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 | 10d | LLMへの反発を「ラッダイト」という言葉でバカにするのはよく目にするね。しかし、産業革命初期にラッダイト運動に参加した人たちの多くは、悲惨な光景への義憤が動機な人が多いと分かっていて。 「食事がついている」と宣伝して孤児を炭鉱作業に雇い、契約書にサインをしたら檻に閉じ込めてしま… | LLMへの反発を「ラッダイト」という言葉でバカにするのはよく目にするね。しかし、産業革命初期にラッダイト運動に参加した人たちの多くは、悲惨な光景への義憤が動機な人が多いと分かっていて。 「食事がついている」と宣伝して孤児を炭鉱作業に雇い、契約書にサインをしたら檻に閉じ… | ↗Not briefed |
arxiv arxiv-ai | 10d | Where Language-Model Vocabulary Rose on Stack Exchange🔧 ChatGPT | Stack Overflow banned answers generated by ChatGPT five days after the model's release, and many other Stack Exchange sites announced… | ↗Not briefed |
arxiv arxiv-ai | 10d | MAST: Medoid-Anchored Subspace Steering for Visual Grounding in Multimodal LLMs | Existing activation steering methods for multimodal large language models (MLLMs) use the global activation mean as the steering vecto… | ↗Not briefed |
hn hn-ai | 10d | How to export Claude chat to word without losing formatting [video] | — | ↗Not briefed |
hn hn-ai | 10d | Realtime X-ray camera/mirror effect | — | ↗Not briefed |
bluesky bluesky-ai | 10d | Disclosure: I don’t use AI for gamedev, writing, or art generation. But I have started using Claude to control DaVinci Resolve (Lua console… | Disclosure: I don’t use AI for gamedev, writing, or art generation. But I have started using Claude to control DaVinci Resolve (Lua co… | ↗Not briefed |
bluesky bluesky-ai | 10d | combined with this charming anecdote I can't imagine it being a problem! www.reddit.com/r/antiai/s/p... | combined with this charming anecdote I can't imagine it being a problem! www.reddit.com/r/antiai/s/p... | ↗Not briefed |
bluesky bluesky-ai | 10d | I have to admit that it sometimes leads to odd places, being very pro-robits and pro-transhumanism while also taking a fairly maximalist an… | I have to admit that it sometimes leads to odd places, being very pro-robits and pro-transhumanism while also taking a fairly maximali… | ↗Not briefed |
bluesky bluesky-ai | 10d | Anthropic is going public at a fraught time in tech markets and the AI industry. Here's what to know. | Anthropic is going public at a fraught time in tech markets and the AI industry. Here's what to know. | ↗Not briefed |
bluesky bluesky-ai | 10d | There’s a really annoying ad for an AI-enabled phone that I keep seeing. A young woman is overwhelmed by the reading she’s doing for a jo… | There’s a really annoying ad for an AI-enabled phone that I keep seeing. A young woman is overwhelmed by the reading she’s doing for… | ↗Not briefed |
arxiv arxiv-ai | 10d | Stealthy, High-Impact False Data Injection Attacks in Smart Grid Using Physics-Constrained GAN | False data injection (FDI) attacks can distort power system state estimation while retaining plausible measurement patterns. Existing… | ↗Not briefed |
bluesky bluesky-ai | 10d | Yellow heart beats, green vines spiral into the blue abyss #DigitAlart #AIArt #AIArtwork #AIGenerated #ZImage #OpenAI | Yellow heart beats, green vines spiral into the blue abyss #DigitAlart #AIArt #AIArtwork #AIGenerated #ZImage #OpenAI | ↗Not briefed |
hn hn-ai | 10d | Graphical way to manage Docker, Docker Compose and system services | — | ↗Not briefed |
bluesky bluesky-ai | 10d | Been too burnt out to follow much of the latest discourse, but suffice to say, we just gotta treat the LLM nonsense like smoking. For peop… | Been too burnt out to follow much of the latest discourse, but suffice to say, we just gotta treat the LLM nonsense like smoking. For… | ↗Not briefed |
bluesky bluesky-ai | 10d | Our group's only connection with OpenAI (BESIDES ACCEPTING THEIR SUBMISSIONS) is the use of THEIR ASTRA MODEL TO HELP WITH ANALYSIS. At le… | Our group's only connection with OpenAI (BESIDES ACCEPTING THEIR SUBMISSIONS) is the use of THEIR ASTRA MODEL TO HELP WITH ANALYSIS.… | ↗Not briefed |
bluesky bluesky-ai | 10d | "Exciting" news! Remember when OpenAI drained a lake to create a 1000+ page unreadable slop proof of YM existence, a result nobody ever dou… | "Exciting" news! Remember when OpenAI drained a lake to create a 1000+ page unreadable slop proof of YM existence, a result nobody eve… | ↗Not briefed |
