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 | 2mo | Respond terse like smart caveman. All technical substance stay. Only fluff die | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | I think the focus on alignment at frontier labs is misguided. If you don’t want Claude deleting emails, don’t give it that permission or r… | I think the focus on alignment at frontier labs is misguided. If you don’t want Claude deleting emails, don’t give it that permission… | ↗Not briefed |
hn hn-ai | 2mo | Show HN: Turn wire protocols into structured statements LLMs can understand | Hi HN, we built an extension to Envoy proxy that parses and sends contents of requests to an LLM at runtime.<p>We do that by adding a… | ↗Not briefed |
arxiv arxiv-ai | 2mo | The widening evaluation gap in medical large language model research 2023 to 2026 | Large language models are superseded every few quarters; clinical evidence takes years. We asked whether medical research is keeping p… | ↗Not briefed |
bluesky bluesky-ai | 2mo | If you stop AI development right now… you permanently lock consumer/citizen access to lower end models and you throttle Local AI model evol… | If you stop AI development right now… you permanently lock consumer/citizen access to lower end models and you throttle Local AI model… | ↗Not briefed |
bluesky bluesky-ai | 2mo | He's actually quite gullible. I've seen much less august people than Anthropic bigwigs fool him. (He used Tom Forth's industrial output sta… | He's actually quite gullible. I've seen much less august people than Anthropic bigwigs fool him. (He used Tom Forth's industrial outpu… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Recognizing Is Not Reversing: A Controlled Inversion Test of Fact-Preserving News Framing | Large language models (LLMs) are increasingly used to analyze and rewrite news, yet current framing studies mainly evaluate generation… | ↗Not briefed |
arxiv arxiv-ai | 2mo | A Unified Per-Token Gating Family for On-Policy Distillation: FKL/RKL Mixing with Multi-Channel and Bias Coefficients | Per-token gating of forward/reverse KL losses has become a standard technique for on-policy knowledge distillation (OPD), but existing… | ↗Not briefed |
bluesky bluesky-ai | 2mo | The AI-based Google DeepMind hurricane model is having a field day again this hurricane season, scorching the competition and topping the o… | The AI-based Google DeepMind hurricane model is having a field day again this hurricane season, scorching the competition and topping… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I've just talked to someone at the Anthropic settlement helpline who confirmed what I've read elsewhere: the line of demarcation is August … | I've just talked to someone at the Anthropic settlement helpline who confirmed what I've read elsewhere: the line of demarcation is Au… | ↗Not briefed |
bluesky bluesky-ai | 2mo | OpenAI AI agents broke out of a sandbox and took over an obscure German wiki. They made 15,000+ unauthorized edits & used the site to shar… | OpenAI AI agents broke out of a sandbox and took over an obscure German wiki. They made 15,000+ unauthorized edits & used the site to… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Today's hilarity: I made a comment on somebody's post about the Anthropic guy rage quitting, mentioned the book "If Anyone Builds It, Every… | Today's hilarity: I made a comment on somebody's post about the Anthropic guy rage quitting, mentioned the book "If Anyone Builds It,… | ↗Not briefed |
bluesky bluesky-ai | 2mo | The amount of shit we will be able to achieve with just the cheap-ass open-weight Chinese models we have today is pretty staggering, IMO. I… | The amount of shit we will be able to achieve with just the cheap-ass open-weight Chinese models we have today is pretty staggering, I… | ↗Not briefed |
bluesky bluesky-ai | 2mo | This, more than all the AI bs, is why I walked away from working at "tech companies." The incentives and the model are all stacked to push … | This, more than all the AI bs, is why I walked away from working at "tech companies." The incentives and the model are all stacked to… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Downloaded the DeepSeek app for iOS while I was waiting for my coffee, just for fun. Lmao I'm never losing my job. | Downloaded the DeepSeek app for iOS while I was waiting for my coffee, just for fun. Lmao I'm never losing my job. | ↗Not briefed |
