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 | 20d | Integer Multiplication as a Living Research Programme [pdf] | — | ↗Not briefed |
bluesky bluesky-ai | 20d | LLM research on the Hitler vector is suggestive, depending on how much you think human brains are made of vectors. | LLM research on the Hitler vector is suggestive, depending on how much you think human brains are made of vectors. | ↗Not briefed |
bluesky bluesky-ai | 20d | Stop Sending Every AI Request to Your Most Expensive Model! hackernoon.com/stop-sending... | Stop Sending Every AI Request to Your Most Expensive Model! hackernoon.com/stop-sending... | ↗Not briefed |
bluesky bluesky-ai | 20d | Anthropic commits 5% of shares to new public benefit fund that could address AI job disruption | Anthropic commits 5% of shares to new public benefit fund that could address AI job disruption | ↗Not briefed |
hn hn-ai | 20d | Show HN: Gspot, a tool to lint gstack code | Hi HN! I just launched gspot. It's a tool that auto detects languages and frameworks you use in a repo and then sets up linters a… | ↗Not briefed |
bluesky bluesky-ai | 20d | “ 城壁の外から来るクラッカーに備えて防壁を高く積み上げながら、城門の内側では、自分の家に火を放てる権限を、反省のできない働き手に渡している。 ハッキングを恐れる前に、まず社内のエージェントから削除の権限を取り上げ、取り消せない操作のすべてに人間の承認を義務づけるべきだ。 … | “ 城壁の外から来るクラッカーに備えて防壁を高く積み上げながら、城門の内側では、自分の家に火を放てる権限を、反省のできない働き手に渡している。 ハッキングを恐れる前に、まず社内のエージェントから削除の権限を取り上げ、取り消せない操作のすべてに人間の承認を義務づけるべ… | ↗Not briefed |
hn hn-ai | 20d | What Anthropic's 17x quantum claim left out [video] | — | ↗Not briefed |
hn hn-ai | 20d | 2x Tesla p100s, q6_k quant, Qwen 3.8 27B ~60tps🔧 Qwen 3 | — | ↗Not briefed |
bluesky bluesky-ai | 20d | i've hit my personal limit. i think i've been undergoing sunk cost fallacy this entire time thinking there's any merit to LLM-assisted deve… | i've hit my personal limit. i think i've been undergoing sunk cost fallacy this entire time thinking there's any merit to LLM-assisted… | ↗Not briefed |
arxiv arxiv-ai | 20d | Finite Interaction Lens: Structured Precision Sharing for Vision-Language Models | Identifying which visual interventions interact in a vision-language model requires comparing their joint and separate effects. The re… | ↗Not briefed |
hn hn-ai | 20d | Eight-Dimension Audit for AI Governance: Anthropic, EU, China, UK | — | ↗Not briefed |
bluesky bluesky-ai | 20d | If openAI wanted to continue to pretend to care about "the future of humanity" (does any of you suckers still buy that?) they would stick w… | If openAI wanted to continue to pretend to care about "the future of humanity" (does any of you suckers still buy that?) they would st… | ↗Not briefed |
arxiv arxiv-ai | 20d | Benchmark for Structured Retrieval-Augmented Generation in Perovskite Photovoltaics | While retrieval-augmented language models increasingly serve as literature assistants in the physical sciences, existing benchmarks ev… | ↗Not briefed |
bluesky bluesky-ai | 20d | i obviously already hate AI, and one of the things i hate about it is that the name Claude is now related to that water-guzzling, content a… | i obviously already hate AI, and one of the things i hate about it is that the name Claude is now related to that water-guzzling, cont… | ↗Not briefed |
bluesky bluesky-ai | 20d | Worse, does the AI model just...put these things in there now? | Worse, does the AI model just...put these things in there now? | ↗Not briefed |
bluesky bluesky-ai | 20d | The best thing roko’s basilisk ever gave us was being able to mock the shit out of anyone who took it seriously And the gift is really gi… | The best thing roko’s basilisk ever gave us was being able to mock the shit out of anyone who took it seriously And the gift is real… | ↗Not briefed |
bluesky bluesky-ai | 20d | What OpenAI is doing isn't science, and it's barely engineering. The goal isn't knowledge (which comes from understanding), and it isn't e… | What OpenAI is doing isn't science, and it's barely engineering. The goal isn't knowledge (which comes from understanding), and it is… | ↗Not briefed |
