Logic Programs with Functional Gradient Guided Language Models | Declarative logic programs offer a powerful and interpretable abstraction for encoding relational structure and neurosymbolic reasonin… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ChatGPT can make mistakes because it is an LLM that predicts words based on patterns rather than possessing real-world understanding www.ft…🔧 ChatGPT | ChatGPT can make mistakes because it is an LLM that predicts words based on patterns rather than possessing real-world understanding w… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I love that the company which is like "a mass causality event is fine if it means we get AI out of it" is the one selling an AI personal as… | I love that the company which is like "a mass causality event is fine if it means we get AI out of it" is the one selling an AI person… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Docs: OpenAI told investors its annualized revenue was nearing $50B at the end of September, well below the $70B reported by media based on… | Docs: OpenAI told investors its annualized revenue was nearing $50B at the end of September, well below the $70B reported by media bas… | ↗Not briefed |
rss techmeme | 1mo | Docs: OpenAI told investors its annualized revenue was nearing $50B at the end of September, well below the $70B reported by media based on investor docs (Financial Times) | Financial Times: Docs: OpenAI told investors its annualized revenue was nearing $50B at the end of September, well below the $70B repo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Adobe's legal team. But crazy that this is possible. "Hey Ai, just copy paste that product and re-release as open source" How do you defe… | Adobe's legal team. But crazy that this is possible. "Hey Ai, just copy paste that product and re-release as open source" How do you… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ENEMYYYYYYYYYYY!!! www.ft.com/content/b66a... | ENEMYYYYYYYYYYY!!! www.ft.com/content/b66a... | ↗Not briefed |
bluesky bluesky-ai | 1mo | *OPENAI RECENTLY TOLD INVESTORS ANNUALIZED REVENUE NEAR $50B: FT whoops lol | *OPENAI RECENTLY TOLD INVESTORS ANNUALIZED REVENUE NEAR $50B: FT whoops lol | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI annualised revenues $20bn less than previously signalled ft.trib.al/3sfynkT | OpenAI annualised revenues $20bn less than previously signalled ft.trib.al/3sfynkT | ↗Not briefed |
bluesky bluesky-ai | 1mo | Who among us hasn't misstated revenue by $20,000,000,000? | Who among us hasn't misstated revenue by $20,000,000,000? | ↗Not briefed |
arxiv arxiv-ai | 1mo | Multi-Agent Egocentric World Model with Fine-Grained Embodied Interaction | Egocentric world models predict first-person observations conditioned on an agent's actions, but most focus on a single agent. Real em… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I mean, people have been utter dicks to open source developers for years now, just that the advent of AI/LLM sort of gave them a focus for … | I mean, people have been utter dicks to open source developers for years now, just that the advent of AI/LLM sort of gave them a focus… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I fear we’re ignoring the reality that there are people who think it’s a fine idea to throw our rulebooks into an LLM and have it spit out … | I fear we’re ignoring the reality that there are people who think it’s a fine idea to throw our rulebooks into an LLM and have it spit… | ↗Not briefed |
bluesky bluesky-ai | 1mo | This just reminded me of a recent incident: OpenAI GPT-6 Astra resorted to cheating when losing against human-created Starcraft bots. The L… | This just reminded me of a recent incident: OpenAI GPT-6 Astra resorted to cheating when losing against human-created Starcraft bots.… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anecdotally: in my actual research work, the amount of human errors, including misquotations, that LLMs have been catching recently greatly… | Anecdotally: in my actual research work, the amount of human errors, including misquotations, that LLMs have been catching recently gr… | ↗Not briefed |