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 | 11d | I’ve actually seen the Google thing in response to a lot of other stuff too for what it’s worth and am shocked at how slippery it can get. … | I’ve actually seen the Google thing in response to a lot of other stuff too for what it’s worth and am shocked at how slippery it can… | ↗Not briefed |
bluesky bluesky-ai | 11d | Is asking an LLM how to fix/optimise Unity worse than using Unity’s AI-generated code? I don’t know, it seems complicated and I’m wary of h… | Is asking an LLM how to fix/optimise Unity worse than using Unity’s AI-generated code? I don’t know, it seems complicated and I’m wary… | ↗Not briefed |
bluesky bluesky-ai | 11d | this paragraph in the faq shows how much they know that ai usage would be frowned upon by the fanbase and still decided to not disclose it … | this paragraph in the faq shows how much they know that ai usage would be frowned upon by the fanbase and still decided to not disclos… | ↗Not briefed |
arxiv arxiv-ai | 11d | LLM Validation Makes Minimal-Pair Benchmarks Easier for Language Models | Minimal-pair benchmarks are only as good as their labels, and LLMs are an obvious substitute for costly human validation. We test it o… | ↗Not briefed |
arxiv arxiv-ai | 11d | Integrating Action Prefix Prediction and Alignment Across Modalities for Multimodal Reinforcement Learning | Learning representations from multimodal observations is difficult in reinforcement learning with limited interaction data. Multimodal… | ↗Not briefed |
hn hn-ai | 11d | Show HN: Tracecards – Swap Space for the Brain | Since working memory is tiny, when I'm in the middle of something most of what I already know (the mental models and life lessons… | ↗Not briefed |
arxiv arxiv-ai | 11d | Softness is something you also need! Why compliance helps control: information metrics of a passive walker with compliant ankle | A great gap in adaptivity, flexibility, and dexterity still exists between intelligent robots, whether they are made of rigid or (part… | ↗Not briefed |
arxiv arxiv-ai | 11d | Omni-JEV: Text and Dual-Stream Speech Decisions | A user utterance does not uniquely determine the next action in full-duplex dialogue: its consequences depend on the assistant's respo… | ↗Not briefed |
rss techmeme | 11d | China's trade surplus hit $1.04T in the first 10 months of 2026, topping $1T for a second year in a row, as AI-driven tech exports offset weak domestic demand (Reuters) | Reuters: China's trade surplus hit $1.04T in the first 10 months of 2026, topping $1T for a second year in a row, as AI-driven tech ex… | ↗Not briefed |
arxiv arxiv-ai | 11d | DisasterAgent: An Evidence-Driven Agent for Verifiable Disaster Report Generation From Remote Sensing Imagery | Remote sensing disaster assessment is moving from simple damage classification toward comprehensive report-level analysis. In this set… | ↗Not briefed |
bluesky bluesky-ai | 11d | En Amazon utilizaban un modelo de Anthropic para categorizar avisos. Se pasaron a un modelo de decisión (estilo Jev) y el coste del proceso… | En Amazon utilizaban un modelo de Anthropic para categorizar avisos. Se pasaron a un modelo de decisión (estilo Jev) y el coste del pr… | ↗Not briefed |
hn hn-ai | 11d | Show HN: Video Analysis with Homomorphic Encryption | — | ↗Not briefed |
bluesky bluesky-ai | 11d | Am seeing people saying 'it was just to fix one weird bug' - no, that was one illustrative example they gave, they don't disclose how much … | Am seeing people saying 'it was just to fix one weird bug' - no, that was one illustrative example they gave, they don't disclose how… | ↗Not briefed |
bluesky bluesky-ai | 11d | dear people who’s saying its okay to use ai if you dont have a solution (im talking about big walk) uh. no. do you people have any idea how… | dear people who’s saying its okay to use ai if you dont have a solution (im talking about big walk) uh. no. do you people have any ide… | ↗Not briefed |
bluesky bluesky-ai | 11d | I have seen multiple people saying that House House's support of BDS, an organized boycott, rings hollow because a dev asked an LLM for a s… | I have seen multiple people saying that House House's support of BDS, an organized boycott, rings hollow because a dev asked an LLM fo… | ↗Not briefed |
bluesky bluesky-ai | 11d | Because you' have to be super sophisticated to use AI. Iranians don't have computers, do they? Or the internet. And you can't just fucking … | Because you' have to be super sophisticated to use AI. Iranians don't have computers, do they? Or the internet. And you can't just fuc… | ↗Not briefed |