hn hn-ai | 20d | Examples and information relating to using Web Assembly WAT | — | ↗Not briefed |
arxiv arxiv-ai | 20d | Representations in Motion: Layer-wise Rotation and Trajectory Divergence in LLMs | Recent work suggests that activation dynamics--how representations change during inference--might be more revealing about language mod… | ↗Not briefed |
hn hn-ai | 20d | Tkmd – A local CLI task manager and MCP server with AI agent review gates` | — | ↗Not briefed |
bluesky bluesky-ai | 20d | I'll see ads for this LLM shit and it's like "generate me this line of text on a plain background for my very important presentation slide!… | I'll see ads for this LLM shit and it's like "generate me this line of text on a plain background for my very important presentation s… | ↗Not briefed |
bluesky bluesky-ai | 20d | regresamos.substack.com/p/i-will-rev... predictably i watched the new street fighter. unpredictably this resulted in me finally "talking" … | regresamos.substack.com/p/i-will-rev... predictably i watched the new street fighter. unpredictably this resulted in me finally "talk… | ↗Not briefed |
bluesky bluesky-ai | 20d | lol how does it know | lol how does it know | ↗Not briefed |
rss techmeme | 20d | Sources: Nscale faces investor pushback on its ~$30B IPO valuation, with some urging a lower valuation of $16B to $19B, as Microsoft exits a UK data center deal (Bailey Lipschultz/Bloomberg) | Bailey Lipschultz / Bloomberg: Sources: Nscale faces investor pushback on its ~$30B IPO valuation, with some urging a lower valuation… | ↗Not briefed |
bluesky bluesky-ai | 20d | I was having a conversation last week with someone about the LLM dev "endgame" I personally do not believe we simply transition to a state… | I was having a conversation last week with someone about the LLM dev "endgame" I personally do not believe we simply transition to a… | ↗Not briefed |
arxiv arxiv-ai | 20d | Design and Cross-Domain Evaluation of an Agent Execution Harness | Completing complex tasks with language-model agents requires coordinating planning, execution, and review so the final result addresse… | ↗Not briefed |
hn hn-ai | 20d | Ask HN: Is Agentic Coding really faster? | I am trying to get a grasp of what we can do with agents. I am an avid user of agentic coding. But somehow I think I code faster by ha… | ↗Not briefed |
hn hn-ai | 20d | Show HN: CADR Mono – a programming font designed in Lisp | Just another monospaced vanity font. Among other things, this has a strong backquote to help with reading CL code. | ↗Not briefed |
arxiv arxiv-ai | 20d | Flattening Fantasies: Analyzing Stereotypes in Pornographic Discourse | This study presents a large-scale quantitative analysis of 5.4 million video titles from xHamster, the world's fourth most visited por… | ↗Not briefed |
arxiv arxiv-ai | 20d | Concurrent two-scale topology optimization of printable bone scaffolds via rectified-flow generation and exact latent-space refinement | Bone scaffolds must reconcile a macroscale mechanical response with microscale geometric feasibility, yet conventional approaches to t… | ↗Not briefed |
arxiv arxiv-ai | 20d | Learning Which Device to Call: Online Tool Grounding for LLM Agents in IoT | The Internet of Things (IoT) gives large language model (LLM) agents access to physical observations, but one requested capability is… | ↗Not briefed |
bluesky bluesky-ai | 20d | Quando apareceu a sugestão de usar genAI pra fazer o L, me pareceu ideia ruim, mais com cara de propaganda da openai do que de qualquer out… | Quando apareceu a sugestão de usar genAI pra fazer o L, me pareceu ideia ruim, mais com cara de propaganda da openai do que de qualque… | ↗Not briefed |
arxiv arxiv-ai | 20d | FTFM: Factorized Tabular Foundation Models | Tabular foundation models (TFMs) are becoming a powerful alternative to conventional tabular learning, adapting to new datasets in con… | ↗Not briefed |
hn hn-ai | 20d | ShortSmash – prompt-to-published short videos, 50-language dubbing, open source | — | ↗Not briefed |