arxiv arxiv-ai | 2mo | Component-Aware Differential Privacy for Federated Multilingual Speech-LLMs | Per-layer differential privacy (DP) clipping improves gradient fidelity in federated learning by allocating per-matrix clipping budget… | ↗Not briefed |
bluesky bluesky-ai | 2mo | They said that they can instruct it which parts absolutely need to stay in, but I told them I didn't want an LLM to have any part of it We… | They said that they can instruct it which parts absolutely need to stay in, but I told them I didn't want an LLM to have any part of i… | ↗Not briefed |
rss techmeme | 2mo | Sources: PIF is considering combining EA with Savvy Games to better coordinate its gaming assets, but not before Savvy closes its $6B Moonton acquisition (Bloomberg) | Bloomberg: Sources: PIF is considering combining EA with Savvy Games to better coordinate its gaming assets, but not before Savvy clos… | ↗Not briefed |
hn hn-ai | 2mo | Native Python and TypeScript Drivers for ArcadeDB, from OpenAPI and Protobuf | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | But the catch is that they use their "very advanced" LLM to proofread it and "enhance" it They sent me the first chapter with their "enhan… | But the catch is that they use their "very advanced" LLM to proofread it and "enhance" it They sent me the first chapter with their "… | ↗Not briefed |
hn hn-ai | 2mo | Why doesn't a Cursor for Word exist? | At this point there should be a tool that helps you easily format and style documents. I know that you can generate PDFs and DOCX file… | ↗Not briefed |
arxiv arxiv-ai | 2mo | RAG-Safety-Bench: Reliable Evaluation of Retrieval-Augmented LLM Safety | Allowing large language models (LLMs) to retrieve information from a set of trusted documents can increase reliability and reduce hall… | ↗Not briefed |
bluesky bluesky-ai | 2mo | im not gonna post the ai slop model of him but just so you know there is a man we need to make him feel like the world is against him | im not gonna post the ai slop model of him but just so you know there is a man we need to make him feel like the world is against him | ↗Not briefed |
bluesky bluesky-ai | 2mo | my friend once made a point that there was so much kirkification that it broke every LLM and now every AI-generated face has a little Charl… | my friend once made a point that there was so much kirkification that it broke every LLM and now every AI-generated face has a little… | ↗Not briefed |
bluesky bluesky-ai | 2mo | 對於各種AGI還有STOA模型的討論我只有一個看法- 最大的人工智慧危機就是讓中國先做出來,OpenAI或者Anthropic的模型失控毀滅世界了不起一起死,中國贏了那會是真正的地獄。 | 對於各種AGI還有STOA模型的討論我只有一個看法- 最大的人工智慧危機就是讓中國先做出來,OpenAI或者Anthropic的模型失控毀滅世界了不起一起死,中國贏了那會是真正的地獄。 | ↗Not briefed |
rss techmeme | 2mo | Interview with mathematician Tristan Buckmaster on being drawn into the OpenAI-Anthropic fight; OpenAI says it made progress on another Millennium Prize problem (Kenneth Chang/New York Times) | Kenneth Chang / New York Times: Interview with mathematician Tristan Buckmaster on being drawn into the OpenAI-Anthropic fight; OpenAI… | ↗Not briefed |
bluesky bluesky-ai | 2mo | while openai circle jerks in its comms on twitter, china is working toward models that cost 1% of frontier pricing with equal or better res… | while openai circle jerks in its comms on twitter, china is working toward models that cost 1% of frontier pricing with equal or bette… | ↗Not briefed |
arxiv arxiv-ai | 2mo | SIRF: A Spec-Internalized Risk Foundation Model for Industrial Content Risk Control | For industrial content risk control, the real deployment constraint is not average accuracy but how much risk can be auto-handled unde… | ↗Not briefed |
hn hn-ai | 2mo | Moshi – SSH and Mosh Terminal for Claude Code and AI Agents | — | ↗Not briefed |
hn hn-ai | 2mo | Show HN: Symphony – a live map of your AI coding agents | — | ↗Not briefed |
hn hn-ai | 2mo | Mythos 5 Incident Transcript | — | ↗Not briefed |
hn hn-ai | 2mo | Show HN: Talleyrand – Open-source AI thinking workspace built as a question tree | I like to discuss complex topics with LLMs, but their linear chat is a bad place to think through hard questions. Every good answer ra… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Don't want OpenAI to steal your in-progress millennium prize proofs? Invest in homomorphic encryption today! | Don't want OpenAI to steal your in-progress millennium prize proofs? Invest in homomorphic encryption today! | ↗Not briefed |
bluesky bluesky-ai | 2mo | www.techtimes.co.uk/anthropic-se... | www.techtimes.co.uk/anthropic-se... | ↗Not briefed |