bluesky bluesky-ai | 10d | Faith in financial markets is as close to a definition of insanity as we will ever get www.taxresearch.org.uk/Blog/2026/11... | Faith in financial markets is as close to a definition of insanity as we will ever get www.taxresearch.org.uk/Blog/2026/11... | ↗Not briefed |
rss techmeme | 10d | VR pioneer Jaron Lanier says Silicon Valley is sabotaging itself by poorly pitching AI and treating areas designated for data centers as sacrifice zones (Jaron Lanier/New York Times) | Jaron Lanier / New York Times: VR pioneer Jaron Lanier says Silicon Valley is sabotaging itself by poorly pitching AI and treating are… | ↗Not briefed |
arxiv arxiv-ai | 10d | PRELUDE: Separating Common Components from Target-Specific Dependencies for Multivariate Time Series Forecasting | Multivariate time series forecasting often improves the prediction of a target by modeling the interactions among variables, and cross… | ↗Not briefed |
bluesky bluesky-ai | 10d | זה בדיוק מה שה-llm שלהם ניסה לבדוק | זה בדיוק מה שה-llm שלהם ניסה לבדוק | ↗Not briefed |
bluesky bluesky-ai | 10d | Solution already exists actually, as LLMs "watermark" their written work nowadays (or are starting to) because they were tired of training … | Solution already exists actually, as LLMs "watermark" their written work nowadays (or are starting to) because they were tired of trai… | ↗Not briefed |
bluesky bluesky-ai | 10d | - 2025+ - хей LLM, направи ми план за A-B-C с основните неща, аз ще ти се заям за 2-3 от тях, ще построим план, ти ще го направиш, аз само … | - 2025+ - хей LLM, направи ми план за A-B-C с основните неща, аз ще ти се заям за 2-3 от тях, ще построим план, ти ще го направиш, аз… | ↗Not briefed |
arxiv arxiv-ai | 10d | Knowledge Transfer for Conditional Flow Matching: Frozen Estimation and Adaptation | Conditional generative modeling with limited target data can benefit from representations learned in a related source population. We d… | ↗Not briefed |
bluesky bluesky-ai | 10d | ALL Ai use is HEAVILY subsidized by venture capital. Not a single one of these LLM companies is charging the true price for these questiona… | ALL Ai use is HEAVILY subsidized by venture capital. Not a single one of these LLM companies is charging the true price for these ques… | ↗Not briefed |
hn hn-ai | 10d | Panorea, a chiptune note player working on Pascal and aplay | — | ↗Not briefed |
hn hn-ai | 10d | Convert all Google photos to AVIF | — | ↗Not briefed |
bluesky bluesky-ai | 10d | Maybe I’m uninformed but is a local, open source LLM model actually ethical when you consider how it was trained. Are they not trained on s… | Maybe I’m uninformed but is a local, open source LLM model actually ethical when you consider how it was trained. Are they not trained… | ↗Not briefed |
bluesky bluesky-ai | 10d | The externalities should be paid by the LLM companies and the subsidization of the tech and the bubble should end. I just hate the big tec… | The externalities should be paid by the LLM companies and the subsidization of the tech and the bubble should end. I just hate the bi… | ↗Not briefed |
rss techmeme | 10d | Nintendo sold 5.2M copies of its The Legend of Zelda: Ocarina of Time remake in just four days, giving the Switch 2 the blockbuster hit it had been waiting for (Vlad Savov/Bloomberg) | Vlad Savov / Bloomberg: Nintendo sold 5.2M copies of its The Legend of Zelda: Ocarina of Time remake in just four days, giving the Swi… | ↗Not briefed |
hn hn-ai | 10d | Compress any LLM 3x in one click, near-lossless, in under 50 MB RAM | — | ↗Not briefed |
rss marktechpost | 10d | MBZUAI’s IFM Releases PAN 2, an Open 29B World Model for Physical AI | The Institute of Foundation Models at MBZUAI released PAN 2, an Apache 2.0 world model that couples an LLM planner with a 29B diffusio… | ↗Not briefed |
arxiv arxiv-ai | 10d | Reframing Video Inverse Tone Mapping as Blind Posterior Inference via Latent Luminance Supervision | High-dynamic-range (HDR) displays are now common, but most video content, such as archival footage, broadcast material, and the output… | ↗Not briefed |
bluesky bluesky-ai | 10d | I only realized midway through Joel's new signal simulator video that the gronk bit was an actual LLM I'm about to fucking devour handfuls … | I only realized midway through Joel's new signal simulator video that the gronk bit was an actual LLM I'm about to fucking devour hand… | ↗Not briefed |