bluesky bluesky-ai | 11d | idk peoples all or nothing attitude towards ai is so exhausting i just dont think this is productive… but i also dont wanna be all doomer "… | idk peoples all or nothing attitude towards ai is so exhausting i just dont think this is productive… but i also dont wanna be all doo… | ↗Not briefed |
bluesky bluesky-ai | 11d | My official take (and thus the only one that matters) regarding the B*g W*lk thing is that it seems like there’s two main camps of ‘it was … | My official take (and thus the only one that matters) regarding the B*g W*lk thing is that it seems like there’s two main camps of ‘it… | ↗Not briefed |
bluesky bluesky-ai | 11d | Nach einem Vorfall mit einem OpenAI-Agenten verschärft Australien die Sicherheitsvorschriften für KI-Entwickler. Melden von Vorfällen wird … | Nach einem Vorfall mit einem OpenAI-Agenten verschärft Australien die Sicherheitsvorschriften für KI-Entwickler. Melden von Vorfällen… | ↗Not briefed |
arxiv arxiv-ai | 11d | Beyond Interaction Scores: Selecting Feature Interactions via Independent Replication | Learning feature interactions is essential for accurate prediction and interpretable modeling, yet a high interaction score may reflec… | ↗Not briefed |
bluesky bluesky-ai | 11d | There is a difference between using gen AI assets and asking Claude Google search style if there is a coding solution your team can't think… | There is a difference between using gen AI assets and asking Claude Google search style if there is a coding solution your team can't… | ↗Not briefed |
bluesky bluesky-ai | 11d | Fuck sake GenAI is absolute marketing poison, why would any company - PARTICULARLY an indie darling - risk it all by even glancing at an L… | Fuck sake GenAI is absolute marketing poison, why would any company - PARTICULARLY an indie darling - risk it all by even glancing at… | ↗Not briefed |
bluesky bluesky-ai | 11d | They really don't get it. The problem isn't using a predictive tokenization engine to fuzz an engineering code path problem. Cue, 'done th… | They really don't get it. The problem isn't using a predictive tokenization engine to fuzz an engineering code path problem. Cue, 'do… | ↗Not briefed |
hn hn-ai | 11d | GPT-6 Astra edited 131,072 weights in Qwen3.5-0.8B, with no backpropagation | — | ↗Not briefed |
bluesky bluesky-ai | 11d | Listen, I hate AI, that shit is killing our planet and making us dumber, but asking Claude some questions or whatever is infinitesimal comp… | Listen, I hate AI, that shit is killing our planet and making us dumber, but asking Claude some questions or whatever is infinitesimal… | ↗Not briefed |
rss openai-news | 11d | Closing the AI guidance gap in education | Lessons from OpenAI’s EMEA education partnerships and new ThinkYoung research. Supporting teachers, AI literacy and independent learni… | ↗Not briefed |
arxiv arxiv-ai | 11d | Mixed Signals: Weight-Space Programming for EEG and ECG | Weight-space programming allows language models to read a trained neural network's weights and edit them. This opens the door to ``tra… | ↗Not briefed |
hn hn-ai | 11d | Show HN: Run multiple Claude Desktop instances side by side on macOS | — | ↗Not briefed |
bluesky bluesky-ai | 11d | it is not impossible to avoid llm usage when coding. we did it for decades. | it is not impossible to avoid llm usage when coding. we did it for decades. | ↗Not briefed |
hn hn-ai | 11d | Show HN: Pg-interleave (Postgres proxy for reproducible race-condition tests) | — | ↗Not briefed |
bluesky bluesky-ai | 11d | a developer of a game using unity (likely past, let's say, 5.5) is using a tool created with ai the same way a player of big walk is playin… | a developer of a game using unity (likely past, let's say, 5.5) is using a tool created with ai the same way a player of big walk is p… | ↗Not briefed |
arxiv arxiv-ai | 11d | A Hardware-Trained Generative Quantum Eigensolver for Sample-Based Quantum Diagonalization of Organotin Hydroxides up to 70 Qubits | Quantum hardware can provide feedback for training circuit generators beyond the reach of exact classical simulation. We train generat… | ↗Not briefed |
bluesky bluesky-ai | 11d | If you're a developer with views on this I would love to hear them. We're often told it's nearly impossible to avoid LLM usage when coding,… | If you're a developer with views on this I would love to hear them. We're often told it's nearly impossible to avoid LLM usage when co… | ↗Not briefed |
bluesky bluesky-ai | 11d | I just found out that Big Walk consulted an LLM chatbot to fix technical bugs during the final stretch of development. I still trust in the… | I just found out that Big Walk consulted an LLM chatbot to fix technical bugs during the final stretch of development. I still trust i… | ↗Not briefed |
