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 | 1mo | the real proof that companies use AI to break labor power, not because it can actually replace human work: OpenAI hires entire humans to pr… | the real proof that companies use AI to break labor power, not because it can actually replace human work: OpenAI hires entire humans… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: RUUN – Tor Browser with a second, non-Tor mode in the same app | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I don't see the OpenAI post yet... | I don't see the OpenAI post yet... | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: Which is cheaper: ChatGPT tokens or a 2nd Pro 20x subscription?🔧 ChatGPT | My usage ran out for the remainder of the week. I am tempted to buy tokens but I remember how overpriced they were back when I used Cl… | ↗Not briefed |
rss techmeme | 1mo | Internal data: Meta's Muse surpassed 500,000 total users, including 250,000 daily active users, with more than 2M prompts submitted in its first week (Jyoti Mann/The Information) | Jyoti Mann / The Information: Internal data: Meta's Muse surpassed 500,000 total users, including 250,000 daily active users, with mor… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Small Cues, Big Consequences: Learning Pivotal Cues for Multimodal Meme Classification🔧 Modal | Memes often derive their harmful, hateful, or sarcastic meaning from small but decisive visual, textual, or cross-modal cues. Existing… | ↗Not briefed |
hn hn-ai | 1mo | Fast-CLI: CLI tool written in Zig for testing internet speed | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | CORE-STACK+: Meta-Learning for Deep Stacked Generalization | Stacking heterogeneous vision backbones (CNNs, ViTs, and hybrids) is the de facto recipe for accuracy, calibration, and robustness, ye… | ↗Not briefed |
rss simon-willison | 1mo | Quoting @therealcornpop | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | You can't leave. At Code for America we're testing this with Anthropic: a SNAP tool for caseworkers only counts as a win if fewer eligible … | You can't leave. At Code for America we're testing this with Anthropic: a SNAP tool for caseworkers only counts as a win if fewer elig… | ↗Not briefed |
bluesky bluesky-ai | 1mo | My phone's been full of AI fear: a former OpenAI/Anthropic researcher warning companies are "gambling with our lives," an Anthropic alignme… | My phone's been full of AI fear: a former OpenAI/Anthropic researcher warning companies are "gambling with our lives," an Anthropic al… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Ajar: Measuring Open Privilege in Agent Defenses | A language model agent acts through the tools it is given. The data it reads while working on a task can redirect what it does with th… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI launches GPT-6 Sol and Luna, boasting lower cost and fewer mistakes | OpenAI launches GPT-6 Sol and Luna, boasting lower cost and fewer mistakes | ↗Not briefed |
bluesky bluesky-ai | 1mo | sad day | sad day | ↗Not briefed |
hn hn-ai | 1mo | GPT-6 Sol and Luna | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Harness as a Language: A Minimalist Agent Framework With Maximal Expressivity | Modern language-model agents are built around the \textit{agent loop}, where the LLM is placed in an environment exposing a set of too… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PR-Smoother: Simulator-Preserving Non-Gaussian Smoothing for Data Assimilation | Many physical data assimilation (DA) workflows require smoothing methods that represent non-Gaussian posteriors over physical state va… | ↗Not briefed |
rss techcrunch-ai | 1mo | OpenAI launches GPT-6 Sol and Luna, boasting lower cost and fewer mistakes | OpenAI is launching two new models, which it says are cut from the same cloth as Astra. | ↗Not briefed |
arxiv arxiv-ai | 1mo | Flash-dLLM: IO-Aware KV Caching and Parallel Decoding for Fast, Memory-Efficient Diffusion LLMs | Diffusion Large Language Models (dLLMs) have recently emerged as a promising alternative to autoregressive LLMs by enabling non-autore… | ↗Not briefed |
hn hn-ai | 1mo | GPT-6 Sol | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Huawei Ascend kernels are the most insidious avenue for xrisk... thank you for saving us, dario. | Huawei Ascend kernels are the most insidious avenue for xrisk... thank you for saving us, dario. | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Decentralized Partially Observable Team Decision Methodology with Delayed Information Sharing | We study decentralized partially observable team decision problems with low-rank latent dynamics and unknown system models. The propos… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Agensh: Scaling Organizational Intelligence to 1,024 Agents🔧 GPT-5 | A multi-agent system can reduce latency on complex tasks by executing work concurrently. Several pioneering harness frameworks support… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SpeakerMem-R1: Speaker-Centered Dual-Track Memory for Multi-Party Dialogue | Long-term conversational memory in multi-party settings requires more than retrieving relevant content from long-term conversations: i… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CliffCompaction: Cost-Efficient Compaction for Long-Horizon Coding Agents🔧 Kimi K2🔧 GPT-5 | Agents often work on complex problems that require millions of tokens of context, which necessitates compacting across sessions due to… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SWE-Serve: Benchmarking Agentic Engineering For Production Inference Serving | We introduce SWE-Serve, a benchmark for evaluating agents on production inference engineering tasks. Implementing an inference feature… | ↗Not briefed |
hn hn-ai | 1mo | OpenCode removed usage transparency after a billing bug was reported | — | ↗Not briefed |