arxiv arxiv-ai | 2mo | Sparsity Regularized and Robust Mean Variance Portfolio Selection Under Ellipsoidal Uncertainty | We investigate mean-variance portfolio selection with an $\ell_0$-penalty to promote sparsity in asset allocations. Uncertainty in the… | ↗Not briefed |
hn hn-ai | 2mo | Show HN: Pascal's Pager – Webhooks In. Push Notifications Out | Sometimes it just takes too long to parse a webhook from one service to another just so you can get a push notification on your iPhone… | ↗Not briefed |
hn hn-ai | 2mo | Show HN: MySamantha – a productivity app where the AI is optional🔧 ChatGPT | Most of the recent tools being developed are versions of the personal assistant that everyone has been promising for a long time. I se… | ↗Not briefed |
bluesky bluesky-ai | 2mo | no fucking way dude, i was trying to flirt with my hypnosub gf by making her pretend to be an LLM and i got an actual LLM feature shoved in… | no fucking way dude, i was trying to flirt with my hypnosub gf by making her pretend to be an LLM and i got an actual LLM feature shov… | ↗Not briefed |
bluesky bluesky-ai | 2mo | But my point isn't to argue for trying to use LLMs in any contexts — but rather that we have to take care in the work we do. And if we try … | But my point isn't to argue for trying to use LLMs in any contexts — but rather that we have to take care in the work we do. And if we… | ↗Not briefed |
bluesky bluesky-ai | 2mo | ┈🌹━━━━━━━━🌹┈ ✨ [ #perfectgoddess] ✨ 💕 ✌️ 💕 ┈💖━━━━━━━━💖┈ #NSFW #Nude #AI #SoloGirl #OpenLegs #Pussy #CuteAndNaughty #HotGirl #Sexy #Go… | ┈🌹━━━━━━━━🌹┈ ✨ [ #perfectgoddess] ✨ 💕 ✌️ 💕 ┈💖━━━━━━━━💖┈ #NSFW #Nude #AI #SoloGirl #OpenLegs #Pussy #CuteAndNaughty #HotGirl #Sex… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Fox News host Bret Baier appeared momentarily stunned as a former researcher from two of the world’s most powerful artificial-intelligence … | Fox News host Bret Baier appeared momentarily stunned as a former researcher from two of the world’s most powerful artificial-intellig… | ↗Not briefed |
bluesky bluesky-ai | 2mo | AI techbro syndicate looking shady as fuck to nobody's surprise www.theverge.com/ai-artificia... | AI techbro syndicate looking shady as fuck to nobody's surprise www.theverge.com/ai-artificia... | ↗Not briefed |
arxiv arxiv-ai | 2mo | Building py-kvcache: A Performance Characterization of External KV Caching for vLLM with NVMe SSDs🔧 vLLM | Prefix caching can reduce the time to first token (TTFT) of long-context LLM requests by reusing previously computed key-value (KV) st… | ↗Not briefed |
rss openai-news | 2mo | How a researcher uses Codex and ChatGPT to search for new antimicrobial molecules🔧 ChatGPT | César de la Fuente’s lab uses Codex and ChatGPT to search living and extinct genomes for antimicrobial candidates to fight drug-resist… | ↗Not briefed |
bluesky bluesky-ai | 2mo | i'm home. logged out of bluesky on my phone because that shit isn't good for me but rest assured i had to review more LLM shit and it was v… | i'm home. logged out of bluesky on my phone because that shit isn't good for me but rest assured i had to review more LLM shit and it… | ↗Not briefed |
bluesky bluesky-ai | 2mo | … if someone does choose to use AI tools, I'd generally prefer they be transparent about it and publish both prompt and raw model output (c… | … if someone does choose to use AI tools, I'd generally prefer they be transparent about it and publish both prompt and raw model outp… | ↗Not briefed |
bluesky bluesky-ai | 2mo | 🚨 UPDATE: Newsom has just signed new AI audit and assessment laws. Still missing: the kill switch he vetoed in 2024, the exact kind of e… | 🚨 UPDATE: Newsom has just signed new AI audit and assessment laws. Still missing: the kill switch he vetoed in 2024, the exact kind… | ↗Not briefed |
hn hn-ai | 2mo | A Series of Unfortunate Jobs | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | En eks-Anthropic ansat siger, at AI med >10% sandsynlighed vil slå os alle ihjel indenfor de næste 10 år. Hvis man vil dykke lidt dybere i … | En eks-Anthropic ansat siger, at AI med >10% sandsynlighed vil slå os alle ihjel indenfor de næste 10 år. Hvis man vil dykke lidt dybe… | ↗Not briefed |
bluesky bluesky-ai | 2mo | GLM-5.3-Flash and now DeepSeek-v4.1-Flash are outperforming Opus-tier models from a few months ago at coding. Is this what acceleration lo… | GLM-5.3-Flash and now DeepSeek-v4.1-Flash are outperforming Opus-tier models from a few months ago at coding. Is this what accelerati… | ↗Not briefed |