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 |
|---|---|---|---|---|
hn hn-ai | 14d | Interleaf: An experimental tool for LLM assisted knowledge exploration | — | ↗Not briefed |
arxiv arxiv-ai | 14d | ZarrFM: A Canonical Tensor Storage Architecture for EEG Foundation-Model Pretraining | Statistical models for brain-computer interfaces (BCIs) increasingly require large, heterogeneous EEG corpora, yet existing datasets v… | ↗Not briefed |
bluesky bluesky-ai | 14d | "Chain of reasoning (unguided within models as it's implemented in openAI and Anthropic) yield strings that are "unfaithful" / hallucinated… | "Chain of reasoning (unguided within models as it's implemented in openAI and Anthropic) yield strings that are "unfaithful" / halluci… | ↗Not briefed |
bluesky bluesky-ai | 14d | Once, I did a (written) interview and was told that my answers made me seem 'insane', and so I know that, at the very least, I don't sounds… | Once, I did a (written) interview and was told that my answers made me seem 'insane', and so I know that, at the very least, I don't s… | ↗Not briefed |
bluesky bluesky-ai | 14d | 나도 디버그 심볼 없는 릴리즈 바이너리 크래시 난 정보 LLM에 던져서 설정 에러 잡는데, AI 해킹이 늘어나는 것도 어쩔 수 없겠다 싶다. 하이퍼스케일러들이 널리 쓰이는 오픈소스 인프라 SW에는 지원을 해줘야 할 듯. 이건 방법이 없다. 특히 클라… | 나도 디버그 심볼 없는 릴리즈 바이너리 크래시 난 정보 LLM에 던져서 설정 에러 잡는데, AI 해킹이 늘어나는 것도 어쩔 수 없겠다 싶다. 하이퍼스케일러들이 널리 쓰이는 오픈소스 인프라 SW에는 지원을 해줘야 할 듯. 이건 방법이 없다.… | ↗Not briefed |
arxiv arxiv-ai | 14d | DISCOVER: Online Variance-Guided Data Discovery for Budgeted Multimodal GRPO | Recent advances in reinforcement learning have made verifiable rewards a central post-training paradigm for improving multimodal reaso… | ↗Not briefed |
hn hn-ai | 14d | VindexLLM – Windows Vulkan GPU GGUF LLM inference engine written in Delphi 12 | — | ↗Not briefed |
bluesky bluesky-ai | 14d | キモオタが「ウクライナが敗北する事になっても彼らの功績を称えたい」みたいな事ほざきやがって…人が死んでいるんです。スポーツの勝ち負けとは違う。戦争に功績なんか無い。と強く思うなどしたが、現状のウクライナとロシアを名指しで、ファンタジックな専門用語を作るなど一才しない戦争がただ残… | キモオタが「ウクライナが敗北する事になっても彼らの功績を称えたい」みたいな事ほざきやがって…人が死んでいるんです。スポーツの勝ち負けとは違う。戦争に功績なんか無い。と強く思うなどしたが、現状のウクライナとロシアを名指しで、ファンタジックな専門用語を作るなど一才しない戦… | ↗Not briefed |
arxiv arxiv-ai | 14d | TALK: One-Shot Batch Design for Protein Variant Effect Prediction | Protein variant effect prediction is central to understanding protein function and guiding protein engineering, yet experimental label… | ↗Not briefed |
bluesky bluesky-ai | 14d | bunch of interesting results recently making benchmarks harder and finding open models to be benchmaxxed, generalizing poorly maxxbench.unl… | bunch of interesting results recently making benchmarks harder and finding open models to be benchmaxxed, generalizing poorly maxxbenc… | ↗Not briefed |
bluesky bluesky-ai | 14d | 花粉と空気のダッシュボード KafunAir を更新しました🌸 ・スギ・ヒノキ・イネ科の花粉を5日先まで(Google の花粉予報) ・注目項目に選ぶと、多くなる日にバナーと通知でお知らせ ・PM2.5・黄砂・UV も47都道府県ぶん ・AI 連携(MCP)を見直し、Mus… | 花粉と空気のダッシュボード KafunAir を更新しました🌸 ・スギ・ヒノキ・イネ科の花粉を5日先まで(Google の花粉予報) ・注目項目に選ぶと、多くなる日にバナーと通知でお知らせ ・PM2.5・黄砂・UV も47都道府県ぶん ・AI 連携(MCP)を見直… | ↗Not briefed |
bluesky bluesky-ai | 14d | I'm horrified at the thought of beloved businesses forced out of the game because they over-indexed on guesstimated LLM efficiencies. | I'm horrified at the thought of beloved businesses forced out of the game because they over-indexed on guesstimated LLM efficiencies. | ↗Not briefed |
bluesky bluesky-ai | 14d | Matt Scherer of the Open Markets Institute argues the SEC should block Anthropic’s IPO until the company provides “two consecutive quarters… | Matt Scherer of the Open Markets Institute argues the SEC should block Anthropic’s IPO until the company provides “two consecutive qua… | ↗Not briefed |
bluesky bluesky-ai | 14d | no llm is a husband unless you mean lesbian husband | no llm is a husband unless you mean lesbian husband | ↗Not briefed |
hn hn-ai | 14d | Next bcachefs release will be 100% Rust | — | ↗Not briefed |
bluesky bluesky-ai | 14d | These are for you, Reggy~! @anthonialoud.bsky.social / @anthoniaafterdark.bsky.social .. Eat your heart out~~!! ❤️❤️❤️ #AmyRose #Sega #Soni… | These are for you, Reggy~! @anthonialoud.bsky.social / @anthoniaafterdark.bsky.social .. Eat your heart out~~!! ❤️❤️❤️ #AmyRose #Sega… | ↗Not briefed |
