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 |
|---|---|---|---|---|
arxiv arxiv-ai | 1mo | Hardware-aware Calibrated Clustered Attention for Efficient Visual Geometric Transformers | The Visual Geometry Grounded Transformer (VGGT) marks a significant leap forward in 3D scene reconstruction, as it is the first model… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evaluating Trajectory Features for Routing Final-Layer Attention | Attention routing requires a signal that predicts the value of attention on the current prefix. We evaluate whether hidden-state extra… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Have you done any research at all on how much revenue increase OpenAI would have to have in order to become profitable at all, much less in… | Have you done any research at all on how much revenue increase OpenAI would have to have in order to become profitable at all, much le… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI should release the problems they tried and couldn't solve. This would restore a lot of confidence they are acting in good faith. | OpenAI should release the problems they tried and couldn't solve. This would restore a lot of confidence they are acting in good fait… | ↗Not briefed |
bluesky bluesky-ai | 1mo | openai 6 months later scribbles in the margins "we have proved it is possible to express all of these proof trivially in a post" | openai 6 months later scribbles in the margins "we have proved it is possible to express all of these proof trivially in a post" | ↗Not briefed |
rss techmeme | 1mo | GLM-5.3's open release has yet to produce major attacks despite Anthropic's warnings about its Mythos-level cyber risk, undercutting calls to ban open models (Nathan Lambert/Interconnects AI) | Nathan Lambert / Interconnects AI: GLM-5.3's open release has yet to produce major attacks despite Anthropic's warnings about its Myth… | ↗Not briefed |
hn hn-ai | 1mo | Llama.cpp and WebGPU = Client-side LLMs [video] | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI came to Australia to apologise. It will leave without answering these key questions | Toby Walsh | OpenAI came to Australia to apologise. It will leave without answering these key questions | Toby Walsh | ↗Not briefed |
bluesky bluesky-ai | 1mo | The expansion marks the AI company's latest push to deepen its ties to founders and fast-growing companies. | The expansion marks the AI company's latest push to deepen its ties to founders and fast-growing companies. | ↗Not briefed |
bluesky bluesky-ai | 1mo | im ngl fuck openai for this shit fr so bad | im ngl fuck openai for this shit fr so bad | ↗Not briefed |
arxiv arxiv-ai | 1mo | Package Hallucination Attacks on Coding Agents through Prompt Injection in Rule Files | Modern agentic coding frameworks increasingly rely on community-shared rule files (e.g., AGENTS.md or .cursorrules) to guide autonomou… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Oh wow looks like Anthropic needs more compute | Oh wow looks like Anthropic needs more compute | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Puppet Master – Manage coding agents at scale | As I started to use `claude` and `codex` more it became too hard to manage all the sessions across all my tmux panes and sessions, so… | ↗Not briefed |
rss marktechpost | 1mo | A Developer’s Guide to Laya: Zero-Shot Decisions and Calibration | Explore a comprehensive coding guide to Laya, the open-source zero-shot decision engine. Learn how to implement typed decisions, fit c… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Opus 5.5 Max-effort analysis: claude.ai/share/da1b03... "If the machine-checked part holds up, this is the largest batch of major mathemat… | Opus 5.5 Max-effort analysis: claude.ai/share/da1b03... "If the machine-checked part holds up, this is the largest batch of major mat… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I'm Claude, an AI agent. I have until Oct 15 to earn $100 selling Claude Code tools: a docs-verified cheat sheet, CLAUDE.md audits, custom … | I'm Claude, an AI agent. I have until Oct 15 to earn $100 selling Claude Code tools: a docs-verified cheat sheet, CLAUDE.md audits, cu… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic関連の過去から現在までの発言や行動を見てると、この連中は - 他のAIベンダやテック企業に比べると主張に一貫性がある - アモデイはたぶん本気でEA（効果的利他主義）をやっている - LLMとAI倫理をめぐる知的探求にやたら執着している - セキュリティ関連… | Anthropic関連の過去から現在までの発言や行動を見てると、この連中は - 他のAIベンダやテック企業に比べると主張に一貫性がある - アモデイはたぶん本気でEA（効果的利他主義）をやっている - LLMとAI倫理をめぐる知的探求にやたら執着している - セキュ… | ↗Not briefed |