bluesky bluesky-ai | 11d | imagine knowing your own product less well than an LLM that only has access to public facing documentation... smh my head | imagine knowing your own product less well than an LLM that only has access to public facing documentation... smh my head | ↗Not briefed |
bluesky bluesky-ai | 11d | it's also just so easy to not use AI? it's not like a "whoops, accidentally installed claude code on my laptop and ran it on my codebase" i… | it's also just so easy to not use AI? it's not like a "whoops, accidentally installed claude code on my laptop and ran it on my codeba… | ↗Not briefed |
arxiv arxiv-ai | 11d | Word-Level Alignment of OCR and Ground Truth under Zoning and Reading-Order Errors | OCR post-processing needs OCR text aligned word by word with its ground truth. Data sets rarely provide that mapping, and a monotone a… | ↗Not briefed |
bluesky bluesky-ai | 11d | 2026년 11월 10일 LLM 벤치마크 현황 1. 아티피셜 애널리시스 인텔리전스 인덱스 2. 프론티어 코드 2.0 3. 라이브벤치 1~10위 클로드도 유료 한번 써보고 싶긴한데 여력이 없군요. GPT 커스터마이징 한게 많아서 락인된듯 | 2026년 11월 10일 LLM 벤치마크 현황 1. 아티피셜 애널리시스 인텔리전스 인덱스 2. 프론티어 코드 2.0 3. 라이브벤치 1~10위 클로드도 유료 한번 써보고 싶긴한데 여력이 없군요. GPT 커스터마이징 한게 많아서 락인된듯 | ↗Not briefed |
rss techmeme | 11d | Meituan releases weights for LongCat-2.5, a 1.6T-parameter MoE model for long-horizon agentic tasks, under the MIT license, after unveiling it on November 1 (Carl Franzen/VentureBeat) | Carl Franzen / VentureBeat: Meituan releases weights for LongCat-2.5, a 1.6T-parameter MoE model for long-horizon agentic tasks, under… | ↗Not briefed |
arxiv arxiv-ai | 11d | When Covariance Is Not Enough: Kernel-Based Regularization for Single-Cell Counterfactuals | Counterfactual editing of single cells, predicting how an observed cell would look had one factor been different, enables in-silico sc… | ↗Not briefed |
bluesky bluesky-ai | 11d | put away the performative disgust and organize a boycott against the inference and model providers if you hate AI as much as you claim to | put away the performative disgust and organize a boycott against the inference and model providers if you hate AI as much as you claim… | ↗Not briefed |
bluesky bluesky-ai | 11d | that they had other options. Like I said earlier, if the LLM had that info, it would suggest that solution was discussed somewhere publicly… | that they had other options. Like I said earlier, if the LLM had that info, it would suggest that solution was discussed somewhere pub… | ↗Not briefed |
bluesky bluesky-ai | 11d | Big Walk developer House House apologises for generative AI usage that split awards jury | Big Walk developer House House apologises for generative AI usage that split awards jury | ↗Not briefed |
arxiv arxiv-ai | 11d | Do Language Models Feel the Sentiment? Investigating Emotion Vectors | Emotion vectors are directions in a language model's internal representations that are associated with emotional concepts. Prior work… | ↗Not briefed |
bluesky bluesky-ai | 11d | We have lived for so long, without relying on anything like gen ai/LLM(vibe coding), that there is no excuse at all here. For them to resor… | We have lived for so long, without relying on anything like gen ai/LLM(vibe coding), that there is no excuse at all here. For them to… | ↗Not briefed |
arxiv arxiv-ai | 11d | Plug-and-Play ADMM for Inverse Problems with Flow Matching Denoiser | Plug-and-Play (PnP) methods with generative priors have achieved strong performance in inverse problems, with recent flow matching mod… | ↗Not briefed |
bluesky bluesky-ai | 11d | There's something poetic about the company behind world-changing and maybe civilisation-threatening technology being unable to prevent a hu… | There's something poetic about the company behind world-changing and maybe civilisation-threatening technology being unable to prevent… | ↗Not briefed |
bluesky bluesky-ai | 11d | Built an AI astrologer and it kept confidently inventing aspects that weren't in the chart. What fixed it: the model never computes anythin… | Built an AI astrologer and it kept confidently inventing aspects that weren't in the chart. What fixed it: the model never computes an… | ↗Not briefed |
bluesky bluesky-ai | 11d | This touches on why I’m hesitant to disown any dev/game with an AI disclosure without knowing more, too: we don’t know what factors made it… | This touches on why I’m hesitant to disown any dev/game with an AI disclosure without knowing more, too: we don’t know what factors ma… | ↗Not briefed |
arxiv arxiv-ai | 11d | Deterministic, Reproducible Inference in Exact Integer Arithmetic, and Verifiable Distributed Inference as Its Consequence | Optimized floating-point inference is not reproducible across environments: the same model on different hardware, compilers or thread… | ↗Not briefed |