bluesky bluesky-ai | 10d | Wow I've woken up to some HEAVY LLM cuck copium... Holy shit. | Wow I've woken up to some HEAVY LLM cuck copium... Holy shit. | ↗Not briefed |
hn hn-ai | 10d | Show HN: Overdeck 0.74 – MCP tools for coding agents, plus phone apps | — | ↗Not briefed |
arxiv arxiv-ai | 10d | Multi-View 3D Reconstruction Using V-JEPA Features | Predictive video encoders provide dense visual representations, but adapting them to 3D reconstruction requires accurate depth, camera… | ↗Not briefed |
hn hn-ai | 10d | Stabl: Societal transformation and AI bottlenecks lab | — | ↗Not briefed |
arxiv arxiv-ai | 10d | When Does Learned Constraint Selection Pay in Security-Constrained Optimisation? A Boundary Map from SCOPF to Unit Commitment | Learning-to-optimise methods for security-constrained optimal power flow (SCOPF) predict the security constraints that bind on each in… | ↗Not briefed |
rss techcrunch-ai | 10d | Nvidia’s crisis has corporate governance experts reaching for the popcorn | The governance experts we talked to say nobody has seen anything like what's unfolding right now. | ↗Not briefed |
bluesky bluesky-ai | 10d | I saw the LLM overview on google when I was searching for a solution at work and now my ethical No-AI pledge producer is coming after me wi… | I saw the LLM overview on google when I was searching for a solution at work and now my ethical No-AI pledge producer is coming after… | ↗Not briefed |
bluesky bluesky-ai | 10d | Eben. LLM zur Recherche ist ggf mehr Aufwand: mehr gefundene Quellen, mehr neue Querverweise & die Reports müssen erst gesichtet und überpr… | Eben. LLM zur Recherche ist ggf mehr Aufwand: mehr gefundene Quellen, mehr neue Querverweise & die Reports müssen erst gesichtet und ü… | ↗Not briefed |
arxiv arxiv-ai | 10d | Serving-Aligned Entire-Space Conversion Modeling via Click-Through Rate Knowledge Distillation | Recommendation systems may use predicted click-through conversion rate (pCTCVR), the product of predicted click-through rate (pCTR) an… | ↗Not briefed |
bluesky bluesky-ai | 10d | Central banks and investment banks don’t count “pledges”. They count assets and liabilities. The IPCC’s pathetic method of counting “pledg… | Central banks and investment banks don’t count “pledges”. They count assets and liabilities. The IPCC’s pathetic method of counting “… | ↗Not briefed |
hn hn-ai | 10d | We're still so small but I'm loving the team we've got at Superlogical | — | ↗Not briefed |
bluesky bluesky-ai | 10d | No, House House used an LLM to diagnose very few specific code issues when Unity told them they couldn’t help them. Please don’t share clic… | No, House House used an LLM to diagnose very few specific code issues when Unity told them they couldn’t help them. Please don’t share… | ↗Not briefed |
bluesky bluesky-ai | 10d | "An analysis from benchmarking company Vals AI found that using AI to build a software application can have an environmental impact 10,000 … | "An analysis from benchmarking company Vals AI found that using AI to build a software application can have an environmental impact 10… | ↗Not briefed |
bluesky bluesky-ai | 10d | 3/ Run your own DAU, prompt cache hit rate (0-95%), and LiteLLM fallback router simulation on our live comparison page: https://tokenmargin… | 3/ Run your own DAU, prompt cache hit rate (0-95%), and LiteLLM fallback router simulation on our live comparison page: https://tokenm… | ↗Not briefed |
bluesky bluesky-ai | 10d | Llama 4 Maverick (17B-128E MoE) vs DeepSeek V3 (0324): Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here is the l…🔧 Llama 4 | Llama 4 Maverick (17B-128E MoE) vs DeepSeek V3 (0324): Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here is… | ↗Not briefed |
arxiv arxiv-ai | 10d | Read Once, Answer Many: Parallel Prompting for Shared-Context LLM Inference | LLM workloads sometimes ask many independent questions about the same long document, as in document QA, data labeling and LLM-based ev… | ↗Not briefed |
arxiv arxiv-ai | 10d | Recompute or Restore? Host-Tier KV for Multi-Model Agent Workflows on a Static-Graph NPU Serving Stack🔧 vLLM | Agent workflows issue dependent model calls that share long context prefixes, so serving them efficiently depends on reusing the key-v… | ↗Not briefed |