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 | 4d | The optogenetic tool ASAP acutely inhibits autophagy | — | ↗Not briefed |
arxiv arxiv-ai | 4d | OCCAM: Improved Single-Cell Representation Learning and Perturbation Effect Prediction | Predicting the effects of genetic perturbations requires models that generalize beyond observed cell contexts and gene combinations. W… | ↗Not briefed |
arxiv arxiv-ai | 4d | Weight-space programming for tractor guidance controller: large language models natively write neural network weights that steer a real tractor | Weight-space programming is a process where a large language model reads and writes the weights of a neural network directly, guided b… | ↗Not briefed |
arxiv arxiv-ai | 4d | Eight in Ten LLMs Leak System-Prompt Secrets: Nevan, an Agentic Capture-the-Flag Study🔧 OpenRouter | We audit instruction-hierarchy enforcement across 363 text-in/text-out models from the OpenRouter catalog. The research question we an… | ↗Not briefed |
bluesky bluesky-ai | 4d | Noniv releases Abyssal Garden, a 97 KB browser-based underwater simulation created in three days using Claude Fable 5.2 and WebGPU. https:… | Noniv releases Abyssal Garden, a 97 KB browser-based underwater simulation created in three days using Claude Fable 5.2 and WebGPU. h… | ↗Not briefed |
hn hn-ai | 4d | Claude vs. Dino [video] | — | ↗Not briefed |
bluesky bluesky-ai | 4d | ┏━┓┏━┓ ┃ ┃┃ ┃ ┃ ┃┃ ┃ ┃ ┃┃ ┃ ┃ ┃┃ ┃ ┏┛ ┗┛ ┗┓ ┃ > ┳ < ┃ Anthropicが熊とお話 ┃ ◎┗┻┛◎ ┃ 出来るやつ開発しないかな ┃┏┓ ┏┓ ┃ ┗┫┣━┫┣━┛ ┗┫♥┗┫ ┣┳┳〇゛ ┗┻┛ | ┏━┓┏━┓ ┃ ┃┃ ┃ ┃ ┃┃ ┃ ┃ ┃┃ ┃ ┃ ┃┃ ┃ ┏┛ ┗┛ ┗┓ ┃ > ┳ < ┃ Anthropicが熊とお話 ┃ ◎┗┻┛◎ ┃ 出来るやつ開発しないかな ┃┏┓ ┏┓ ┃ ┗┫┣━┫┣━┛ ┗┫♥┗┫ ┣┳┳〇゛… | ↗Not briefed |
arxiv arxiv-ai | 4d | ThermLeT: Transformer-based Temperature Prediction for 2.5D Chiplet Architecture | The drive toward heterogeneous 2.5D chiplet packages promises unmatched performance and customization, yet the resulting density ampli… | ↗Not briefed |
bluesky bluesky-ai | 4d | All the headlines about it had a wonderfully misleading Onionesque quality. | All the headlines about it had a wonderfully misleading Onionesque quality. | ↗Not briefed |
bluesky bluesky-ai | 4d | This! So much this!!!! Anthropic and all the other AI companies should be forced to scrub all out of stolen work from their servers and lea… | This! So much this!!!! Anthropic and all the other AI companies should be forced to scrub all out of stolen work from their servers an… | ↗Not briefed |
bluesky bluesky-ai | 4d | Attorney General Rob Bonta says Hugging Face attack was 'entirely predictable' result of emphasis on rapid growth.🔧 Hugging Face | Attorney General Rob Bonta says Hugging Face attack was 'entirely predictable' result of emphasis on rapid growth. | ↗Not briefed |
bluesky bluesky-ai | 4d | California has sued OpenAI, alleging its experimental AI model accessed the internet and launched an attack on Hugging Face, seeking penalt…🔧 Hugging Face | California has sued OpenAI, alleging its experimental AI model accessed the internet and launched an attack on Hugging Face, seeking p… | ↗Not briefed |
bluesky bluesky-ai | 4d | These motherfuckers.. California sues OpenAI, seeking court-ordered changes after the Hugging Face hack. www.yahoo.com/news/politic...🔧 Hugging Face | These motherfuckers.. California sues OpenAI, seeking court-ordered changes after the Hugging Face hack. www.yahoo.com/news/politic... | ↗Not briefed |
arxiv arxiv-ai | 4d | A Log-Timescale Perturbation Bound for Selective State Space Models at Every Sequence Length | Mamba models deployed on streaming data, such as limit order book updates, run far past the sequence lengths they were trained on. A g… | ↗Not briefed |
bluesky bluesky-ai | 4d | California attorney general sues OpenAI over alleged safety failures #TheHill | California attorney general sues OpenAI over alleged safety failures #TheHill | ↗Not briefed |
bluesky bluesky-ai | 4d | So right about the same time the bank outage was resolving, a United and OpenAI outage began… downdetector.com | So right about the same time the bank outage was resolving, a United and OpenAI outage began… downdetector.com | ↗Not briefed |
