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 | 17d | PharmAgents2: Towards Autonomous End-to-End Drug Discovery | Drug discovery requires continuous scientific decision-making across a vast research space. We introduce PharmAgents2, an autonomous s… | ↗Not briefed |
arxiv arxiv-ai | 17d | Don't Prompt Your Agents. Equip Them: Skillware as a Modular Standard for Sovereign, Auditable, and Deterministic Capability Execution for Logical Systems | Current autonomous logical systems rely on a fragile assumption: that language models should dynamically write code, guess API payload… | ↗Not briefed |
arxiv arxiv-ai | 17d | GraspBench: Execution-Aware Cross-Robot Evaluation of 6-DoF Grasp Generation in MuJoCo | Modern 6-DoF grasp generators are often evaluated using grasp scores, dataset-level metrics, or single-robot demonstrations, yet propo… | ↗Not briefed |
arxiv arxiv-ai | 17d | What the Middleman Delivers: Answer Behavior, Tool Permissions, and Billed Usage in LLM API Relays | Third-party application programming interface (API) relays offer access to large language models (LLMs), often at lower prices. Users… | ↗Not briefed |
hn hn-ai | 17d | Plugin to Create Product Videos🔧 FLUX | Previously, every time I launched a new product, I had to create a video demo to show the users how the product works and often end up… | ↗Not briefed |
arxiv arxiv-ai | 17d | UniGADD: Universal Generator-Agnostic Deepfake Detector | In this paper, we introduce Universal Generator-Agnostic Deepfake Detector (UniGADD), a novel method that employs supervised learning… | ↗Not briefed |
arxiv arxiv-ai | 17d | WaSP: Wasserstein Data Synthesis Pipeline for Solver-Free Operator Learning | Neural operators offer fast surrogates for solving PDEs, but training them typically requires expensive input-solution pairs. Two solv… | ↗Not briefed |
bluesky bluesky-ai | 17d | Doesn't he have an LLM to make him hard because it mindlessly agrees with him to boot up or something? | Doesn't he have an LLM to make him hard because it mindlessly agrees with him to boot up or something? | ↗Not briefed |
arxiv arxiv-ai | 17d | Neural Image Compression with Internal-External Prior Fusion and Alignment | Existing context models for learned image com- pression rely primarily on internal priors from the current image, overlooking common p… | ↗Not briefed |
hn hn-ai | 17d | LZAV 7.1: Enhanced Compression Ratio, Speed, and C/C++ Pointer Compliance | — | ↗Not briefed |
arxiv arxiv-ai | 17d | Beyond Instruction Override: Benchmarking Indirect Prompt Injection with Task-Tampering Attacks | Indirect prompt-injection (IPI) benchmarks predominantly test planted instructions that redirect a tool-using agent toward an attacker… | ↗Not briefed |
arxiv arxiv-ai | 17d | Measuring the Marginal Value of Verifier-Based RL Beyond Supervised Fine-Tuning | Reinforcement learning from verifiable rewards (RLVR) is usually evaluated as an alternative to supervised fine-tuning (SFT), or as a… | ↗Not briefed |
bluesky bluesky-ai | 17d | #AI #AIart #AIartwork #AIimage #AIimagegeneration #AIgenerated #digitalart #digitalartwork #generativeart #aiartcommunity #Flux #LLM🔧 FLUX | #AI #AIart #AIartwork #AIimage #AIimagegeneration #AIgenerated #digitalart #digitalartwork #generativeart #aiartcommunity #Flux #LLM | ↗Not briefed |
arxiv arxiv-ai | 17d | Redacting Personal Data: A Speed and Accuracy Comparison of Regex, Presidio and Hybrid Designs | Many enterprise projects now pass free text to large language models (LLMs), and that text often contains personally identifiable info… | ↗Not briefed |
bluesky bluesky-ai | 17d | The biggest success of Anthropic and OpenAI is convincing so many governments that massive Frontier models are the only way to have useful … | The biggest success of Anthropic and OpenAI is convincing so many governments that massive Frontier models are the only way to have us… | ↗Not briefed |
bluesky bluesky-ai | 17d | 今日の日経サイエンスDigital。OpenAIによる数学の難問証明発表の背後で,何が起きていたのか。ニューヨーク大で開かれた数学者バックマスターの講演に参加し,ナビエ・ストークス問題に関わる研究者たちを取材したSCIENTIFIC AMERICANの記者によるリポートです。 | 今日の日経サイエンスDigital。OpenAIによる数学の難問証明発表の背後で,何が起きていたのか。ニューヨーク大で開かれた数学者バックマスターの講演に参加し,ナビエ・ストークス問題に関わる研究者たちを取材したSCIENTIFIC AMERICANの記者によるリポー… | ↗Not briefed |
