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 | dIon: Fragmentation-Based Invariance for Self-Supervised Learning of Tandem Mass Spectra | We introduce a novel invariance for peptide tandem mass spectrometry data, unlocking self-supervised representation learning that impr… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Future Anchored Verification and Online Recovery for World Action Models | World action models (WAMs) have emerged as a promising paradigm for robotic manipulation. They act by first predicting how a task shou… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Browsentic Open-source Claude for Chrome alternative for any agent CLI | A personal agent harness lives in your existing browser, connects to your claude-code (or any other major harness codex or cursor-cli.… | ↗Not briefed |
arxiv arxiv-ai | 1mo | What May an Agent Change About Itself? A Containment Floor for Self-Configuring Agent Runtimes | Many agent runtimes give the agent a tool for editing its own configuration. Some of that configuration grants abilities, such as enab… | ↗Not briefed |
bluesky bluesky-ai | 1mo | talking to someone who told me their friend used Claude to plan their vacation. And that feels inhuman to me. I understand the impetus to u… | talking to someone who told me their friend used Claude to plan their vacation. And that feels inhuman to me. I understand the impetus… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Parakeetogo, a portable ASR binary faster than Photon/torch for CPU | I wanted an easily deployable way of bringing transcription to an underpowered server I have. I used Moondream's excellent parake… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Probabilistic Race and Ethnicity Prediction Using Group-Specific Name Lists | Statistically valid estimation of racial and ethnic disparities often requires inferring the probability that an individual belongs to… | ↗Not briefed |
arxiv arxiv-ai | 1mo | VLA-ZO: Fast Zeroth-Order Adaptation for Vision-Language-Action Models | Adapting vision-language-action (VLA) models to deployment-time distribution shifts is important for reliable robotic operation, but c… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DPNL: A DPLL-based Algorithm for Probabilistic Neurosymbolic Learning | Probabilistic Neurosymbolic Learning (PNL) combines neural predictions with symbolic reasoning, enabling end-to-end learning from fina… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evolving in Thought Space: Training a Small Model at Test Time Unlocks Better Discoveries | Open-ended scientific discovery often requires repeatedly proposing and evaluating candidate solutions. LLM-based systems can support… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Ramp Metering Control via Hybrid State Deep Reinforcement Learning in Partially Observable Connected Vehicle Environments | Freeway on-ramp merges are major sources of congestion, causing significant economic and environmental costs. While Deep Reinforcement… | ↗Not briefed |
hn hn-ai | 1mo | How to keep learning in the age of LLMs | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Mcpward – black-box contract and security testing for MCP servers | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | write an LLM harness to crossbreed past tweet themes to generate new candidate psychoses and tournament rank them | write an LLM harness to crossbreed past tweet themes to generate new candidate psychoses and tournament rank them | ↗Not briefed |
hn hn-ai | 1mo | Gitframes | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Don’t forget who this degenerate is? www.reuters.com/legal/govern... | Don’t forget who this degenerate is? www.reuters.com/legal/govern... | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI is going to kill us but also string theory is bullshit and the rainforest is gay. This guy's a real winner, huh? futurism.com/artificial… | AI is going to kill us but also string theory is bullshit and the rainforest is gay. This guy's a real winner, huh? futurism.com/artif… | ↗Not briefed |
hn hn-ai | 1mo | What Still Gets Paid If AI Changes the Interface? | — | ↗Not briefed |
hn hn-ai | 1mo | Accept 'bad things' in return for benefits of AI, says Sam Altman | — | ↗Not briefed |
hn hn-ai | 1mo | Subql/common 5.8.3 compromised: postinstall stealer in 18k-star SubQuery repo | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think regulation should directly prohibit first-person LLM outputs because we cannot escape the anthropizing tendency of the brain. We sh… | I think regulation should directly prohibit first-person LLM outputs because we cannot escape the anthropizing tendency of the brain.… | ↗Not briefed |
bluesky bluesky-ai | 1mo | -Rebecca Solnit "Two weeks ago an Anthropic safety staffer quit with a public statement. Now one from OpenAI has done the same. " www.the… | -Rebecca Solnit "Two weeks ago an Anthropic safety staffer quit with a public statement. Now one from OpenAI has done the same. " ww… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CentriQ: Calibration-Free Quantization of Diffusion Transformers via Exact Mean Centering | Diffusion transformers (DiTs) achieve state-of-the-art image generation, but their sampling cost limits deployment. Quantizing both we… | ↗Not briefed |
bluesky bluesky-ai | 1mo | when emerging technologies are guided by long-termist utilitarianism, the public gets the deaths and doesn't get the benefits, while the Mu… | when emerging technologies are guided by long-termist utilitarianism, the public gets the deaths and doesn't get the benefits, while t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | OCL-PDE: A Generative Framework for PDE Inverse Problems with Observation-Complementary Latents | Partial differential equation (PDE) inverse problems are often ill-posed, making fine-scale details difficult to recover. We address t… | ↗Not briefed |
hn hn-ai | 1mo | OctLLM – Octrees as an Explicit 3D Language | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Radio Free America: Enforce existing laws against AI companies | Radio Free America: Enforce existing laws against AI companies | ↗Not briefed |