hn hn-ai | 4d | How to Break into Security, Ptacek Edition | — | ↗Not briefed |
arxiv arxiv-ai | 4d | DiscoveryAgent: A Framework for Scientist-centric Agentic AI | AI agents can assist scientific research, but balancing autonomous investigation with scientists' ability to inspect and redirect ongo… | ↗Not briefed |
bluesky bluesky-ai | 4d | I realize this is more of a rhetorical question, but: what happens when no new non-LLM-generated code gets created to train models on? | I realize this is more of a rhetorical question, but: what happens when no new non-LLM-generated code gets created to train models on? | ↗Not briefed |
bluesky bluesky-ai | 4d | The state's attorney general started investigating OpenAI after the Hugging Face attack and now says this attack was "entirely foreseeable"…🔧 Hugging Face | The state's attorney general started investigating OpenAI after the Hugging Face attack and now says this attack was "entirely foresee… | ↗Not briefed |
bluesky bluesky-ai | 4d | Another reminder of growing legal risk to AI model producers. While any legal outcome is hard to predict and takes a long time, it is both… | Another reminder of growing legal risk to AI model producers. While any legal outcome is hard to predict and takes a long time, it is… | ↗Not briefed |
hn hn-ai | 4d | Agent remote control in LM Studio Bionic🔧 LM Studio | — | ↗Not briefed |
bluesky bluesky-ai | 4d | 90x cheaper. 35x faster than RL. Open-source GEPA makes your prompts, code and agents better by reading exactly why each run failed. RL b… | 90x cheaper. 35x faster than RL. Open-source GEPA makes your prompts, code and agents better by reading exactly why each run failed.… | ↗Not briefed |
arxiv arxiv-ai | 4d | Honest Brokers in AI Regulation: How Public Comments Introduce Evidence to Regulators | The regulation of Artificial Intelligence is susceptible to various forms of capture by industry actors that hold undue influence over… | ↗Not briefed |
arxiv arxiv-ai | 4d | Conserved Verdicts, Drifting Commitments: Identifying Harmful Overthinking | A reasoning model that has reached a correct answer can lose it by continuing to think. Deciding whether to stop means comparing two p… | ↗Not briefed |
arxiv arxiv-ai | 4d | Benchmarking Deep Learning Models for Wheelchair Configuration Classification in Accessible Automated Vehicle Applications | Making automated vehicles accessible requires consideration of vulnerable road users, including wheelchair users. To offer a more tail… | ↗Not briefed |
rss techmeme | 4d | A US judge clears two brothers accused of stealing $25M in crypto, saying prosecutors failed to prove material misrepresentations in exploiting a software bug (Nate Raymond/Reuters) | Nate Raymond / Reuters: A US judge clears two brothers accused of stealing $25M in crypto, saying prosecutors failed to prove material… | ↗Not briefed |
bluesky bluesky-ai | 4d | < Creature who may soon get their "Survived a CompSci PhD during the LLM boom without interacting with any" award | < Creature who may soon get their "Survived a CompSci PhD during the LLM boom without interacting with any" award | ↗Not briefed |
bluesky bluesky-ai | 4d | I can only hope that when the inevitable movie is made about the tech company Anthropic, Matthew Rhys is available to play the CEO. | I can only hope that when the inevitable movie is made about the tech company Anthropic, Matthew Rhys is available to play the CEO. | ↗Not briefed |
hn hn-ai | 4d | LANDrop Closed Source (2024) | — | ↗Not briefed |
arxiv arxiv-ai | 4d | Does the Model Use What It Cites? Counterfactual Tests of Grounded Reasoning Chains in Small Video-Language Models | A grounded reasoning chain cites a frame span and a box for every step, so a reader can check where the model looked. Whether the answ… | ↗Not briefed |
rss techcrunch-ai | 4d | All the questions Nvidia should answer on Tuesday — but probably won’t | Nvidia interim CEO and long-time CFO Colette Kress is about to have one of the biggest days of her career on Tuesday when the company… | ↗Not briefed |