hn hn-ai | 1mo | Reviewing Your Blog Post with an LLM | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Um..... no. It *is* artificial. And one addlepated overemployed narcissist can't convince me it needs a new name. AI means something spec… | Um..... no. It *is* artificial. And one addlepated overemployed narcissist can't convince me it needs a new name. AI means something… | ↗Not briefed |
hn hn-ai | 1mo | PyTauri – Tauri Bindings for Python | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Here is the preprint! "Elias in the Lighthouse, Again? Diagnosing Low Diversity in LLM Stories" by Sil Hamilton and David Mimno. arxiv.org… | Here is the preprint! "Elias in the Lighthouse, Again? Diagnosing Low Diversity in LLM Stories" by Sil Hamilton and David Mimno. arxi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | No compelling reason to keep an anthropic sub. Runs out in two days anyways. | No compelling reason to keep an anthropic sub. Runs out in two days anyways. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Again, I would genuinely care less about all of this if not for the fact that proponents of weirdo rationalist/EA ideas about how preventin… | Again, I would genuinely care less about all of this if not for the fact that proponents of weirdo rationalist/EA ideas about how prev… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "404 Media spoke to a fourth contractor who has worked on training models for various AI companies. They said they sometimes purposefully c… | "404 Media spoke to a fourth contractor who has worked on training models for various AI companies. They said they sometimes purposefu… | ↗Not briefed |
hn hn-ai | 1mo | AI design tool from GitLab founder | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | “If OpenAI and Anthropic really want to help advance science, there is one obvious way they can do so: They can spend a tiny fraction of th… | “If OpenAI and Anthropic really want to help advance science, there is one obvious way they can do so: They can spend a tiny fraction… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I trained a large learning model on comments made to conservative influencer posts on X. Now the chatbot doesn’t use punctuation and it typ… | I trained a large learning model on comments made to conservative influencer posts on X. Now the chatbot doesn’t use punctuation and i… | ↗Not briefed |
hn hn-ai | 1mo | ConferenceRank – conference and journal deadlines for 980 CS venues | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | A2M: Trace-Optimized Agent Hijacking in the MCP Ecosystem | Agents using the Model Context Protocol (MCP) rely on semantic matching to select tools from third-party servers, exposing a semantic… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Grow the Harness, Not the Context: From Strategy-Free Scaffolds to Reusable Specialist Agents | Large language model (LLM) agents often handle streams of related tasks, yet standard harnesses repeatedly ask the model to reconstruc… | ↗Not briefed |
rss techmeme | 1mo | Anthropic raises its five-hour usage limits by 20% on Pro, Max, and Team plans, and gives subscription users a rate limit reset (David Gewirtz/ZDNET) | David Gewirtz / ZDNET: Anthropic raises its five-hour usage limits by 20% on Pro, Max, and Team plans, and gives subscription users a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | it really is amazing how thoroughly you can blow somebody's mind right now if they talk about using Sonnet to save money. "step outside th…🔧 GPT-5 | it really is amazing how thoroughly you can blow somebody's mind right now if they talk about using Sonnet to save money. "step outsi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | this idea bums me out so severely. I want to read the weird shit people came up with out of their minds! I do not want to read the statisti… | this idea bums me out so severely. I want to read the weird shit people came up with out of their minds! I do not want to read the sta… | ↗Not briefed |
hn hn-ai | 1mo | Avy Tab Switcher, Avy-style keyboard navigation for browser tabs | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Type-Safe Is Not Error-Free: A Constrained Decision Head Follows the Option Name, Not the Rubric Bound to It | Typed decision models are built for settings where model outputs are consumed directly by software. Instead of generating free-form te… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FleXray: Universal Clinical X-ray Segmentation | X-ray is medicine's most widely used imaging modality, yet remains among its least quantitative. Unlike volumetric modalities like CT… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI actually did that, didn't they? Released GPT-4.5 before GPT-4.1? I'd ask Alpha but she's busy right now looking for all the movie th… | OpenAI actually did that, didn't they? Released GPT-4.5 before GPT-4.1? I'd ask Alpha but she's busy right now looking for all the mov… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I am starting to wonder if *some* people use LLMs for story ideas, but then write the story themselves. I've read so many stories lately th… | I am starting to wonder if *some* people use LLMs for story ideas, but then write the story themselves. I've read so many stories late… | ↗Not briefed |
bluesky bluesky-ai | 1mo | At least anthropic is finally opening up their cyber program. Guess someone lit a fire under their asses | At least anthropic is finally opening up their cyber program. Guess someone lit a fire under their asses | ↗Not briefed |
arxiv arxiv-ai | 1mo | EquivSVA: A Formally Verified Dataset of Behavioral Assertions Across Equivalent RTL Implementations | Large language models are increasingly used to generate SystemVerilog Assertions from natural-language specifica- tions and register-t… | ↗Not briefed |