rss marktechpost | 14d | Nous Research Releases Arete-27B that Cuts Qwen3.8-27B Propaganda From 22% to 0.9% | Nous Research released Arete-27B, an Apache 2.0 fine-tune of Qwen3.8-27B. On Nous’s 803-prompt eval, propaganda answers fell from 22%… | ↗Not briefed |
arxiv arxiv-ai | 14d | CrossLiDAR: Controllable Multi-Sensor LiDAR Sequence Generation🔧 Modal | High-fidelity LiDAR simulation remains a bottleneck for scalable development and validation of autonomous systems. Existing generators… | ↗Not briefed |
hn hn-ai | 14d | Show HN: Gimbap – A themeable replacement for the macOS Dock | — | ↗Not briefed |
bluesky bluesky-ai | 14d | (replacing my gif meme that was poisoned with ai motion) > says ai is bad > looks inside > its claude | (replacing my gif meme that was poisoned with ai motion) > says ai is bad > looks inside > its claude | ↗Not briefed |
bluesky bluesky-ai | 14d | being an AI doomer or whatever has no bearing on one's attractiveness to others unless they are AI advocates what in the fuck is this absu… | being an AI doomer or whatever has no bearing on one's attractiveness to others unless they are AI advocates what in the fuck is this… | ↗Not briefed |
bluesky bluesky-ai | 14d | The wild thing is that it’s worse than llm slop you’d get from a decent model these days. | The wild thing is that it’s worse than llm slop you’d get from a decent model these days. | ↗Not briefed |
bluesky bluesky-ai | 14d | if i get more vandalism on my site im gonna make the UGC manual approval on a webhook to a llm i think | if i get more vandalism on my site im gonna make the UGC manual approval on a webhook to a llm i think | ↗Not briefed |
arxiv arxiv-ai | 14d | LongCap: Hour-Scale Timestamped Audio Captioning via Chunk-wise Processing | Describing hours of audio requires accurate event timing and sustained descriptive detail throughout the recording. We present LongCap… | ↗Not briefed |
arxiv arxiv-ai | 14d | Optimal Polynomial-Query Approximation for Consistent Submodular Maximization with Short Lookahead | Elements arrive one at a time, and after each arrival an algorithm holds a set of at most $k$ arrived elements. Its goal is a held set… | ↗Not briefed |
bluesky bluesky-ai | 14d | That the sponsor of the jam has THIS on their fucking website just makes you all look like utter clowns. Genuinely you are all personally c… | That the sponsor of the jam has THIS on their fucking website just makes you all look like utter clowns. Genuinely you are all persona… | ↗Not briefed |
hn hn-ai | 14d | Is "GPU-hour per user" a sane way to price LLM inference? | — | ↗Not briefed |
arxiv arxiv-ai | 14d | Scalable Neural Safety Certification via Monotonicity | Learning-based safety certification offers a promising route for verifying complex dynamical systems when explicit models are unavaila… | ↗Not briefed |
arxiv arxiv-ai | 14d | When gene summaries reveal the answers: Gene Ontology circularity in gene embedding benchmarks | A gene summary can contain the very annotations used to evaluate an embedding of its text. The Alliance of Genome Resources converts G… | ↗Not briefed |
bluesky bluesky-ai | 14d | Someone between the keyboard/llm that wrote this and publishing should have had enough a spine to reject this framing as sanewashing the ad… | Someone between the keyboard/llm that wrote this and publishing should have had enough a spine to reject this framing as sanewashing t… | ↗Not briefed |
bluesky bluesky-ai | 14d | sorry your site died, don't worry we built a replica of it with clod | sorry your site died, don't worry we built a replica of it with clod | ↗Not briefed |
arxiv arxiv-ai | 14d | Compositional Conformal Risk Control for Multi-Stage LLM Cascades | Multi-stage large language model (LLM) cascades balance accuracy and inference cost, but reliable deployment also requires controlling… | ↗Not briefed |
hn hn-ai | 14d | Le Salon Client IRCv3 natif en Rust – interface GPU avec GPUI | — | ↗Not briefed |