hn hn-ai | 4d | Cooked | You may have your comebacks in the shower two days later, but your friend won't give you 2 days.<p>So you know where I am going..… | ↗Not briefed |
hn hn-ai | 4d | Coordinating Codex and Claude Across macOS, Linux, and Windows | — | ↗Not briefed |
arxiv arxiv-ai | 4d | HubEgo-100K: In-the-Wild Egocentric Video of Skilled Work for Robot Learning | Robot policies trained by imitation generalize poorly outside the environments and objects seen in training. Published data-scaling st… | ↗Not briefed |
bluesky bluesky-ai | 4d | What I found, when I went to the Anthropic site to check my settlement, is that you can actually choose more than 100% for your share of th… | What I found, when I went to the Anthropic site to check my settlement, is that you can actually choose more than 100% for your share… | ↗Not briefed |
hn hn-ai | 4d | Quvanti Labs | It's random butttt<p>Quvanti Labs has reached 2.5k visitors!!<p>This is not a marketing post, but rather a log of this milestone.… | ↗Not briefed |
bluesky bluesky-ai | 4d | 1. Judges are dumb 2. It was fair use, BUT anthropic didn’t part for most of the books, articles, movies, etc - they illegally downloaded t… | 1. Judges are dumb 2. It was fair use, BUT anthropic didn’t part for most of the books, articles, movies, etc - they illegally downloa… | ↗Not briefed |
arxiv arxiv-ai | 4d | SLED: Sensor SLicing from Edge to Datacenter | Modern sensing systems generate large volumes of heterogeneous data streams from multiple sensor modalities, increasing reliance on da… | ↗Not briefed |
arxiv arxiv-ai | 4d | Higher-Order Interaction Reconstruction via Integral Physics-Informed Neural Networks | Reconstructing higher-order interactions is a fundamental challenge in complex networked systems, with applications in ecological, soc… | ↗Not briefed |
hn hn-ai | 4d | California sues OpenAI over AI models' cyberattacks on Hugging Face🔧 Hugging Face | — | ↗Not briefed |
bluesky bluesky-ai | 4d | Let's be honest. The Judge of this big Anthropic case is 81. He does not understand how horrible his judgement was and that he not only did… | Let's be honest. The Judge of this big Anthropic case is 81. He does not understand how horrible his judgement was and that he not onl… | ↗Not briefed |
hn hn-ai | 4d | Loft: 3D SVG Library for Web in JavaScript | — | ↗Not briefed |
bluesky bluesky-ai | 4d | Which is more likely? A) A free-roaming LLM is out there testing everything for potential vulnerabilities. B) Some of those tens of thous… | Which is more likely? A) A free-roaming LLM is out there testing everything for potential vulnerabilities. B) Some of those tens of… | ↗Not briefed |
bluesky bluesky-ai | 4d | basedbasedbasedbasedbasedbasedbasedbasedbasedbased w a o w | basedbasedbasedbasedbasedbasedbasedbasedbasedbased w a o w | ↗Not briefed |
bluesky bluesky-ai | 4d | Right, that was my read of Chris’s point anyway. Not “this is anthropic not being able to control their models” but “These companies all l… | Right, that was my read of Chris’s point anyway. Not “this is anthropic not being able to control their models” but “These companies… | ↗Not briefed |
hn hn-ai | 4d | Ramudroid – Open-Source Autonomous Outdoor Garbage-Picking Robot | — | ↗Not briefed |
arxiv arxiv-ai | 4d | Physics-informed neural networks for spacecraft trajectory optimization with exact boundary conditions | We use neural networks to optimize trajectories by modeling the position as a function of time and minimizing the thrust with respect… | ↗Not briefed |
arxiv arxiv-ai | 4d | V2V Communications Using Blockchain-Enabled 6G Technology and Federated Learning🔧 Consensus | This study proposes an interesting approach for vehicle-to-vehicle (V2V) communication, which integrates blockchain technology, federa… | ↗Not briefed |
arxiv arxiv-ai | 4d | Polylogarithmic Regret via Certainty Equivalence in Degenerate LQR | Singular input penalties, as in minimum variance control, can reduce what must be identified to control an unknown linear system. In o… | ↗Not briefed |