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 | 1mo | Gemini 3.8 text-to-speech says hello | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anisota / dame(dot)is / bsky38 / and more Using LLM's and coding and moderation on their websites and promotes their gen-ai work for hire. … | Anisota / dame(dot)is / bsky38 / and more Using LLM's and coding and moderation on their websites and promotes their gen-ai work for h… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Shutdown Sabotage Propensities in Multi-Agent Systems | The final safeguard against rogue AI behavior is the human ability to shut systems down. It has been theorized that when an AI is inst… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Non-Commutative State Tracking with Input-Dependent Low-Rank Updates in Mamba-3 | State tracking from sequential observations can require both retaining information and updating it by composing observed operations. W… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Towards Efficient Reasoning: Learning Causal Shortcuts for Diffusion Language Models | Diffusion Language Models (DLMs) have attracted significant attention for their strong reasoning ability. However, under a bidirection… | ↗Not briefed |
hn hn-ai | 1mo | Valve's Linux Steam client now runs on an Android phone. No root | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | we'll never know how stable and functional this site could be if the devs weren't more interested in addressing product issues and bugs ins… | we'll never know how stable and functional this site could be if the devs weren't more interested in addressing product issues and bug… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Predicting Quantization Price for Selecting PTQ Configurations Before Deployment | Weight-space post-training quantization (PTQ) must choose finite formats, granularities, quantizer families, transformations, and bits… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Open Code for Jev | Hi everyone, I built this TUI for jev with the main goal to make Jev immediately accessible and to be able to have decision templates… | ↗Not briefed |
rss deepmind-blog | 1mo | Gemini 3.8 text-to-speech says hello | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Takže stále ještě vede GPT nad Meta ;) Díky za odpovědi. Nicméně Meta AI hodně pokročila, zvlášť v posledním týdnu, kdy přišel model Muse. … | Takže stále ještě vede GPT nad Meta ;) Díky za odpovědi. Nicméně Meta AI hodně pokročila, zvlášť v posledním týdnu, kdy přišel model M… | ↗Not briefed |
bluesky bluesky-ai | 1mo | quietly letting a business deeply dependent on an llm for virtually everything a human did last year know that their model of choice confid… | quietly letting a business deeply dependent on an llm for virtually everything a human did last year know that their model of choice c… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Resource-Adaptive Stochastic Gradient Descent for Online Linear Programming without Re-solving | The growth of large language model (LLM) inference and search services increases the scale of online linear programming problems, moti… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RAMP: Robust Adaptive Mixed-Precision Quantization for Edge CPU Vision Models | Deploying deep learning models on edge CPUs is bottlenecked by computational and memory constraints. Mixed-precision quantization prom… | ↗Not briefed |
hn hn-ai | 1mo | Amazon opens its seller tools to outside AI agents, starting with Anthropic | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Generalizable Robotic Insertion with World Models | Robotic assembly in high-mixture settings requires adaptable systems that can handle diverse parts, yet current approaches typically r… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic refused to let its models be used for mass surveillance or autonomous weapons. The Pentagon called the company a national securit… | Anthropic refused to let its models be used for mass surveillance or autonomous weapons. The Pentagon called the company a national se… | ↗Not briefed |
rss techmeme | 1mo | Q&A with Jensen Huang on AI creating more jobs than it destroys, pushing back against AI doomerism, Chinese open models, the Hugging Face acquisition, and more (Ezra Klein/New York Times)🔧 Hugging Face | Ezra Klein / New York Times: Q&A with Jensen Huang on AI creating more jobs than it destroys, pushing back against AI doomerism, Chine… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Attn – An open-source local Gmail client | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | "You can choose voices from leading [thieving audio slopfarm] models!" The theft here is triple: the "models" were made by stealing voice … | "You can choose voices from leading [thieving audio slopfarm] models!" The theft here is triple: the "models" were made by stealing v… | ↗Not briefed |
bluesky bluesky-ai | 1mo | They get it! > How do you enforce this? > You [can't]. People will still send LLM generated code. This policy makes it explicit that we d… | They get it! > How do you enforce this? > You [can't]. People will still send LLM generated code. This policy makes it explicit that… | ↗Not briefed |
bluesky bluesky-ai | 1mo | yeah, sure, let's throw buckets of money at 20-somethings claiming to build LLM-powered mind-reading technology, why not | yeah, sure, let's throw buckets of money at 20-somethings claiming to build LLM-powered mind-reading technology, why not | ↗Not briefed |
rss techcrunch-ai | 1mo | YouTube Music gets more conversational with new AI features | YouTube is adding new AI-powered features to YouTube Music that aim to make it easier for listeners to discover music, explore artists… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MemBodied: Recurrent Associative Memory for Vision-Language-Action Models | Vision-Language-Action models provide a strong foundation for general-purpose robot control, yet a vast majority of policies do not pr… | ↗Not briefed |
