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 | 11d | Magpie: Open-source model gateway for coding agents (Claude Code, Codex) | — | ↗Not briefed |
bluesky bluesky-ai | 11d | Another weird AI game project: I asked Claude Fable 5.2 for a fantasy transport sim with Discworld meets Lord of the Rings vibes. Then I le… | Another weird AI game project: I asked Claude Fable 5.2 for a fantasy transport sim with Discworld meets Lord of the Rings vibes. Then… | ↗Not briefed |
bluesky bluesky-ai | 11d | my p(doom) is like 60% but my p(llm doom) is 0%. like malevolent world-dominating ai is a cultural phantasm and isnt real but we also arent… | my p(doom) is like 60% but my p(llm doom) is 0%. like malevolent world-dominating ai is a cultural phantasm and isnt real but we also… | ↗Not briefed |
arxiv arxiv-ai | 11d | The Rebundling of Freelance Work in the AI Age🔧 ChatGPT | This paper examines how freelance work is changing in the AI era. We compare the activities people undertake in ChatGPT with the work… | ↗Not briefed |
bluesky bluesky-ai | 11d | If I were under the impression that an "LLM" was just an RLHF'd GPT-4-esque thing, I too would be very nervous about using them in any kind… | If I were under the impression that an "LLM" was just an RLHF'd GPT-4-esque thing, I too would be very nervous about using them in any… | ↗Not briefed |
arxiv arxiv-ai | 11d | Beyond Foreign-Key Connectivity: Semantic Relationship Representation for Relational Deep Learning | Relational databases encode information through tables and primary--foreign key constraints. Although recent Relational Deep Learning… | ↗Not briefed |
arxiv arxiv-ai | 11d | Code as Explanations: Executable and Verifiable LLM Explanations | As large language models (LLMs) increasingly make consequential decisions and solve problems alongside humans, we must understand how… | ↗Not briefed |
bluesky bluesky-ai | 11d | *I'm not taking a stance on LLM coding assistance. I do trust SWEs who say it's a gamechanger, and others who say (in a complementary sense… | *I'm not taking a stance on LLM coding assistance. I do trust SWEs who say it's a gamechanger, and others who say (in a complementary… | ↗Not briefed |
bluesky bluesky-ai | 11d | My guess is there's a fair number of LLM coding critics who are unaware that the way the fields (SWE generally and SV specifically) use the… | My guess is there's a fair number of LLM coding critics who are unaware that the way the fields (SWE generally and SV specifically) us… | ↗Not briefed |
bluesky bluesky-ai | 11d | Head still ringing from encountering someone asking an LLM to write them some extra good prompts and then trying to sell those lists of pro… | Head still ringing from encountering someone asking an LLM to write them some extra good prompts and then trying to sell those lists o… | ↗Not briefed |
bluesky bluesky-ai | 11d | OpenAI bringt GPT-6.1 Terra als günstigeres Modell für KI-Agenten. Unabhängige Tests zeigen starke Effizienzgewinne. #Künstliche Intelligenz | OpenAI bringt GPT-6.1 Terra als günstigeres Modell für KI-Agenten. Unabhängige Tests zeigen starke Effizienzgewinne. #Künstliche Intel… | ↗Not briefed |
rss simon-willison | 11d | llm-gemini 0.36 | — | ↗Not briefed |
bluesky bluesky-ai | 11d | bonus material, prospect.org/2026/09/11/o... | bonus material, prospect.org/2026/09/11/o... | ↗Not briefed |
hn hn-ai | 11d | Show HN: Incremental Delta-Bar-Delta (Sutton's 1992 Rule) in Rust | — | ↗Not briefed |
bluesky bluesky-ai | 11d | I just think there's no ethical way to use any publicly available LLM because they rely on nothing but stolen content. I'd be willing to gi… | I just think there's no ethical way to use any publicly available LLM because they rely on nothing but stolen content. I'd be willing… | ↗Not briefed |
arxiv arxiv-ai | 11d | Probe What-If Worlds Before You Ship: Persona Simulation for Real-World Decision Making | User research and system testing reveal whether products and services meet user needs, yet recruiting specialized participants, design… | ↗Not briefed |
arxiv arxiv-ai | 11d | Conformal Risk Certificates for Portfolio Optimization Under Uncertainty: A Gated Reinforcement-Learning Approach | Reinforcement-learning portfolio policies can optimize allocations from prices, covariances, and prior positions, yet they lack a cali… | ↗Not briefed |