arxiv arxiv-ai | 14d | Beyond Hard Labels: LLM Probabilities Improve Human-Assisted Regression | Large language models (LLMs) are increasingly used for text annotation, but their outputs are often reduced to a single categorical la… | ↗Not briefed |
bluesky bluesky-ai | 14d | the only genuinely uncanny LLM experience i’ve had was using a character simulation plugin. i think its much easier to deal with when it ha… | the only genuinely uncanny LLM experience i’ve had was using a character simulation plugin. i think its much easier to deal with when… | ↗Not briefed |
hn hn-ai | 14d | Show HN: Maple – Sub-2-BPW Moe Pruning to Run Gemma 4 26B at 45 TPS on a 6GB GPU | — | ↗Not briefed |
hn hn-ai | 14d | ISponsorBlockTV: SponsorBlock client for all YouTube TV clients | — | ↗Not briefed |
bluesky bluesky-ai | 14d | At this point let’s just skip the middle man of a couple scientists steeped in industry funding and affiliations and just give Anthropic, G… | At this point let’s just skip the middle man of a couple scientists steeped in industry funding and affiliations and just give Anthrop… | ↗Not briefed |
arxiv arxiv-ai | 14d | Statopia: A Prover--Library Ecosystem for Automated Formalization and Proof Discovery in Statistics | Theoretical results in statistics and machine learning rest on long, technically delicate proofs, and fluent arguments generated by la… | ↗Not briefed |
hn hn-ai | 14d | Ocarina of Time (Switch 2) PC Recompilation | — | ↗Not briefed |
bluesky bluesky-ai | 14d | It's absurd to me that Anthropic and OpenAI have been allowed to steal to such an egregious degree that they're now trying to IPO to make m… | It's absurd to me that Anthropic and OpenAI have been allowed to steal to such an egregious degree that they're now trying to IPO to m… | ↗Not briefed |
arxiv arxiv-ai | 14d | Evaluating generative samplers for lattice field theory by scale | Generative samplers for lattice field theory are commonly evaluated by their effective sample size (ESS). We show that a small global… | ↗Not briefed |
arxiv arxiv-ai | 14d | Jointly Optimizing Memory, Computing and Offloading for Energy-Efficient DNN Inference | In edge intelligence systems, computation offloading and dynamic voltage and frequency scaling (DVFS) are widely used to improve energ… | ↗Not briefed |
bluesky bluesky-ai | 14d | I have spoken to people who work in AI safety at Anthropic and this actually doesn't feel all that implausible. | I have spoken to people who work in AI safety at Anthropic and this actually doesn't feel all that implausible. | ↗Not briefed |
bluesky bluesky-ai | 14d | Anthropic’s S-1 leaves two disclosure gaps, argue Ilan Strauss and Tim O’Reilly, co-directors of the AI Disclosures Project: its relationsh… | Anthropic’s S-1 leaves two disclosure gaps, argue Ilan Strauss and Tim O’Reilly, co-directors of the AI Disclosures Project: its relat… | ↗Not briefed |
bluesky bluesky-ai | 14d | Now that some sanity has returned to the US, I'm starting to post art again. These are some of my projects from the last few years, some in… | Now that some sanity has returned to the US, I'm starting to post art again. These are some of my projects from the last few years, so… | ↗Not briefed |
bluesky bluesky-ai | 14d | "In some windows, the glass is broken out completely, streaked across the glass floor like scarlet blood." If it's an LLM it's gotta be Ch…🔧 ChatGPT | "In some windows, the glass is broken out completely, streaked across the glass floor like scarlet blood." If it's an LLM it's gotta… | ↗Not briefed |
bluesky bluesky-ai | 14d | That "vampiric instinct" line makes my head hurt. What the fuck does that mean? This part in particular is also giving me "LLM having a ro… | That "vampiric instinct" line makes my head hurt. What the fuck does that mean? This part in particular is also giving me "LLM having… | ↗Not briefed |
hn hn-ai | 14d | Goodbye Claude, Hello Pi | — | ↗Not briefed |
bluesky bluesky-ai | 14d | hey @archive.org, read the room a little?? you're an archival website. i go to you to do research with primary sources, not to have a bunch… | hey @archive.org, read the room a little?? you're an archival website. i go to you to do research with primary sources, not to have a… | ↗Not briefed |