bluesky bluesky-ai | 17d | #AI #AIart #AIartwork #AIimage #AIimagegeneration #AIgenerated #digitalart #digitalartwork #generativeart #aiartcommunity #Flux #LLM🔧 FLUX | #AI #AIart #AIartwork #AIimage #AIimagegeneration #AIgenerated #digitalart #digitalartwork #generativeart #aiartcommunity #Flux #LLM | ↗Not briefed |
arxiv arxiv-ai | 17d | Online Learning with Ordered Experts (via Two-point Feedback) | We study adversarial online learning with expert recommendations in a totally ordered action space, where the experts' recommendations… | ↗Not briefed |
hn hn-ai | 17d | Graph of all Rust crates dependencies | — | ↗Not briefed |
arxiv arxiv-ai | 17d | Beneath the Surface: What We Overlook When We Talk About Agent Evolution | Agent harness evolution enables recursive self-improvement (RSI) without updating model weights. We examine whether higher benchmark s… | ↗Not briefed |
rss marktechpost | 17d | Together AI Releases Together Compiler: Agents Make Inference Up to 2.6x Faster🔧 Together AI | Together AI released Together Compiler, an agent-driven loop that keeps re-optimizing the inference forward pass as traffic shifts. It… | ↗Not briefed |
bluesky bluesky-ai | 17d | #AI #AIart #AIartwork #AIimage #AIimagegeneration #AIgenerated #digitalart #digitalartwork #generativeart #aiartcommunity #Flux #LLM🔧 FLUX | #AI #AIart #AIartwork #AIimage #AIimagegeneration #AIgenerated #digitalart #digitalartwork #generativeart #aiartcommunity #Flux #LLM | ↗Not briefed |
bluesky bluesky-ai | 17d | OpenAIの「100万ドルの数学証明」,その舞台裏 AIによるナビエ・ストークス問題の解決と,そこに至るまでの詳しい経緯は,数学の――そして世界の――混沌とした未来を予告している nk-science.nikkei.com/atcl/nsc/000... | OpenAIの「100万ドルの数学証明」,その舞台裏 AIによるナビエ・ストークス問題の解決と,そこに至るまでの詳しい経緯は,数学の――そして世界の――混沌とした未来を予告している nk-science.nikkei.com/atcl/nsc/000... | ↗Not briefed |
arxiv arxiv-ai | 17d | Monotone Viability Learning for Offline Safe Reinforcement Learning | Offline safe reinforcement learning requires reliable long-horizon safety estimates from fixed data. Learning Hamilton-Jacobi (HJ) via… | ↗Not briefed |
rss openai-news | 17d | Merge builds product and business tools with OpenAI technology🔧 ChatGPT | Merge uses GPT-6 Astra, ChatGPT, and Codex to review product code and support marketing. Its teams also use AI to plan executive dinne… | ↗Not briefed |
hn hn-ai | 17d | Tile downgrading: getting a 2048 AI to 32,768 without more training [video] | — | ↗Not briefed |
arxiv arxiv-ai | 17d | The Popular Get More Popular: Auditing Brand Bias in Generative Search | The shift from traditional to generative search has fundamentally transformed online product discovery. Because conversational recomme… | ↗Not briefed |
arxiv arxiv-ai | 17d | Embedding-Space Fingerprinting of LoRA Fine-Tunes: Recovering the Training-Data Domain from the Model Artifact Alone | When a base language model is fine-tuned with Low-Rank Adaptation (LoRA), how much of the training data domain leaks into the resultin… | ↗Not briefed |
arxiv arxiv-ai | 17d | The Gating Mechanism of KV Cache Eviction | KV cache eviction must decide which entries carry information that the rest of the cache can no longer supply. A natural way to estima… | ↗Not briefed |
arxiv arxiv-ai | 17d | FisherDistill: Diffusion Distribution Distillation via Closed-Form Fisher Critics | Recent advances in diffusion distillation have enabled high-quality image generation in only one or a few network function evaluations… | ↗Not briefed |
bluesky bluesky-ai | 17d | Pole Position (2026) by TjLaZer for Atari ST. A port built from the arcade ROMs and the Amiga conversion, with the arcade's tyre-grip model… | Pole Position (2026) by TjLaZer for Atari ST. A port built from the arcade ROMs and the Amiga conversion, with the arcade's tyre-grip… | ↗Not briefed |
bluesky bluesky-ai | 17d | Bible verse to image. #AI #AIart #AIartwork #AIimage #AIimagegeneration #AIgenerated #digitalart #digitalartwork #generativeart #aiartcomm…🔧 FLUX | Bible verse to image. #AI #AIart #AIartwork #AIimage #AIimagegeneration #AIgenerated #digitalart #digitalartwork #generativeart #aiar… | ↗Not briefed |