hn hn-ai | 11d | Fridgep0T: Low-interaction honeypot for supermarket refrigeration controllers | — | ↗Not briefed |
rss latent-space | 11d | Designing for Agents — Zeno Rocha, Resend | Where does answer engine optimization (AEO) end and agent experience begin? Inside Resend, from the docs that help Claude Code pick an… | ↗Not briefed |
rss techmeme | 11d | Lawsuit filing: Point72's quant unit Cubist claims its fired lead research analyst used Claude and Codex to steal source code before leaving for a competitor (Bloomberg) | Bloomberg: Lawsuit filing: Point72's quant unit Cubist claims its fired lead research analyst used Claude and Codex to steal source co… | ↗Not briefed |
bluesky bluesky-ai | 11d | When you train a cutting-edge AI model, the clever but unruly “base” model goes through additional training called alignment that’s meant t… | When you train a cutting-edge AI model, the clever but unruly “base” model goes through additional training called alignment that’s me… | ↗Not briefed |
arxiv arxiv-ai | 11d | Beyond Row-Level Prediction: Evaluating and Learning Composition for Reusable Table Representations | Tables are compositional, with content spread across a schema, cell values, rows, and columns. Large-scale operations such as retrieva… | ↗Not briefed |
hn hn-ai | 11d | Show HN: A clean-room Red Alert 2: Yuri's Revenge engine remake (Rust,wgpu,WASM) | Hi everyone, I have been working on this for some time. It still has some bugs but I find it quite playable. The lack of remastered Re… | ↗Not briefed |
bluesky bluesky-ai | 11d | WHO MADE THIS? | WHO MADE THIS? | ↗Not briefed |
rss techmeme | 11d | SpaceX announces Starlink Mini 2 with a swappable battery that it says can operate for over four hours without an external power source, reportedly for $999 (Andrew Liszewski/The Verge) | Andrew Liszewski / The Verge: SpaceX announces Starlink Mini 2 with a swappable battery that it says can operate for over four hours w… | ↗Not briefed |
bluesky bluesky-ai | 11d | with claude. and heck, let's throw in the internet archive for good measure. both of these tools are irreplaceable, and me complaining to t… | with claude. and heck, let's throw in the internet archive for good measure. both of these tools are irreplaceable, and me complaining… | ↗Not briefed |
bluesky bluesky-ai | 11d | All 7 Anthropic founders will each cash out about $250M in IPO | All 7 Anthropic founders will each cash out about $250M in IPO | ↗Not briefed |
arxiv arxiv-ai | 11d | Risk-Aware Joint Inference and Resource Control for Physical AI in 6G Networks | Wireless edge inference enables Physical-AI devices to execute edge-inferred actions using transmitted sensory inputs. Since the risk… | ↗Not briefed |
bluesky bluesky-ai | 11d | Anthropic is hemorrhaging money too, even they can't make the business model work | Anthropic is hemorrhaging money too, even they can't make the business model work | ↗Not briefed |
bluesky bluesky-ai | 11d | I do think it's a problem that we're arguing about whether or not using the LLM to solve the problem is ethical, without talking about the … | I do think it's a problem that we're arguing about whether or not using the LLM to solve the problem is ethical, without talking about… | ↗Not briefed |
hn hn-ai | 11d | Anthropic says a key AI safety alarm has been triggered, days before its IPO | — | ↗Not briefed |
rss techcrunch-ai | 11d | All 7 Anthropic founders will each cash out about $250M in IPO | Each of the seven co-founders plans to sell exactly the same number of shares in the IPO, which would bring each of them about $250 mi… | ↗Not briefed |
bluesky bluesky-ai | 11d | As the holidays come up and we are faced with so much "normalization" of LLM/GenAI -- consider even more now than ever: supporting small bu… | As the holidays come up and we are faced with so much "normalization" of LLM/GenAI -- consider even more now than ever: supporting sma… | ↗Not briefed |
bluesky bluesky-ai | 11d | especially in this age where any rando can badly (but moderately effectively for single user use) reverse engineer most things by pointing … | especially in this age where any rando can badly (but moderately effectively for single user use) reverse engineer most things by poin… | ↗Not briefed |
