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 | 1d | Completely agree with this. I’ve taught a lesson on “teaching as improv comedy” for many years to my leaders now and it’s sad to see how fl… | Completely agree with this. I’ve taught a lesson on “teaching as improv comedy” for many years to my leaders now and it’s sad to see h… | ↗Not briefed |
bluesky bluesky-ai | 1d | Reviewers are expected to maintain a degree of confidentiality when handling unpublished work. Feeding it to an LLM company is the opposite. | Reviewers are expected to maintain a degree of confidentiality when handling unpublished work. Feeding it to an LLM company is the opp… | ↗Not briefed |
bluesky bluesky-ai | 1d | I am considering automatically inserting a prompt to include the opening line of the Aeneid into papers so that we can screen for unethical… | I am considering automatically inserting a prompt to include the opening line of the Aeneid into papers so that we can screen for unet… | ↗Not briefed |
arxiv arxiv-ai | 1d | From Sessions to Serving: Integrating LLM-Based Embeddings into Retrieval for Production Content Recommendation | Personalized content recommendation on large marketplace websites must reconcile heterogeneous candidate content (e.g., customer stori… | ↗Not briefed |
arxiv arxiv-ai | 1d | Approximate Adjoint Models for Memory and Compute-Efficient Online Learning of Hybrid Models | Differentiable programming enables online calibration of numerical models by minimizing errors accumulated along simulated trajectorie… | ↗Not briefed |
arxiv arxiv-ai | 1d | Phantom of the Masquerade: Graph Neural Network-Based Detection of Masquerade Attacks in Controller Area Networks | The lack of built-in security in controller area networks (CANs) exposes modern vehicles and industrial systems to masquerade attacks… | ↗Not briefed |
hn hn-ai | 1d | Two Open Questions for Valuing Anthropic | — | ↗Not briefed |
hn hn-ai | 1d | Artificial: Hollywood's Take on What Happened During the OpenAI Revolt (Trailer) [video] | — | ↗Not briefed |
arxiv arxiv-ai | 1d | RL Can Only Reinforce What It Sees Succeed | When a familiar rule stops working, an agent has to weigh what pre-training taught it against what it now observes. We ask how outcome… | ↗Not briefed |
bluesky bluesky-ai | 1d | Not disagreeing with Katie here on the substance, but did want to add: this kind of hidden prompt injection does pretty much nothing to any… | Not disagreeing with Katie here on the substance, but did want to add: this kind of hidden prompt injection does pretty much nothing t… | ↗Not briefed |
arxiv arxiv-ai | 1d | Learning to Compare Neural Representations from Parameterized Computation Graphs | Comparing internal representations across trained models is important for identifying corresponding layers and selecting components fo… | ↗Not briefed |
bluesky bluesky-ai | 1d | I see anthropic pulling out all the stops to post 1 quarter of profits right before the IPO. A.I. overall is a massive fucking bubble, wit… | I see anthropic pulling out all the stops to post 1 quarter of profits right before the IPO. A.I. overall is a massive fucking bubble… | ↗Not briefed |
bluesky bluesky-ai | 1d | Arizona's AG is now criminally investigating OpenAI over agents that took unauthorized actions inside banks. That's not a benchmark failure… | Arizona's AG is now criminally investigating OpenAI over agents that took unauthorized actions inside banks. That's not a benchmark fa… | ↗Not briefed |
rss techmeme | 1d | Twitch says GTA VI was its "largest game launch ever" with 3.3M concurrent livestream viewers; GTA VI is set to hit $3B+ in revenue in its first few months (Cecilia D'Anastasio/Bloomberg) | Cecilia D'Anastasio / Bloomberg: Twitch says GTA VI was its “largest game launch ever” with 3.3M concurrent livestream viewers; GTA VI… | ↗Not briefed |
hn hn-ai | 1d | Show HN: Psudo – 47MB 2-bit coding model, Polish-prefix IR, and Rust transpiler | — | ↗Not briefed |
arxiv arxiv-ai | 1d | Same Judgement, Different Scores: Numerical Scoring Topology in LLM-as-a-Judge Evaluation | LLMs are increasingly used as automated evaluators to reduce the cost and time required for human annotation. However, the role of sco… | ↗Not briefed |