bluesky bluesky-ai | 1mo | David Robinson, un responsable de la sécurité de l’entreprise a annoncé sa démission : «j’ai quitté OpenAI car sa culture d’entreprise est … | David Robinson, un responsable de la sécurité de l’entreprise a annoncé sa démission : «j’ai quitté OpenAI car sa culture d’entreprise… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SimAuthor: Harnessing Foundation Models for Persistent Scientific Simulator Authoring | Foundation models can generate scientific code, but authoring a scientific simulator (an executable program encoding hypotheses about… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Americans are turning on AI. Anthropic wants $2 trillion anyway Plus: Confidence in Big Tech just hit a new record low @mia-silverio.bsky… | Americans are turning on AI. Anthropic wants $2 trillion anyway Plus: Confidence in Big Tech just hit a new record low @mia-silverio… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think that we’ve also lost something by collapsing ‘AI’ and generative AI/LLM together. When I was doing some light ‘digital humanities’ … | I think that we’ve also lost something by collapsing ‘AI’ and generative AI/LLM together. When I was doing some light ‘digital humanit… | ↗Not briefed |
hn hn-ai | 1mo | Pivot – Real-time analytics on Iceberg and Delta, faster than ClickHouse | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Lipschitz Thinking: Ten Years of Certifiable-by-Design Robust Neural Networks | The Lipschitz property of a deep neural network provides a direct measure of its sensitivity to input perturbations and, when explicit… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Shared Stopping Decisions Change Answers in HQQ Cache Quantization | Language-model systems batch questions for throughput, but unrelated questions should not change a target's answer when its input and… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Generative World Models Enable Predictive Control of Laser Melt Pool Dynamics | World models, which learn how environments respond to actions, are emerging as a powerful paradigm for planning through imagined futur… | ↗Not briefed |
arxiv arxiv-ai | 1mo | From Papers to Mechanisms: An Evidence-Grounded Knowledge Substrate for Scientific Language Models | Scientific language models often access literature through untyped text chunks, which fragment the functional and evidential structure… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Web4kit, web pages that pick what to show each visitor | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Constrained Goal-directed Planar Graph Generation with Grammar-based Reinforcement Learning | Planar graphs are central to applications across science and engineering, yet existing generators provide limited support for goal-dir… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RoSA: Rotational Sparse Adaptation for Memory-Efficient Fine-Tuning | Parameter-efficient fine-tuning (PEFT) reduces the cost of adapting foundation models by focusing training on a small parameter subset… | ↗Not briefed |
hn hn-ai | 1mo | Build and fly drones with a C flight controller running as WebAssembly | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Так-то я переусложненный аппарат для копирования нуклеотидных последовательностей, а копирование ответов LLM из окна в окно это для души. | Так-то я переусложненный аппарат для копирования нуклеотидных последовательностей, а копирование ответов LLM из окна в окно это для ду… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Few-Shot Prototype Head Adaptation for On-Device ECG Personalization on PSoC~6 | Wearable and bedside electrocardiogram (ECG) monitors must adapt to patient-specific morphology to maintain arrhythmia detection accur… | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLM-generated grift emails do get funny. I just got an email saying my podcast, Out of the Software Crisis podcasts.apple.com/us/podcast/o.… | LLM-generated grift emails do get funny. I just got an email saying my podcast, Out of the Software Crisis podcasts.apple.com/us/podca… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Artificial Intelligence and the New Science of Culture | Cognition and culture have long been treated as parallel objects of study, joined more by metaphor than by mechanism. We argue they ar… | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI tech bros really said "what if we took the uncanny valley and made it a business model" and now they're mad that people are creeped out … | AI tech bros really said "what if we took the uncanny valley and made it a business model" and now they're mad that people are creeped… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Certification-Enhanced Generalization Bounds | We investigate the use of formal methods to provide tight and sound generalization bounds for learning algorithms. By casting the trad… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI CEO Sam Altman: “We believe the world should accept some bad things happening for the benefits of this technology.” | OpenAI CEO Sam Altman: “We believe the world should accept some bad things happening for the benefits of this technology.” | ↗Not briefed |
rss techmeme | 1mo | Survey: only 11% of ~400 businesses could forecast AI spending; a study finds lower-priced models cost more than higher-priced models on 32% of 6,800+ tasks (Wall Street Journal) | Wall Street Journal: Survey: only 11% of ~400 businesses could forecast AI spending; a study finds lower-priced models cost more than… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 'The Atlas model places Amazon workers at the center of this global fight against the AI billionaires. And it highlights that union organiz… | 'The Atlas model places Amazon workers at the center of this global fight against the AI billionaires. And it highlights that union or… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DP-ES: Differentially Private Evolution Strategies for Prompt Optimization | Token-level differentially private (DP) prompt optimization methods such as DP-OPT can become unstable under tight privacy budgets: on… | ↗Not briefed |