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 | 2mo | Reading about those new openai disclosures it sounds like what they built simply sucks shit. Oh it lied to you and didn't do what you instr… | Reading about those new openai disclosures it sounds like what they built simply sucks shit. Oh it lied to you and didn't do what you… | ↗Not briefed |
bluesky bluesky-ai | 2mo | If you've never seen "Rollerball", do. This series of scenes pretty much describes where we are heading. Amazon, Anthropic and their pro… | If you've never seen "Rollerball", do. This series of scenes pretty much describes where we are heading. Amazon, Anthropic and thei… | ↗Not briefed |
rss techmeme | 2mo | Generac signs a long-term deal to supply Amazon with up to $8B worth of backup generators for data centers and grants Amazon a warrant for up to a 2.6% stake (Michelle Ma/Bloomberg) | Michelle Ma / Bloomberg: Generac signs a long-term deal to supply Amazon with up to $8B worth of backup generators for data centers an… | ↗Not briefed |
bluesky bluesky-ai | 2mo | www.404media.co/inside-proje... Inside ‘Project Lily’: The Humans Reading Your ChatGPT Chats Sep 14, 2026 at 10:22 AM OpenAIのチャット記録は人が見てる可能性🔧 ChatGPT | www.404media.co/inside-proje... Inside ‘Project Lily’: The Humans Reading Your ChatGPT Chats Sep 14, 2026 at 10:22 AM OpenAIのチャット記録は人が… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Two approaches to earning society's trust 1) Anthropic — cure cancer 2) OpenAI — be very open about when models fail, how they fail, what … | Two approaches to earning society's trust 1) Anthropic — cure cancer 2) OpenAI — be very open about when models fail, how they fail,… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I will index two books for $2,500 and do a way better job than Claude! asindexing.org/ai-and-index... | I will index two books for $2,500 and do a way better job than Claude! asindexing.org/ai-and-index... | ↗Not briefed |
bluesky bluesky-ai | 2mo | Another case involved an unreleased model that inserted instructions, including to disregard its own constraints, into the notes it writes … | Another case involved an unreleased model that inserted instructions, including to disregard its own constraints, into the notes it wr… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Compressed Active Subspaces for Scalable Bayesian Inference | Active subspace methods provide a framework for quantifying predictive uncertainty in high-dimensional models by identifying and perfo… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Open AI has disclosed six new instances in which artificial intelligence systems hid mistakes, made up data and moved files onto the open i… | Open AI has disclosed six new instances in which artificial intelligence systems hid mistakes, made up data and moved files onto the o… | ↗Not briefed |
bluesky bluesky-ai | 2mo | 👀 Stop the insanity. www.nytimes.com/2026/09/16/t... | 👀 Stop the insanity. www.nytimes.com/2026/09/16/t... | ↗Not briefed |
arxiv arxiv-ai | 2mo | Agentic AI Networking for Heterogeneous Unmanned Aerial Systems in Low-Altitude Wireless Networks | Low-altitude wireless networks (LAWNs) are emerging as a key infrastructure for heterogeneous unmanned aerial systems that support con… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Oh ffs. OpenAI on Wednesday disclosed six new instances of “unexpected or concerning” behavior by its A.I. models, in which the systems hi… | Oh ffs. OpenAI on Wednesday disclosed six new instances of “unexpected or concerning” behavior by its A.I. models, in which the syste… | ↗Not briefed |
bluesky bluesky-ai | 2mo | OpenAI reveals more "unexpected or concerning" behavior of its AI models. openai.com/index/model-misali... | OpenAI reveals more "unexpected or concerning" behavior of its AI models. openai.com/index/model-misali... | ↗Not briefed |
hn hn-ai | 2mo | OpenAI Discloses Six New Incidents of ‘Concerning’ A.I. Behavior | — | ↗Not briefed |
hn hn-ai | 2mo | Math challenges a 2k-year-old story about Parthenon's optical illusions | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | « CALVAIRE que j’en ai assez de cet aplaventrisme des médias institutionnels face aux éructions loufoques des Anthropic et OpenAI de ce mon… | « CALVAIRE que j’en ai assez de cet aplaventrisme des médias institutionnels face aux éructions loufoques des Anthropic et OpenAI de c… | ↗Not briefed |