hn hn-ai | 1mo | Evaluator Optimizer | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | An explainer answering quite *how* AI is supposed to be able to kill us all, why Anthropic is pushing for regulation, what to really worry … | An explainer answering quite *how* AI is supposed to be able to kill us all, why Anthropic is pushing for regulation, what to really w… | ↗Not briefed |
hn hn-ai | 1mo | Why are Opus 5.5 and GPT-6 Sol cheaper? | — | ↗Not briefed |
rss the-decoder | 1mo | Anthropic engineer explains why Claude's writing got worse although the model got smarter | Anthropic employee Jackson Kernion explains why newer Claude models write so oddly. Optimizing for math, code, and technical explanati… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Managing in the age of AI "will only work if you talk to humans more than you talk to Claude" @skamille.themanagerswrath.com skamille.mediu… | Managing in the age of AI "will only work if you talk to humans more than you talk to Claude" @skamille.themanagerswrath.com skamille.… | ↗Not briefed |
hn hn-ai | 1mo | Jev does not play dice | — | ↗Not briefed |
hn hn-ai | 1mo | Parakeet.java: Automatic Speech Recognition in Pure Java | — | ↗Not briefed |
hn hn-ai | 1mo | Step Code: MIT-licensed coding agent CLI built around Step 5 Preview | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Complementary Roles of Activation and Parametric Memory in Few-Shot Learning | At test time, large language models (LLMs) can encode historical information in activation memory (i.e., KV caches) and parametric mem… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: I Built a GIF Decider with Jev | Hey everyone!<p>I wanted to try out a fun use case for Jev and had this idea before where I could use AI to decide what GIF is best to… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: A San Francisco cable car ride as a procedural risograph film | I saw a risograph render with Claude on Reddit and wanted to try it on the Powell/Hyde cable car. It's a 97s film covering U… | ↗Not briefed |
arxiv arxiv-ai | 1mo | hyperbolix: Hyperbolic Deep Learning in JAX | We present hyperbolix, an open-source library for hyperbolic deep learning in JAX, built on Flax NNX. To our knowledge, it is the firs… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Controlling Collectives of AI Agents in Reasoning Space with Spatial Transformers | Large Language Models (LLMs) introduce an exciting new paradigm for planning and navigation in robotics, but fail on even simple multi… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond Poetry: Can Large Language Models Generate Classical Arabic Maqamat?🔧 GPT-5 | Large language models (LLMs) have shown strong performance in creative text generation, yet their ability to produce culturally ground… | ↗Not briefed |
bluesky bluesky-ai | 1mo | SoftBank leiht sich wohl weitere 11 Milliarden Dollar von Investoren, um seine dritte Teilzahlung an OpenAI zu finanzieren. Die Anleihen ge… | SoftBank leiht sich wohl weitere 11 Milliarden Dollar von Investoren, um seine dritte Teilzahlung an OpenAI zu finanzieren. Die Anleih… | ↗Not briefed |
rss techmeme | 1mo | At its Accelerate conference, Amazon says it is opening its seller tools to third-party AI agents, starting with Claude, in beta for US merchants (Todd Bishop/GeekWire) | Todd Bishop / GeekWire: At its Accelerate conference, Amazon says it is opening its seller tools to third-party AI agents, starting wi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Patreon's co-founder, head of product and head of engineering are all leaving to join OpenAI to work on a new "creator product". This read… | Patreon's co-founder, head of product and head of engineering are all leaving to join OpenAI to work on a new "creator product". This… | ↗Not briefed |
hn hn-ai | 1mo | Anonymous Git Pull Requests | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yeah, see, the thing is, more experts have since weighed in on the 100+ *additional* unsolved math problems that Astra has subsequently sol… | Yeah, see, the thing is, more experts have since weighed in on the 100+ *additional* unsolved math problems that Astra has subsequentl… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Do Center Biases Propagate? Robustness of Pathology Foundation Models in Whole-Slide Image Classification | Pathology foundation models (PFMs) have transformed computational pathology through powerful representation learning from histopatholo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | agreed, but that's more about them sucking at sandboxing than them just giving the model internet access like anthropic accidentally did, a… | agreed, but that's more about them sucking at sandboxing than them just giving the model internet access like anthropic accidentally d… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Build a CPU from gates, in TypeScript, in the browser | — | ↗Not briefed |
hn hn-ai | 1mo | Riso-windowseat – Procedural risograph films and prints | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | From Agent Output to Authorized Transition | Agentic engineering systems can edit repositories, run tools and tests, build firmware, synthesize schematics, and prepare deployable… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI's Next Marketing Stunt [video] | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Log-Depth Recurrent Language Modeling | Language modeling using Transformers has become commonplace despite their fixed computational depth and quadratic runtime with respect… | ↗Not briefed |