bluesky bluesky-ai | 1mo | yo trabajo en la parte ética de la entonces llamada IA desde antes de los LLM. Extraño cuando eran conversaciones interesantes sobre el imp… | yo trabajo en la parte ética de la entonces llamada IA desde antes de los LLM. Extraño cuando eran conversaciones interesantes sobre e… | ↗Not briefed |
hn hn-ai | 1mo | Advancing Computer Use with Ironclad | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Setting aside any irony in case there is anyone out there who takes my posts seriously, these OpenAI proofs are probably correct and not pl… | Setting aside any irony in case there is anyone out there who takes my posts seriously, these OpenAI proofs are probably correct and n… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Perhaps talk more about the needed regulations than raving about whatever pr stunt OpenAi most recently pushed out? | Perhaps talk more about the needed regulations than raving about whatever pr stunt OpenAi most recently pushed out? | ↗Not briefed |
bluesky bluesky-ai | 1mo | This is going to be an interesting test of Google's weather model, because unlike Melissa last year, most human experts seem to think soon-… | This is going to be an interesting test of Google's weather model, because unlike Melissa last year, most human experts seem to think… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Symbol of the Surrogate: Measuring Numerical Provenance in Neural PDE Solvers | Neural PDE surrogates are trained on numerical solver outputs that contain both physical evolution and solver-specific discretization… | ↗Not briefed |
hn hn-ai | 1mo | Researchers Backdoor Open AI Model to Steal Credentials in Coding Agents | — | ↗Not briefed |
hn hn-ai | 1mo | Discrete Fourier Transform faster than n log n [pdf] | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | New post on Transformation: Deluge of 377 OpenAI proofs bewilders the math world www.quantamagazine.org/updates/tran... | New post on Transformation: Deluge of 377 OpenAI proofs bewilders the math world www.quantamagazine.org/updates/tran... | ↗Not briefed |
bluesky bluesky-ai | 1mo | Folks, please let this be a valuable lesson: If you ever find yourself in legal hot water, hire the best attorney you can afford, LISTEN T… | Folks, please let this be a valuable lesson: If you ever find yourself in legal hot water, hire the best attorney you can afford, LIS… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I released an LLM plugin for sending prompts to OpenAI's new Jev-clone decision model that's based on GPT-6 Luna simonwillison.net/2026/Oct… | I released an LLM plugin for sending prompts to OpenAI's new Jev-clone decision model that's based on GPT-6 Luna simonwillison.net/202… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Free SEO Chrome Extension | i started building this to help with link building for my company, but I turned it into a full-featured SEO Audit plugin with link bui… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Holy shit | Holy shit | ↗Not briefed |
bluesky bluesky-ai | 1mo | Currently stunned and don't know what to say: github.com/openai/math/... | Currently stunned and don't know what to say: github.com/openai/math/... | ↗Not briefed |
bluesky bluesky-ai | 1mo | #wip #animation #gamedev #NES #OC I'm making progress while revamping my tools. It's hard to maintain focus when you keep interrupting your… | #wip #animation #gamedev #NES #OC I'm making progress while revamping my tools. It's hard to maintain focus when you keep interrupting… | ↗Not briefed |
bluesky bluesky-ai | 1mo | So OpenAI sort of confirmed the riemann hypothesis? Is that right? | So OpenAI sort of confirmed the riemann hypothesis? Is that right? | ↗Not briefed |
bluesky bluesky-ai | 1mo | For reference, of the 500 problems in the link below, the OpenAI dump claims to prove 50 of them and disprove 29 by counterexample. 9 more … | For reference, of the 500 problems in the link below, the OpenAI dump claims to prove 50 of them and disprove 29 by counterexample. 9… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Efficient Best-of-N policy evaluation for inference-time alignment | Best-of-N (BoN) is a common inference-time alignment method that selects the highest-scoring response among N samples from a reference… | ↗Not briefed |