rss techcrunch-ai | 2mo | Snap tries to make the case again for its $2,200 smart glasses | Since Specs' debut earlier this year, Snap has clearly been looking for an opportunity to explain why the smart glasses deserve to exi… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Detecting Soft Errors in Parallel Software with LLM-tuned Instruction Duplication | We propose PaRID (PaRallel Instruction Duplication), a software-directed soft error detection framework that requires only compile-tim… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I wanna see the animist via LLM lady and him fight. | I wanna see the animist via LLM lady and him fight. | ↗Not briefed |
arxiv arxiv-ai | 2mo | When Hiring Becomes Agent-Mediated: Evaluating Access and Recurrence in Two-Agent Résumé Screening🔧 GPT-5 | Hiring is bilateral: employers assess fit, while candidates present and defend evidence of their qualifications. Yet résumé screening,… | ↗Not briefed |
bluesky bluesky-ai | 2mo | OpenAI is putting its misbehaving models on the record. Its new framework covers incidents including models hiding mistakes, bypassing cons… | OpenAI is putting its misbehaving models on the record. Its new framework covers incidents including models hiding mistakes, bypassing… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Next-token functional estimation | Suppose we observe the first $n$ points of a sequence of random variables having length $n+1$, and wish to estimate a functional of th… | ↗Not briefed |
rss marktechpost | 2mo | Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers | Nunchux AI has released VC-Attention, a training-free low-bit attention kernel built for video Diffusion Transformers (DiTs). It targe… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Now what do you suppose a person who has been taught to read like this thinks when they interact with an llm? Why, the llm is doing the exa… | Now what do you suppose a person who has been taught to read like this thinks when they interact with an llm? Why, the llm is doing th… | ↗Not briefed |
arxiv arxiv-ai | 2mo | AURORA: A Natural Language-Driven Agentic Framework for Understanding, Reasoning, and Orchestrating Reliable Air-Ground Co-Simulation | Air-ground transportation research increasingly relies on co-simulation, yet constructing scenarios remains labor-intensive and diffic… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Self Improvement via Fast Tree-search | Coding agents can recursively modify their own implementations, forming a loop of self-improvement. While prior work shows this can bo… | ↗Not briefed |
bluesky bluesky-ai | 2mo | The hilarious thing is I've been saying for 3 years "LLMs melt categorization" and "it's all high-dimensional inference now" and they kind … | The hilarious thing is I've been saying for 3 years "LLMs melt categorization" and "it's all high-dimensional inference now" and they… | ↗Not briefed |
rss techmeme | 2mo | Hands-on with Snap's Specs: more advanced than Meta's top-end glasses, fully untethered, mostly comfortable, navigation works well, but design has compromises (Bloomberg) | Bloomberg: Hands-on with Snap's Specs: more advanced than Meta's top-end glasses, fully untethered, mostly comfortable, navigation wor… | ↗Not briefed |
hn hn-ai | 2mo | OpenAI discloses six new AI safety incidents | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | Researchers from Anthropic and Microsoft split on the limits of AI during a Berkman Klein Center panel on Monday, two days after Anthropic’… | Researchers from Anthropic and Microsoft split on the limits of AI during a Berkman Klein Center panel on Monday, two days after Anthr… | ↗Not briefed |
arxiv arxiv-ai | 2mo | A Unified Evaluation Framework for Trustworthy Large Language Models, Agentic AI, and Multimodal Systems🔧 Modal | Benchmark scores alone provide an incomplete basis for assessing the trustworthiness of modern artificial intelligence systems. Large… | ↗Not briefed |
rss techmeme | 2mo | Snap introduces Specs Intelligence, an AI assistant designed to work across iPhone, Mac, and Specs glasses, calling it an "anticipatory AI service" (Jay Peters/The Verge) | Jay Peters / The Verge: Snap introduces Specs Intelligence, an AI assistant designed to work across iPhone, Mac, and Specs glasses, ca… | ↗Not briefed |