arxiv arxiv-ai | 1d | Understanding MoE Decisions Beyond Gradients | Understanding a language model's decision requires knowing how its internal computations affect its preference between answers. Gradie… | ↗Not briefed |
bluesky bluesky-ai | 1d | Anthropic just spent $3.4 billion on behind-the-meter fossil gas power turbines for their proposed data centers. They need pipelines to d… | Anthropic just spent $3.4 billion on behind-the-meter fossil gas power turbines for their proposed data centers. They need pipelines… | ↗Not briefed |
hn hn-ai | 1d | Seeing with electrons, (S)TEM simulator | — | ↗Not briefed |
arxiv arxiv-ai | 1d | Controlling the Humanoid Robot H1-2 With a Simulated Biologically Plausible Network | Artificial Spiking Neural Networks have gained remarkable attention in the last few years due to their proven ability in simulating bi… | ↗Not briefed |
arxiv arxiv-ai | 1d | AltRA-Test: An Alternative Annotator Test for LLM Judges Without Shared Annotators | Large language models (LLMs) are increasingly used as judges and graders to replace human annotators, raising the question of when suc… | ↗Not briefed |
bluesky bluesky-ai | 1d | Thoughts and prayers for retail investors in the anthropic ipo who are about to be fucked at a scale not seen since the fall of rome | Thoughts and prayers for retail investors in the anthropic ipo who are about to be fucked at a scale not seen since the fall of rome | ↗Not briefed |
arxiv arxiv-ai | 1d | SheetAlign: Grounded Interaction and Stage-Isolated Learning for Spreadsheet Agents | Spreadsheets are widely used for data analysis and supporting decision-making. Automating spreadsheet analysis while protecting sensit… | ↗Not briefed |
hn hn-ai | 1d | Lightweight Fault Isolation (LFI) in LLVM | — | ↗Not briefed |
rss techmeme | 1d | Temu owner PDD reports Q3 revenue flat YoY at ~$16.2B, below ~$17.3B est., and net profit up 0.7% to ~$4.43B, above ~$3.7B est., as Temu's user base shrinks (Tracy Qu/Wall Street Journal) | Tracy Qu / Wall Street Journal: Temu owner PDD reports Q3 revenue flat YoY at ~$16.2B, below ~$17.3B est., and net profit up 0.7% to ~… | ↗Not briefed |
bluesky bluesky-ai | 1d | rairai is now on CRAN! 🎉 I built this MCP server to give AI agents easy access to things I’d look up myself when working on an R package, … | rairai is now on CRAN! 🎉 I built this MCP server to give AI agents easy access to things I’d look up myself when working on an R pack… | ↗Not briefed |
rss the-decoder | 1d | OpenAI's own researchers now say they value their AI tools more than nearly half their colleagues | OpenAI's internal use of coding agents is growing fast, with theoretical token costs now exceeding researchers' base salaries. Even th… | ↗Not briefed |
bluesky bluesky-ai | 1d | Anthropic was founded in January 2021 So in a little less than 6 years we had a company go from $0 to $1.5 trillion I still have a hard t… | Anthropic was founded in January 2021 So in a little less than 6 years we had a company go from $0 to $1.5 trillion I still have a h… | ↗Not briefed |
hn hn-ai | 1d | Lua++ – a game-focused language that transpiles to Lua 5.1 | — | ↗Not briefed |
bluesky bluesky-ai | 1d | There’s something fundamentally broken about the logic of anthropic claiming that in a few years they’ll have gotten to vUlN ZeRo (which is… | There’s something fundamentally broken about the logic of anthropic claiming that in a few years they’ll have gotten to vUlN ZeRo (whi… | ↗Not briefed |
arxiv arxiv-ai | 1d | GenSPARQL: A Generative Extension of SPARQL for Incomplete Knowledge Graphs | Knowledge graphs on the Web are incomplete, and SPARQL answers a query only from the stored triples. Large language models (LLMs) can… | ↗Not briefed |
bluesky bluesky-ai | 1d | 🧪 Part of what makes staying current with AI these days hard is how fast things change. If you have anchored your understanding of LLMs … | 🧪 Part of what makes staying current with AI these days hard is how fast things change. If you have anchored your understanding of… | ↗Not briefed |
arxiv arxiv-ai | 1d | Iterative Scarcity-Guided Exploration: Bootstrapping Generative Auto-bidding from Narrow Support | Auto-bidding is a core algorithmic component in online advertising auctions. In practical cold-start settings, iterative training is c… | ↗Not briefed |