arxiv arxiv-ai | 20d | Narrow Decision Models: A System-1 Architecture for Calibrated, Typed Decisions | System One models, introduced with TypeSafe's Jev, answer typed questions about a state with calibrated probabilities instead of gener… | ↗Not briefed |
arxiv arxiv-ai | 20d | Symbolic Constraint Agents: Mining Formal Tool-Call Constraints to Prevent Agent Errors🔧 GPT-5 | Symbolic constraints (programmatic, deterministic checks for tool calls) enable deterministic verification of LLM agents' tool calls,… | ↗Not briefed |
arxiv arxiv-ai | 20d | Editing as Language: Teaching LLM Agents a Grammar for Native Context Editing | A synopsis in novels folds a story into a few lines; a function in code gathers many steps with a single execution. Context editing is… | ↗Not briefed |
hn hn-ai | 20d | Boeing's 737 MAX 7 gets certification from FAA | — | ↗Not briefed |
bluesky bluesky-ai | 20d | Minor update to #DeepSeek v4.1 Flash 1017 #Model Improvements on benchmarks. 1M context window. #AI #LLM #OpenSource MIT lic 🤗 huggingface… | Minor update to #DeepSeek v4.1 Flash 1017 #Model Improvements on benchmarks. 1M context window. #AI #LLM #OpenSource MIT lic 🤗 huggin… | ↗Not briefed |
hn hn-ai | 20d | Zlodev – Local HTTPS reverse proxy with auto-certificates, custom DNS, and a TUI | — | ↗Not briefed |
bluesky bluesky-ai | 20d | Les ravages de la drogue: Anthropic prévoit une croissance du PIB américain de près de 50% pour 2030. Pour comparaison, la FED prévoit pou… | Les ravages de la drogue: Anthropic prévoit une croissance du PIB américain de près de 50% pour 2030. Pour comparaison, la FED prévoi… | ↗Not briefed |
arxiv arxiv-ai | 20d | EmotiQ: A Quantum Residual Specialist Framework for Text Emotion Classification | Emotion classification from text is challenging because many affective states are distinguished by subtle linguistic signals and seman… | ↗Not briefed |
bluesky bluesky-ai | 20d | Here's a great Layman's explanation for why what OpenAI did was so fucking dangerous "Unsolvable problems" are like chew toys for mathemat… | Here's a great Layman's explanation for why what OpenAI did was so fucking dangerous "Unsolvable problems" are like chew toys for mat… | ↗Not briefed |
bluesky bluesky-ai | 20d | before I grant that wish, I have a few clarifying questions for you, so we can be sure you're getting what you want | before I grant that wish, I have a few clarifying questions for you, so we can be sure you're getting what you want | ↗Not briefed |
hn hn-ai | 20d | Show HN: WhatsApp MCP running free on Cloudflare🔧 whatsapp MCP | I wanted my humble Claude on mobile to access my WhatsApp messages but the only options I could find either required WhatsApp Business… | ↗Not briefed |
arxiv arxiv-ai | 20d | Beyond the record: evaluating probabilistic AI weather forecasts deep in the tail | Can probabilistic artificial intelligence (AI) weather forecasts retain skill beyond their training domain? We compare four state-of-t… | ↗Not briefed |
bluesky bluesky-ai | 20d | As Nova Scotia's Premier forces a MAGA owned data center on us, to serve Anthropic. Given that Prime Minister Carney said we are at war, w… | As Nova Scotia's Premier forces a MAGA owned data center on us, to serve Anthropic. Given that Prime Minister Carney said we are at w… | ↗Not briefed |
arxiv arxiv-ai | 20d | Reinforcement learning on the EPRI AI-ccelerating Unit Commitment benchmark: diagnosing over-commitment and guiding MILP with learned policies | Reinforcement learning (RL) is increasingly proposed as a fast solver for unit commitment (UC), but standalone agents remain far more… | ↗Not briefed |
bluesky bluesky-ai | 20d | Zenzap raised $44M, led by Bessemer and Lightspeed, to replace WhatsApp and iMessage for frontline and shift workers. Customers include Fed… | Zenzap raised $44M, led by Bessemer and Lightspeed, to replace WhatsApp and iMessage for frontline and shift workers. Customers includ… | ↗Not briefed |
hn hn-ai | 20d | ParetoOps – CI gate on cost per successful task for LLM evals | — | ↗Not briefed |