2026-10-30 AI 日报

Orkas is an open-source (MIT), local-first AI team for your desktop: a Commander turns your goal into to-dos and coordinates 16 built-in specialist agents…

Lightweight coding agent that runs in your terminal

Open-source Grok Bot alternative with a virtual machine that bots can use

Open-source AI coworkers that each get a computer of their own: a browser, files and tools, with every action decided before it happens and recorded after.…

Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context…

Turn Claude Code into a full game dev studio — 49 AI agents, 72 workflow skills, and a complete coordination system mirroring real studio hierarchy.

Local-first voice dictation for macOS. Hold a shortcut, speak, and paste into any app. Native Rust, on-device speech recognition, and a Linux beta.

DeepTutor: Lifelong Personalized Tutoring. https://deeptutor.info/.

7.4 billion tokens per month. 34 free LLM providers. 635 free model endpoints. All behind one /v1 endpoint, plus any custom OpenAI-compatible endpoint. Smart…

the runtime your coding agents live on

One-step and few-step text-to-image generation must accurately render the details specified in the prompt despite limited opportunities for refinement.…

A benchmark is not merely a collection of test questions; it naturally induces an empirical measure over difficulty. In this work, we develop this…

Diffusion vision-language models (DVLMs) process their entire input at every denoising step. Since visual tokens make up 89-97% of this input on the…

We present Youtu-Parsing-Omni, an omni-modal parser that converts documents, images, audio, and video into structured records grounded in space and time,…

Whether models trained on different modalities converge to the same representation of the world is central to AI. The Platonic Representation Hypothesis (PRH)…

Recursive self-improvement requires models not only to acquire useful training experience, but also to decide when their current experience has ceased to be…

Pixel-space generative Transformers typically use one token per image patch, so capturing finer detail requires more tokens and makes global attention…

Large language model agents have enabled software development through natural-language interaction, but extending this paradigm to end-to-end game development…

Interactive world models generate visual content conditioned on user actions, yet maintaining spatial consistency over long-horizon exploration in real time…

Scaling supervised training for working agents requires diverse professional tasks that yield reliable expert demonstrations. However, missing or inconsistent…

https://cursor.com/blog/claude-code-extension-security-notic...

ANTHROPIC IS CRACKING DOWN ON VIDEO GAME REWRITES 🚨 Claude Code picked up a new rule around Oct 27, sent from Anthropic’s servers under a flag called…

Sharing one of our latest papers on Multi-Token Prediction (MTP). MTP is now a standard recipe for frontier pre-training but turns out, it can be done much…

Multi-Token Prediction (MTP) is now a common ingredient in language-model pre-training, as it improves downstream performance and enables faster inference…

IIRC there was a guy working on this a year or so ago but it didn't seem to go anywhere. Meanwhile there's a new Gemma-4-Max-Fursona-UwU-9000 every day. Is…

Hey everyone! We’re releasing new DeepSeek-V4.1-1017 Dynamic V3.1 (preview) GGUFs which showcases how Unsloth’s GGUFs retain the most accuracy for its size…

I'm finding myself at odds with the vibes of the community here (and the open-weight/self-hosted community in general) when it comes to LLM's - be it on…

The previous released Qwen3.8-27B-Humanlike-Chat 2.0, a fine-tune that texts humanly and does stuff, is now at 136k downloads for the GGUF repo between v1 and…

I'm the developer of TextFold, a free, Apache 2.0 reader. I added a browser extension so a page you can access can be opened as text with five reading depths.…

What My Project Does Cloakwall redacts personal data (emails, cards, SSNs, IBANs, API keys, medical record and NHS numbers, phones, IPs, dates of birth) from…

the litellm 1.107.0 stealer didnt stop at env vars, going by the writeups it also queried the proxys own postgres for saved provider credentials. those are…

OpenAI shares excerpts from court briefs by creators, academics, accessibility advocates, and legal experts, among others, on how fair use supports AI…

Released the Computer Use tool to general availability (GA) across the Interactions and GenerateContent APIs.

We’ve deprecated fast mode for Claude Opus 4.8, with removal on November 30, 2026. After removal, requests to claude-opus-4-8 with speed: "fast" will return…

CodeRabbit uses GPT-6.1 Sol to investigate security risks across complex codebases. Its security agent helps developers review findings and proposed fixes.

Genie 4: our new world model unlocking unprecedented realism in real-time simulation.

DevFest 2026 runs from October 1 to December 31, bringing together nearly a million developers across more than 800 global events centered on the agentic AI…

OpenAI shares model results on the Kakeya conjecture, Exponential Time Hypothesis, and quantum chromodynamics—and how researchers can verify and build on them.

The weights of Mistral Large 4 (Le Chonk) are now open. A 1T parameter multimodal model, and best open-weight model for coding, trained in Europe.

We’ve launched the Usage and Cost Analytics API in beta with the analytics-usage-cost-2026-08-24 beta header. Its /v1/organizations/analytics/usage and…

The Usage and Cost API is now a legacy API, and so are the usage_report, cost_report, user_usage_report, and user_cost_report endpoints of the Claude…