Claude Dut, Anthropic API 更改，Reursive 本地模式

中文版正在随流水线自动整理，稍后刷新即可；可先阅读下方英文原文。

本文由 AI宝盒自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。

TLDR AI 2026-10-22

Claude Duet 🤖, Anthropic API changes 👨💻, Recursive’s local model 💻

Claude Duet (5 minute read)

Anthropic has unveiled Claude Duet, a proactive AI agent powered by Opus 5.5 that monitors connected apps, prioritizes tasks, and works autonomously while requesting approval before certain actions. The research preview initially targeted US Pro and Max subscribers through a waitlist.

OpenAI models are coming to Google Cloud (2 minute read)

OpenAI's frontier models and Codex agentic workflows are coming to Google Cloud's Model Garden. The integration will give customers native access to these models to power complex agentic workflows at scale. It will pair OpenAI's frontier models with Google Cloud's built-in access controls, centralized data governance, and high-performance infrastructure to support demanding enterprise workloads. OpenAI models will soon be available in Google Antigravity and the Gemini Agent.

Introducing Flint (7 minute read)

Recursive Flint is a mixture-of-experts model built for fast agentic coding on local hardware. It has 40 billion total parameters but activates just 2.5 billion per token. It was built from scratch by a small team of people and agents in under four months. Flint generates more than 125 tokens per second on an M4 Pro MacBook Pro and more than 150 tokens per second on a 64GB NVIDIA RTX Spark at short contexts. It can be used as a daily coding model to complete tasks locally without sending data to a remote server.

If Intelligence Is Getting Cheap, What Remains Expensive? (7 minute read)

Long Lake argues that acquiring established service businesses solves AI deployment barriers through customer trust, expert knowledge, proprietary data, and aligned incentives. Backed by $7.5B, the company has partnered with 40 businesses to turn automation into measurable operating gains.

How does Fable 5.1 hide messages in images? (15 minute read)

Language models can encode messages in images. These messages can be decoded by the same language model even if no information about the encoding method is passed. The act of decoding the hidden messages is obvious if the tool calls of the decoding model are monitored. The encoded messaging is surprisingly hard to spot, and it could be used as a precursor to steganography in images.

Claude Mods: Flipping the Harness Inside Out (11 minute read)

Claude Mods flips the agent harness paradigm to face the human, creating a customizable "Attention Interface" for Claude AI. Mods are TypeScript or JavaScript functions that alter Claude Code's behavior in real-time, allowing users to design how Claude interacts with them. This shift lets users program the human-agent interface and even have Claude build mods, enhancing workflow and decision-making.

OpenAI's PHP Patcher (11 minute read)

OpenAI coordinated with PHP maintainers and security partners to patch critical vulnerabilities discovered using GPT-5.6 Sol and independently by Wiz. The company also released a PHP vulnerability checker and proposed a framework for coordinating responses to AI-discovered security flaws.

OpenClaw creator says he spent $10 million in tokens last month (4 minute read)

Peter Steinberger, the creator of OpenClaw and now a member of technical staff at OpenAI, estimates that he spent around $10 million in tokens running AI workloads last month, all from out of investors' pockets. Steinberger has a virtually unlimited token budget, allowing him to work differently from others in the industry. This article collects some of his answers during a conversation with Cloudflare's CTO at the Cloudflare Connect 2026 conference.

Get the most interesting AI stories and breakthroughs delivered in a free daily email.

Join 1,100,000 readers for one daily email

Auto-fetched and translated by AI Gem Box. All rights belong to the original author.