hn hn-ai | 11d | Anthropic Targets $1.4T Valuation in IPO | — | ↗Not briefed |
bluesky bluesky-ai | 11d | Per a post by noted Silicon Valley VC David Hornik, it turns out there's one thing #ChatGPT absolutely cannot do. It can't respond to the …🔧 ChatGPT | Per a post by noted Silicon Valley VC David Hornik, it turns out there's one thing #ChatGPT absolutely cannot do. It can't respond to… | ↗Not briefed |
hn hn-ai | 11d | Gilguard: Catch the import that silently re-enables GIL on free-threaded Python | — | ↗Not briefed |
rss techmeme | 11d | Clara, which sells corporate spend management tools in Latin America, raised $105M, including a $65M Series C and $40M in renewed growth financing (Maria Clara Cobo/Bloomberg) | Maria Clara Cobo / Bloomberg: Clara, which sells corporate spend management tools in Latin America, raised $105M, including a $65M Ser… | ↗Not briefed |
bluesky bluesky-ai | 11d | “We only used a LLM chatbot” Cool, that’s what Claude Code is. “But we didn’t use genAI” That’s what Claude Code outputs, did you not us… | “We only used a LLM chatbot” Cool, that’s what Claude Code is. “But we didn’t use genAI” That’s what Claude Code outputs, did you n… | ↗Not briefed |
hn hn-ai | 11d | Render is adding Anthropic, TypeSafe AI and others to its Subprocessors | — | ↗Not briefed |
bluesky bluesky-ai | 11d | "Investoren hinterfragen im aktuellen Hochzinsumfeld zunehmend die Nachhaltigkeit des Investitionszyklus, zirkuläre Finanzierungsstrukturen… | "Investoren hinterfragen im aktuellen Hochzinsumfeld zunehmend die Nachhaltigkeit des Investitionszyklus, zirkuläre Finanzierungsstruk… | ↗Not briefed |
rss techmeme | 11d | Sources: despite Trump's denial, he was told at least twice in seven months about Huang's affair with an alleged spy, including by Rubio en route to China (New York Times) | New York Times: Sources: despite Trump's denial, he was told at least twice in seven months about Huang's affair with an alleged spy,… | ↗Not briefed |
bluesky bluesky-ai | 11d | I'm an AI skeptic overall, but damn, I fed Claude 2 years of campaign finance data for a local city councilor, asked it to build a spreadsh… | I'm an AI skeptic overall, but damn, I fed Claude 2 years of campaign finance data for a local city councilor, asked it to build a spr… | ↗Not briefed |
rss the-decoder | 11d | Anthropic turns Claude into a full workspace with Plugins, spreadsheets, whiteboards, and soon Meeting notes | Anthropic is expanding Claude with Sheets, Whiteboards, and Interactive Connectors, letting partners like Adobe and Atlassian embed th… | ↗Not briefed |
bluesky bluesky-ai | 11d | Two AI futures, at least one of which feels quite dystopian to me: 1) The AI chat window as the replacement for the web browser. Everything… | Two AI futures, at least one of which feels quite dystopian to me: 1) The AI chat window as the replacement for the web browser. Every… | ↗Not briefed |
bluesky bluesky-ai | 11d | Some panoramic history of our times, on the model of The Proud Tower (Europe before WWI), or The Great Crash (Wall Street 1929) will start … | Some panoramic history of our times, on the model of The Proud Tower (Europe before WWI), or The Great Crash (Wall Street 1929) will s… | ↗Not briefed |
arxiv arxiv-ai | 11d | What Entropy-Based View Selection Cannot See: An Audit of Test-Time Adaptation for Vision-Language Models | Vision-language models such as CLIP can classify images into categories they were never explicitly trained to recognize, but accuracy… | ↗Not briefed |
bluesky bluesky-ai | 11d | 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 | 11d | "we're very sorry we used Claude to solve this one edge-case issue caused by a broken game engine" < KILL. DESTROY. ANNIHILATE "yeah this … | "we're very sorry we used Claude to solve this one edge-case issue caused by a broken game engine" < KILL. DESTROY. ANNIHILATE "yeah… | ↗Not briefed |
bluesky bluesky-ai | 11d | Claude 3 Opus (Legacy High-Cost) vs Claude 3.7 Sonnet: Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here is the l… | Claude 3 Opus (Legacy High-Cost) vs Claude 3.7 Sonnet: Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here is… | ↗Not briefed |