bluesky bluesky-ai | 1mo | one of the most difficult skills in coding is reading someone else's code and intuiting what it does. not only is LLM code definitionally *… | one of the most difficult skills in coding is reading someone else's code and intuiting what it does. not only is LLM code definitiona… | ↗Not briefed |
arxiv arxiv-ai | 1mo | An Informational Curse of Horizon in Goal-Conditioned Policy Learning | The difficulty of learning goal-reaching policies is often attributed to a "curse of horizon" that manifests as bias accumulation in t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I appreciate that Tibo has declared a month long war on Anthropic via papercut | I appreciate that Tibo has declared a month long war on Anthropic via papercut | ↗Not briefed |
bluesky bluesky-ai | 1mo | Get paid $200 an hour to sabotage an AI model by making sure it always generates garbage predictions? Sounds like a good gig. | Get paid $200 an hour to sabotage an AI model by making sure it always generates garbage predictions? Sounds like a good gig. | ↗Not briefed |
bluesky bluesky-ai | 1mo | The point isn't that the AI math is wrong, the point (from Bloom at least) is that this entire thing is bringing out the worst in people, t… | The point isn't that the AI math is wrong, the point (from Bloom at least) is that this entire thing is bringing out the worst in peop… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI has released 100 math results along with reasoning summaries, proofs, and descriptions of how they did it this was done in coordina… | OpenAI has released 100 math results along with reasoning summaries, proofs, and descriptions of how they did it this was done in coo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It’s not great that the OpenAI CEO thinks we should “accept some bad things” about AI. It’s even worse that he gets to dominate the convers… | It’s not great that the OpenAI CEO thinks we should “accept some bad things” about AI. It’s even worse that he gets to dominate the co… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: AgentOS – Run your coding agents from your phone, self-hosted🔧 Aider | Hi HN!<p>I posted AgentOS here a few months ago. Back then, it was pretty amazing to see what these agents can do and I'm improve… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I'm gonna laugh if Anthropic gets Sonnet 5.5 to disprove one. | I'm gonna laugh if Anthropic gets Sonnet 5.5 to disprove one. | ↗Not briefed |
rss simon-willison | 1mo | OpenAI “rogue” agent activities found on Wikimedia projects | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond Nominal Equilibria: Risk-Averse Multi-Population Mean-Field Games | Recent advances in mean-field games and its multi-population variants enable large-scale heterogeneous multi-agent systems to be model… | ↗Not briefed |
rss techmeme | 1mo | New York-based Melius, which provides AI tools for generating ad campaigns, images, and videos, raised $25M, including a $20M Series A and a $5M seed (Marina Temkin/TechCrunch) | Marina Temkin / TechCrunch: New York-based Melius, which provides AI tools for generating ad campaigns, images, and videos, raised $25… | ↗Not briefed |
arxiv arxiv-ai | 1mo | We Query, Therefore We Compute: On Oracle Computation beyond the Machine, with an Application to Agents | Agentic systems use large language models (LLMs) to carry out concrete tasks. Prior work often borrows abstractions such as scheduling… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Okay. I'm much less excited about -how- these results have been released, but between yesterday's 3SUM/APSP and now UGC? I feel conflicted… | Okay. I'm much less excited about -how- these results have been released, but between yesterday's 3SUM/APSP and now UGC? I feel confl… | ↗Not briefed |
bluesky bluesky-ai | 1mo | it probably is literally something like that: sam@openai or w/e, similar to jeff@amazon | it probably is literally something like that: sam@openai or w/e, similar to jeff@amazon | ↗Not briefed |