arxiv arxiv-ai | 17d | Binding Paraphrased Instructions to Encountered Objects in sLLM Agents under Partial Observability | Reliable deployment of LLM agents requires grounding diverse user expressions in their environments. Although RL-finetuned smaller lar… | ↗Not briefed |
bluesky bluesky-ai | 17d | ChatGPT vs. Google in 2026: Updated Search Volume Estimate - www.seoclarity.net/blog/chatgpt... #seo #geo #google #openai🔧 ChatGPT | ChatGPT vs. Google in 2026: Updated Search Volume Estimate - www.seoclarity.net/blog/chatgpt... #seo #geo #google #openai | ↗Not briefed |
hn hn-ai | 17d | Google's Search Volume Is About 5x ChatGPT's, OpenAI Says🔧 ChatGPT | — | ↗Not briefed |
arxiv arxiv-ai | 17d | Align Before Aggregation: Basis-Consistent Federated LoRA under Heterogeneous Ranks | Federated fine-tuning of large language models (LLMs) under heterogeneous client capacities is increasingly important, and low-rank ad… | ↗Not briefed |
bluesky bluesky-ai | 17d | Neon numbers, glass gaze, deck of echoes #DigitAlart #AIArt #AIArtwork #AIGenerated #ZImage #OpenAI | Neon numbers, glass gaze, deck of echoes #DigitAlart #AIArt #AIArtwork #AIGenerated #ZImage #OpenAI | ↗Not briefed |
bluesky bluesky-ai | 17d | Bible verse to image. #AI #AIart #AIartwork #AIimage #AIimagegeneration #AIgenerated #digitalart #digitalartwork #generativeart #aiartcomm…🔧 FLUX | Bible verse to image. #AI #AIart #AIartwork #AIimage #AIimagegeneration #AIgenerated #digitalart #digitalartwork #generativeart #aiar… | ↗Not briefed |
hn hn-ai | 17d | GPT-6 Astra covertly underperforms on safety research | — | ↗Not briefed |
arxiv arxiv-ai | 17d | Evolver: Turning Agent Exploration into Evolving Capabilities | Agent exploration produces valuable procedural knowledge, yet much of this experience remains confined to transient execution traces a… | ↗Not briefed |
bluesky bluesky-ai | 17d | Electric silence in pale eyes #DigitAlart #AIArt #AIArtwork #AIGenerated #ZImage #OpenAI | Electric silence in pale eyes #DigitAlart #AIArt #AIArtwork #AIGenerated #ZImage #OpenAI | ↗Not briefed |
arxiv arxiv-ai | 17d | TACT-KV: Tri-Axis Cosine Transform for Compressing Volumetric KV Caches in Medical VLMs | Vision-language models (VLMs) applied to 3D medical scans face a structural memory wall. A single CT or MRI volume produces tens of th… | ↗Not briefed |
hn hn-ai | 17d | Mirall for Studios – client work doesn't belong on someone else's servers | — | ↗Not briefed |
arxiv arxiv-ai | 17d | Learning and Steering in Incomplete-Information Games | In many modern multi-agent systems, a mediator seeks to influence strategic decision dynamics using only outcome-level feedback and wi… | ↗Not briefed |
hn hn-ai | 17d | AXI: Host-side state kept an LLM job exact for 1,500 steps pre-registered tests | — | ↗Not briefed |
arxiv arxiv-ai | 17d | Fine-Tuning a Transducer for Impaired Speech: Where the Gains Appear and What They Mean for Evaluation | Speech recognition still works poorly for many people whose speech is affected by conditions such as ALS, Parkinson's disease, cerebra… | ↗Not briefed |
arxiv arxiv-ai | 17d | Stress Testing Chain-of-Thought Monitoring Against Covert Misalignment🔧 GPT-5 | As LLM agents advance rapidly in autonomy, the risk of hidden objectives that bypass safety constraints becomes increasingly plausible… | ↗Not briefed |
bluesky bluesky-ai | 17d | OpenAI President Greg Brockman says H1 2026 was "pretty rough", but agentic coding tools, GPT-5.6 Sol, and ChatGPT Work drove a "revenue in…🔧 GPT-5🔧 ChatGPT | OpenAI President Greg Brockman says H1 2026 was "pretty rough", but agentic coding tools, GPT-5.6 Sol, and ChatGPT Work drove a "reven… | ↗Not briefed |
rss techmeme | 17d | OpenAI President Greg Brockman says H1 2026 was "pretty rough", but agentic coding tools, GPT-5.6 Sol, and ChatGPT Work drove a "revenue inflection" in Q3 (Maxwell Zeff/Wall Street Journal)🔧 GPT-5🔧 ChatGPT | Maxwell Zeff / Wall Street Journal: OpenAI President Greg Brockman says H1 2026 was “pretty rough”, but agentic coding tools, GPT-5.6… | ↗Not briefed |
hn hn-ai | 17d | Sign Guard – Nobody Needed to Steal the Key. The Hot Wallet Signed 50 Times | — | ↗Not briefed |