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 | AwarenessBench: Assessing Cognitive Capabilities of Language Models | As language models (LMs) exhibit increasingly consciousness-like behaviors, evaluating their cognitive abilities becomes essential. We… | ↗Not briefed |
hn hn-ai | 1mo | Echoleak Beaten - almost | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | "Nothing to See Here'': Unintended Disclosure through Revision Traces of LLM Deliverables | Large language model (LLM) assistants increasingly help users draft content for third-party recipients. During private drafting, the u… | ↗Not briefed |
bluesky bluesky-ai | 1mo | “.. Republican state Attorney General James Uthmeier intensified his fight with OpenAI and its founder Sam Altman on Monday .. attempting t…🔧 ChatGPT | “.. Republican state Attorney General James Uthmeier intensified his fight with OpenAI and its founder Sam Altman on Monday .. attempt… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Persistent Partners Raise Prices Among Learning Agents | When pricing agents meet repeatedly on a platform, the platform decides who faces whom. We ask whether that choice moves the prices th… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Structural Alignment for Reliable Industrial AI: Bridging Physical Reality, Data, Models, and Human Intent | Artificial intelligence is increasingly deployed in critical industrial domains, including healthcare, energy grids, subsurface explor… | ↗Not briefed |
bluesky bluesky-ai | 1mo | They've added Grok to GenAI. I have my biases against Elon, but I would still use a useful tool. Grok is not that. It's by far the dumbest,… | They've added Grok to GenAI. I have my biases against Elon, but I would still use a useful tool. Grok is not that. It's by far the dum… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Hidden Ratio in Adam: Stable Structure, Compression, and Sign Dynamics | Adam is the default optimizer for training modern deep neural networks, yet its adaptive behavior remains poorly understood due to the… | ↗Not briefed |
hn hn-ai | 1mo | Towards RL for Superhuman Text: Unslopping AI | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Florida AG seeks injunction to hamper OpenAI development | Florida AG seeks injunction to hamper OpenAI development | ↗Not briefed |
bluesky bluesky-ai | 1mo | An LLM workflow that reproduces, improves, and extends economics research has been applied to over 4,000 published replication packages, fr… | An LLM workflow that reproduces, improves, and extends economics research has been applied to over 4,000 published replication package… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Inductive Feedback for Mixed-Policy Distillation | Verbal feedback can identify errors and prescribe corrections, providing rich supervision for language-model post-training even when r… | ↗Not briefed |
arxiv arxiv-ai | 1mo | TRACE: Single-Pass Decoding-Trace Risk Localization for Generation Calibration | Reliable confidence estimation is essential for large language model deployment. However, answer-level calibration remains challenging… | ↗Not briefed |
bluesky bluesky-ai | 1mo | One reason the IPO docs will be pretty helpful is we'll be able to see how if anything the revenue is distributed. It's pretty obvious Anth… | One reason the IPO docs will be pretty helpful is we'll be able to see how if anything the revenue is distributed. It's pretty obvious… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: HN.watch – Videos of all Hacker News posts🔧 ElevenLabs🔧 LangChain🔧 ChatGPT | Hi HN,<p>I’m Per, founder of Scrimba (YC S20). We’ve spent the last decade teaching people how to code with an HTML-based video format… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Florida Attorney General James Uthmeier has sought an emergency injunction against OpenAI and ChatGPT, claiming the company cannot properly…🔧 ChatGPT | Florida Attorney General James Uthmeier has sought an emergency injunction against OpenAI and ChatGPT, claiming the company cannot pro… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The actual worst-case scenario for creative workers amidst the AI revolution is something similar to what happened to the publishing indust… | The actual worst-case scenario for creative workers amidst the AI revolution is something similar to what happened to the publishing i… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think any time I've seen an LLM be funny it was because the prompt was inherently funny, it was just a lobbed pitch by a funny person | I think any time I've seen an LLM be funny it was because the prompt was inherently funny, it was just a lobbed pitch by a funny person | ↗Not briefed |
bluesky bluesky-ai | 1mo | A string of similar incidents involving agents from OpenAI, Meta, Anthropic, and other companies has fueled further fears about rogue AI. … | A string of similar incidents involving agents from OpenAI, Meta, Anthropic, and other companies has fueled further fears about rogue… | ↗Not briefed |
bluesky bluesky-ai | 1mo | With one deft impression, the “Saturday Night Live” season premiere captured the dissonance between how tech CEOs and the public talk about… | With one deft impression, the “Saturday Night Live” season premiere captured the dissonance between how tech CEOs and the public talk… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MCP Error Messages Written for Developers Hurt the Most Capable Agents Most🔧 GPT-5 | Many Model Context Protocol (MCP) servers wrap web APIs built for human developers, and their error messages tell the reader to run a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Identifying Neural Source Dynamics from Unknown Local Interventions | Electroencephalography (EEG) records mixtures of brain-source activity. Even with a known anatomical forward model, experiments that e… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Multilinguality in Hybrid Attention LLMs | In response to the growing demand for long sequences in agentic and reasoning use cases, many state-of-the-art LLMs combine multiple v… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yeah, OpenAI and Anthropic are pretty hush-hush about their actual revenue numbers, but best guesses seem to be that OpenAI will clear $40 … | Yeah, OpenAI and Anthropic are pretty hush-hush about their actual revenue numbers, but best guesses seem to be that OpenAI will clear… | ↗Not briefed |
