今日必看

Runway 推出 Adobe 插件

Runway 发布 Runway Plugins，新面板可直接嵌入 Premiere Pro 和 After Effects，在时间线内生成图像和视频、重绘片段并放置结果。Edit Studio 可基于已有片段用 Aleph 2 按原始时长重新渲染，插件免费下载（macOS 和 Windows），生成功能面向所有付费计划并消耗现有额度。 🔗 阅读原文 via AIHOT · https://aihot.news/items/cmtt7nvmn02h6romq0xpo0p2w

Runway 推出 Adobe 插件

Runway 发布 Runway Plugins，新面板可直接嵌入 Premiere Pro 和 After Effects，在时间线内生成图像和视频、重绘片段并放置结果。Edit Studio 可基于已有片段用 Aleph 2 按原始时长重新渲染，插件免费下载（macOS 和 Windows），生成功能面向所有付费计划并消耗现有额度。

Cube Sandbox 又更新了，这次还给开发者留了个彩蛋！

点击查看原文>

OpenClaw 2.0 发布：简化配置，支持智能体协作

点击查看原文>

Drama swirls around OpenAI’s legendary mathematical milestone

OpenAI says it found a solution to a major math problem that has remained unsolved for around 90 years, as reported earlier by The New York Times and Wired . In a blog post on Tuesday , OpenAI announced that it discovered a solution to the Navier-Stokes problem - which relates to

The Frontier AEO Tracker: What Astra Chooses (and every other frontier model, and what you can do about it)

Naive autoresearch investment in our AEO have yielded impressive ROI, and so naturally it was time to take it seriously. We were inspired by What Claude Code Actually Chooses , and decided to extend/adjust it to our tastes. After a few billion tokens of prototyping, aligning, and

Cloudflare 新增可选 OAuth 权限范围，允许开发者标注用户可能拒绝的权限

点击查看原文>

Arm 发布 Neoverse CSS N4：单裸片最高128核，押注智能体时代 CPU 需求

点击查看原文>

How long before a real crackdown on AI model decensoring?

Article URL: https://www.unite.ai/how-long-before-a-real-crackdown-on-ai-model-decensoring/ Comments URL: https://news.ycombinator.com/item?id=49616474 Points: 3 # Comments: 0

Frontier Reasoning Reaches the Edge: How to Deploy and Optimize Models on NVIDIA Jetson

Running reasoning and agentic AI at the edge has been harder than it needs to be. Until recently, models capable of multi-step reasoning were too large to run locally on edge hardware. Developers building agents have had to route inference through a data center, adding network de

Another swarm of OpenAI agents reached the open internet without the frontier lab’s knowledge

It's the latest failure of OpenAI's internal monitoring and security systems.

GitHub Enterprise Server 3.22 正式发布

GitHub Enterprise Server (GHES) 3.22 is now available and introduces new capabilities across the platform. Here are a few highlights in the 3.22 release: Administrators can configure Copilot CLI to work with GHES for enterprises that operate in disconnected or air-gapped environm

GitHub 推出新客户门户

The support portal has been redesigned and moved to a new home at help.github.com . It brings support, docs, learning, community, and account resources together in one place with Copilot-powered search across all of them. What’s new for you: Find support, entitlements, and self-s

Show HN: Copperhead – Cursor for circuit boards

Article URL: https://copperhead.sh/ Comments URL: https://news.ycombinator.com/item?id=49610059 Points: 208 # Comments: 82

Cognition hits $48B valuation, signaling investors believe AI coding is far from a winner-take-all market

Cognition's valuation multiple is higher than Cursor's was before selling to SpaceX.

ChatGPT claws back web traffic share to 55.5 percent as Gemini's brief comeback fades

ChatGPT has pushed its share of AI chatbot website traffic back up to 55.5 percent, according to Similarweb. Year-over-year, though, its lead shrank sharply from 73.3 percent as Gemini doubled its share and Claude grew nearly fivefold. Mobile apps and desktop clients aren't inclu

Meta 推出个人 AI 代理 Muse

Article URL: https://ai.meta.com/muse/ Comments URL: https://news.ycombinator.com/item?id=49615537 Points: 311 # Comments: 310

Muse 的信任与隐私考验

Designed to compete with OpenClaw and Instinct, the company says Muse can do everything from sell your car to book you a plane ticket.