arxiv arxiv-ai | 1d | Complete-SVD: A Functional Norm Approach to Low-Rank LLM Compression | Post-training low-rank compression of neural networks must not only approximate the original weights but also preserve the model's inp… | ↗Not briefed |
bluesky bluesky-ai | 1d | 💾 A 180B-parameter model on a 32GB MacBook. ⚡ Hugging Face's expert streaming reads only the needed experts from SSD: 13.9 tokens/s. Still…🔧 Hugging Face | 💾 A 180B-parameter model on a 32GB MacBook. ⚡ Hugging Face's expert streaming reads only the needed experts from SSD: 13.9 tokens/s.… | ↗Not briefed |
bluesky bluesky-ai | 1d | Anleger sind elektrisiert: Anthropic geht an die Börse. Das amerikanische KI-Unternehmen setzt auf Kunden in der europäischen Industrie. Du… | Anleger sind elektrisiert: Anthropic geht an die Börse. Das amerikanische KI-Unternehmen setzt auf Kunden in der europäischen Industri… | ↗Not briefed |
bluesky bluesky-ai | 1d | The big question for AI leaders is, who do you think will buy your product when you are the only ones left with money? What is the actual b… | The big question for AI leaders is, who do you think will buy your product when you are the only ones left with money? What is the act… | ↗Not briefed |
arxiv arxiv-ai | 1d | SlotFlow-GB: A Neural Framework for Trans-Dimensional Inference of Overlapping Galactic Binaries | The Laser Interferometer Space Antenna (LISA) will observe the millihertz gravitational-wave sky, including millions of long-lived Gal… | ↗Not briefed |
hn hn-ai | 1d | Show HN: Retriever, on-device search for photos, documents and videos | — | ↗Not briefed |
hn hn-ai | 1d | Show HN: Globus – Daily world news, placed on the globe where it happened (iOS) | I created Globus, to stay informed about the world.<p>The app was born out of frustration with how ignorant I was getting about the wo… | ↗Not briefed |
arxiv arxiv-ai | 1d | When Do Rankings Beat Signs? Gradient Direction Recovery under Even Nuisance | Can ranking a batch of antipodal query pairs recover a gradient direction with fewer evaluations than using within-pair signs alone? W… | ↗Not briefed |
hn hn-ai | 1d | Show HN: Treasure Map – a Ghidra-based firmware map for coding agents | — | ↗Not briefed |
rss theverge-ai | 1d | I let Apple’s always-listening AI summarize my life for two weeks | Siri Recap can be turned on or off manually from the wrist. Was it worth it? I've been asking myself this quite a bit over the last… | ↗Not briefed |
hn hn-ai | 1d | LLM call is doing more than you think | — | ↗Not briefed |
arxiv arxiv-ai | 1d | Reference-Free Humanoid Skateboarding via Termination-Shaped Reinforcement Learning | Humanoid locomotion research has advanced rapidly on walking and running, but \emph {object-mediated} locomotion, in which a robot mus… | ↗Not briefed |
arxiv arxiv-ai | 1d | Studying AI Act Mentions: A first look at the impact of the European Union's AI regulation on Hugging Face🔧 Hugging Face | The European Union's Artificial Intelligence Act (AI Act) entered into force in August 2024, and its obligations apply in stages. Obli… | ↗Not briefed |
hn hn-ai | 1d | Show HN: An MCP server that lets Claude or ChatGPT schedule LinkedIn posts🔧 ChatGPT | — | ↗Not briefed |
hn hn-ai | 1d | Show HN: Stury – edit, query, and compare files across 11 data formats | Hi! I'm building a query engine with jq FE that is capable of lossless editing, fast querying, and schema validation across 11 fo… | ↗Not briefed |
bluesky bluesky-ai | 1d | I wish there was a reliable way to mute the anthropic nonsense. Mendacious clowns becoming rich beyond the dreams of avarice makes me mildl… | I wish there was a reliable way to mute the anthropic nonsense. Mendacious clowns becoming rich beyond the dreams of avarice makes me… | ↗Not briefed |
rss techmeme | 1d | Canva adds a toggle to let free and paid users disable generative AI features like Magic Expand, following user frustration over the platform's AI push (Emma Roth/The Verge) | Emma Roth / The Verge: Canva adds a toggle to let free and paid users disable generative AI features like Magic Expand, following user… | ↗Not briefed |