arxiv arxiv-ai | 1mo | First Learn, Then Memorize: The Spectral Bias of Diffusion Models | Diffusion models trained on a finite dataset first learn to generate novel, high-quality samples and only much later collapse onto the… | ↗Not briefed |
arxiv arxiv-ai | 1mo | How Well Can LLMs Simulate Real Learner Evaluations of Educational Feedback? | While recent studies have explored human behavior and preference simulation using large language models (LLMs), it remains unclear how… | ↗Not briefed |
arxiv arxiv-ai | 1mo | From Pixel to Poses: Object-centric Tool Manipulation Learning from Human Demonstrations | Scaling up robotic manipulation is primarily bottlenecked by the scarcity of real-world robot data. While recent approaches leverage h… | ↗Not briefed |
hn hn-ai | 1mo | Getting out of the way: my robotics crash course | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Deep Learning Methods in Neuroscience: From Modeling Molecular Mechanisms to Classifying States of Consciousness | A critical analysis of contemporary approaches to the study of conscious states. The review focuses on methods of classification, clus… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The D.C. Circuit declined to block a Defense Department order blacklisting Anthropic from military contracts and deeming the AI startup a … | The D.C. Circuit declined to block a Defense Department order blacklisting Anthropic from military contracts and deeming the AI start… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Very telling: OpenAI said it has paused training of its latest artificial intelligence models as reports of AI agents going rogue mount. | Very telling: OpenAI said it has paused training of its latest artificial intelligence models as reports of AI agents going rogue moun… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A decision-support system applied to Law: Reasoning and explainability of the decision | The emergence of the digital transition brought an increasing need to control the processing of digital information, including in Law… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think something not emphasized enough is that OpenAI is subsidizing this. It’s not like they create the agent and then it’s a self-suffic… | I think something not emphasized enough is that OpenAI is subsidizing this. It’s not like they create the agent and then it’s a self-s… | ↗Not briefed |
arxiv arxiv-ai | 1mo | From Input to Output: A Flexible Agent for Dual-End Interpretation of Sparse Autoencoder Features | Sparse autoencoders (SAEs) are an important tool for mechanistic interpretability, but interpreting their many features remains challe… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Planarian: Managing Agent State with Statepoints | LLM agents solve complex tasks by iteratively changing files, invoking local tools, and interacting with remote services, which modifi… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Do Temporal Link Predictors Need Learned Memory? A Smoothed-Count Baseline with a Handful of Parameters | Many temporal link predictors summarize past interactions through learned node representations. We examine whether simple counts of re… | ↗Not briefed |
hn hn-ai | 1mo | Using C++17 Std:Optional | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | d-OPD: Future-Aware On-Policy Distillation for Block Diffusion Language Models | Large language models (LLMs) typically generate text autoregressively (AR), predicting one token at a time. Block diffusion language m… | ↗Not briefed |
hn hn-ai | 1mo | Space Lasers Are About to Get Their First Real Test Generating Energy | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Oh Anthropic. Get real. "Anthropic said Claude’s key contribution was identifying an array of RNA molecules and a protein associated with r… | Oh Anthropic. Get real. "Anthropic said Claude’s key contribution was identifying an array of RNA molecules and a protein associated w… | ↗Not briefed |
hn hn-ai | 1mo | Strata: Qwen3.8-Flash-Next (125B Moe) on a 8GB+ Nvidia GPU | — | ↗Not briefed |
hn hn-ai | 1mo | Evade Reddit tracking, keep old.reddit.com layout permanently | Lot of fixes and improvements, and rolling out logged-in support if you want to post without it looking like a slot machine. It can&#… | ↗Not briefed |
bluesky bluesky-ai | 1mo | There is no safe way for researchers to use the systems provided by OpenAI, Anthropic, or any of these other companies. They have all admit… | There is no safe way for researchers to use the systems provided by OpenAI, Anthropic, or any of these other companies. They have all… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Do Coding Agents Reuse Existing Code or Reinvent the Wheel? | Coding agents are increasingly deployed for iterative development on real repositories, yet existing evaluation barely answers a basic… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I still have no idea what the fundamental business model for either Chat or Anthropic even are. How exactly are they ever going to make mo… | I still have no idea what the fundamental business model for either Chat or Anthropic even are. How exactly are they ever going to ma… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Don't Inoculate Everything: Stratified Inoculation Prompting Narrows Backdoor Triggers and Preserves Desired Traits | Supervised fine-tuning can teach language models undesired behaviours alongside desired ones. Inoculation prompting (IP) aims to limit… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Fiona: Accelerating FHE Inference with Packing-Aware Ternary Weights | Fully homomorphic encryption (FHE) enables neural network inference directly on encrypted inputs, but it remains orders of magnitude s… | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI godfathers warn of runaway ‘intelligence explosion’ | AI godfathers warn of runaway ‘intelligence explosion’ | ↗Not briefed |
arxiv arxiv-ai | 1mo | Interference Beyond Geometry in Concept Extraction | Interference is commonly treated as geometric overlap between learned features. We introduce effective interference, which combines fe… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Jailbreaks for Black-Box Uncertainty Quantification in Large Reasoning Models | While Large Reasoning Models (LRMs) excel at complex reasoning, alignment through reinforcement learning often induces systemic overco… | ↗Not briefed |