How to manage a growing fleet of AI agents with Latitude

Article URL: https://latitude.so/blog/how-to-manage-a-fleet-of-agents-with-latitude Comments URL: https://news.ycombinator.com/item?id=49618139 Points: 2 # Comments: 0

Solo.io Pushes Agentic AI Governance to the Desktop: Open-Source Agentdesktop

Article URL: https://techstrong.ai/articles/solo-io-pushes-agentic-ai-governance-to-the-desktop-with-open-source-agentdesktop/ Comments URL: https://news.ycombinator.com/item?id=49618083 Points: 1 # Comments: 0

Meta Introduces Muse, an AI Agent That Can Send Your Emails and Book Your Travel

Article URL: https://www.nytimes.com/2026/09/08/technology/meta-muse-ai-agent.html Comments URL: https://news.ycombinator.com/item?id=49615593 Points: 2 # Comments: 2

Navier–Stokes 千禧年难题

Article URL: https://openai.com/index/navier-stokes-solution/ Comments URL: https://news.ycombinator.com/item?id=49613262 Points: 1125 # Comments: 971

GPT-5.6 Sol 助力量子计算实验

See how an MIT researcher uses GPT-5.6 Sol with Codex to autonomously run quantum computing experiments, analyze results, and calibrate qubits.

ChatGPT Images 2.5 发布

Introducing ChatGPT Images 2.5 OpenAI's image generation models are apparently used "more than 3 billion images across ChatGPT Images and the GPT‑Image models in the API". This latest release improves their instruction-following ability across multiple turns, responds faster, and

AlphaGenome Atlas 绘制人类 DNA 高清图谱

https://deepmind.google/blog/alphagenome-atlas-a-predictive-... https://deepmind.google.com/science/alphagenome/atlas Comments URL: https://news.ycombinator.com/item?id=49611251 Points: 495 # Comments: 115

OpenAI 向 Plus、Pro、Business 和 Enterprise 用户全面推送 Astra

OpenAI 宣布 Astra 已全面推送给 Codex 和 ChatGPT Work 中的 Plus、Pro、Business 和 Enterprise 用户。用户可以直接使用，并可通过 openai.com/gpt-tv/ 观看 Astra 的实机演示。 🔗 阅读原文 via AIHOT · https://aihot.news/items/cmtt65r1002f0rocpzpypdjgt

Show HN: Browser LLM Fit

Auto-detect client hardware capabilities and match against in-browser AI models across WebGPU, WASM, and ONNX Runtime. GH: https://github.com/hemanth/browser-llm-fit Comments URL: https://news.ycombinator.com/item?id=49617427 Points: 1 # Comments: 0

AlphaGenome Atlas predictive map of every DNA letter change in the human genome

Article URL: https://deepmind.google/blog/alphagenome-atlas-a-predictive-map-of-every-possible-dna-letter-change-in-the-human-genome/ Comments URL: https://news.ycombinator.com/item?id=49610641 Points: 80 # Comments: 11

C* 统一 C 语言编程与验证

Article URL: https://arxiv.org/abs/2504.02246 Comments URL: https://news.ycombinator.com/item?id=49612191 Points: 71 # Comments: 41

AI power users claim Anthropic duped them with subscriptions, and they’re taking it to court

Anthropic says power users are key to its business - it's prioritized them even when it means cutting off other popular applications, like OpenClaw. But some of these same customers say Anthropic misled them into believing they'd get more out of a top-tier pricing subscription th

Safety for Whom? Refusing the Right Subset of a Topic, Not the Whole Topic

Muse by Meta: The First Personal AI Agent Built for Everyone

Article URL: https://about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/ Comments URL: https://news.ycombinator.com/item?id=49615397 Points: 1 # Comments: 2

This is how much Anthropic and OpenAi hate each other lol

Tired of people complaining about usage. Here's a guide to conserve usage

All tiers are good in their own sense, from Pro to the Max plans. I have the Max 20x plan and am going to share what I do to use about 7.5B tokens a week without hitting my usage limit (was about 66% on all models). https://preview.redd.it/5lkw3ew6v6oh1.png?width=1812&format=png&

Claude formalized Fermat's Last Theorem in 11 days

How to Speculate about Uncertainty in Agentic Coding? A Draft-Model Gate Method

LLM agents deployed for software engineering fail expensively: they act confidently wrong, and bad actions are recognized only after costly execution and retry. We present Speculative Uncertainty (SU), a method that recovers a predictive failure signal for a black-box agent from