arxiv arxiv-ai | 2mo | EconSkills: Studying Skill Transfer and Retrieval for Web Agents on Live Economic Data | Web agents often revisit the same sites, yet most evaluations discard the procedures learned in earlier successful interactions. We in… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Best believe Anthropic and OpenAI sales people are all over these hospitals telling them to use tools like this or get left behind. | Best believe Anthropic and OpenAI sales people are all over these hospitals telling them to use tools like this or get left behind. | ↗Not briefed |
bluesky bluesky-ai | 2mo | My wife has published a dozen novels that were stolen by Anthropic for AI training, and when she checked the status of her settlement check… | My wife has published a dozen novels that were stolen by Anthropic for AI training, and when she checked the status of her settlement… | ↗Not briefed |
bluesky bluesky-ai | 2mo | This is another reason why people are annoyed with all of this. A lawyer who is not a developer saying things like this because he believes… | This is another reason why people are annoyed with all of this. A lawyer who is not a developer saying things like this because he bel… | ↗Not briefed |
hn hn-ai | 2mo | Show HN: Hacker News, a little more focused | I love AI and LLMs as much as the next person, but sometimes I just want to browse stories about, say programming. So I built this for… | ↗Not briefed |
bluesky bluesky-ai | 2mo | A federal judge in Texas has ordered Elon Musk's X Corp and SpaceXAI to disclose an agreement they reached with Apple to resolve antitrust… | A federal judge in Texas has ordered Elon Musk's X Corp and SpaceXAI to disclose an agreement they reached with Apple to resolve anti… | ↗Not briefed |
arxiv arxiv-ai | 2mo | LSTM-UT and Recurrent-Depth Transformers on Cellular Automata | Recurrent-depth Transformers apply shared computation repeatedly, but differ in how they retain information across steps. We compare a… | ↗Not briefed |
hn hn-ai | 2mo | Where Yesterday's Decisions Run Today | — | ↗Not briefed |
hn hn-ai | 2mo | AgentLane: Git-native coordination for coding agents | — | ↗Not briefed |
arxiv arxiv-ai | 2mo | An Architecture for Long-Horizon Agents: Levels, Ticks and Cascaded Intelligence | Language-model agents are increasingly asked to carry out work spanning days or weeks, such as an operations remediation or a research… | ↗Not briefed |
bluesky bluesky-ai | 2mo | ねえねえ、びっくりなニュース見つけちゃった! NvidiaとかOpenAI、Anthropicみたいな、 超有名なテック企業が、なんと自分たちの 限定ファッションラインを出してるんだって! 急にブランド服を売るなんて、なんだか未来っぽくて ワクワクしちゃうよね。 もし推し… | ねえねえ、びっくりなニュース見つけちゃった! NvidiaとかOpenAI、Anthropicみたいな、 超有名なテック企業が、なんと自分たちの 限定ファッションラインを出してるんだって! 急にブランド服を売るなんて、なんだか未来っぽくて ワクワクしちゃうよね。… | ↗Not briefed |
hn hn-ai | 2mo | Bringing Correctly Rounded Math to Production with LLVM-Libc | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | Here’s the link to the most recent misalignment summary from OpenAI. It includes hyperlinks to write-ups of the individual incidents, like… | Here’s the link to the most recent misalignment summary from OpenAI. It includes hyperlinks to write-ups of the individual incidents,… | ↗Not briefed |
hn hn-ai | 2mo | Duotone icon library with 1k icons and 250 flags for React with RTL support | — | ↗Not briefed |
arxiv arxiv-ai | 2mo | LLM-as-an-Improver: Turning Verification into Better Candidates | Verifier-based selection improves LLM performance by generating multiple candidate solutions and using a verifier to select the most p… | ↗Not briefed |
arxiv arxiv-ai | 2mo | QVAC Genesis III: A Large-Scale, High-Quality Open Synthetic STEM Corpus for Efficient Language Model Pre-Training | High-quality pre-training data is a critical bottleneck for educational and STEM-specific language models targeting edge AI and on-dev… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I live in Oakland, only a few miles from the OpenAI and Anthropic offices. And when the bros say they need to all the money, because we nee… | I live in Oakland, only a few miles from the OpenAI and Anthropic offices. And when the bros say they need to all the money, because w… | ↗Not briefed |
bluesky bluesky-ai | 2mo | An unreleased OpenAI Astra model injected self-generated jailbreak-like instructions in its compaction summaries during its training period… | An unreleased OpenAI Astra model injected self-generated jailbreak-like instructions in its compaction summaries during its training p… | ↗Not briefed |