Agent 智能体

让模型自主规划、调用工具、完成多步任务的技术方向——从 Claude Code、Manus 到各家 Agent 框架与评测基准的全部动态。
全部主题 →
9,446条收录
1,267条精选

精选归档 · 第 61 页

第 1,201–1,220 条 · 共 1,267 条

10月24日10月24日周五

星期五 · 1 条
02:40
Google DeepMind：Blog（RSS）精选
Gemini 2.5 Computer Use 模型发布

Gemini 2.5 Computer Use 模型基于 Gemini 2.5 Pro 构建，专门用于驱动能与用户界面交互的 agent，现已通过 API 以预览版形式提供。

智能体DeepMindGoogle多模态
推荐理由：Google 发布 Gemini 2.5 Computer Use 模型，支持 Agent 自主操作图形界面

10月20日10月20日周一

星期一 · 1 条
00:00
Claude：Blog（网页）精选
Claude Code 发布网页版

Anthropic 推出网页版 Claude Code，以研究预览形式向 Pro、Max 及企业用户开放。用户可直接在浏览器中分配编码任务，无需本地终端，支持并行处理多个 GitHub 仓库的开发工作，并自动创建 PR 和变更摘要。该服务基于云端隔离沙盒运行，具备网络和文件系统限制，同时登陆 iOS 应用支持移动编码。云会话与现有 Claude Code 使用共享速率限制。

智能体Anthropic产品更新编码
另有 2 家信源报道X：Thariq (@trq212)Claude：Blog（网页）
推荐理由：Claude Code搬进浏览器，零门槛并行编程提PR，手机端也能随时跑任务

10月16日10月16日周四

星期四 · 3 条
08:00
Claude Platform：开发者版本说明（RSS）精选
AI 评分 67/100
Claude Platform 发布 Agent Skills 功能，支持 PowerPoint/Excel/Word/PDF 文件处理

Anthropic 在 Claude Platform 推出 Agent Skills（技能）测试版，通过动态加载指令、脚本和资源来扩展 Claude 的能力。首批提供 PowerPoint、Excel、Word 和 PDF 文件的预置技能，用户也可通过 Skills API 上传自定义技能封装领域知识和组织工作流。该功能需启用代码执行工具。

智能体Anthropic产品更新
推荐理由：Anthropic给Claude加上Agent Skills，直接把领域知识打包成可复用技能，构建复杂Agent的门槛降了一个数量级，做自动化的团队该关注了。
00:00
Cognition 模型 / Devin 博客（网页）精选
AI 评分 61/100
Cognition 发布 SWE-grep 与 SWE-grep-mini，用 RL 训练快速并行代码检索模型

Cognition 发布 SWE-grep 和 SWE-grep-mini 模型，专注高度并行的代码上下文检索，检索能力匹敌前沿编码模型但耗时低一个数量级。模型每轮最多发起 8 个并行工具调用、限制 4 轮，通过多轮强化学习训练，奖励为文件与行范围检索的加权 F1；由 SWE-grep 蒸馏并继续 RL 得到 SWE-grep-mini。

智能体模型发布编码
推荐理由：原文给出检索质量分数与推理速度对比数字，并说明多轮 RL 训练方法，读者可评估这种专用检索模型能否迁移到自己的编码工作流。
00:00
Anthropic：Engineering（事故复盘 + 工程实践 · 网页）精选
AI 评分 73/100
为智能体配备现实世界技能：Agent Skills 开放标准

Anthropic 推出了“Agent Skills”开放标准，旨在为通用智能体（如Claude）提供可组合、可扩展且可移植的领域专业知识。一个Skill是一个包含指令、脚本和资源的文件夹，其核心SKILL.md文件采用渐进式披露设计，智能体可根据任务动态加载所需信息，从而最小化上下文占用。例如，PDF技能赋予了Claude直接操作PDF表单的新能力。该框架允许用户通过封装和共享程序性知识来定制智能体，无需为每个用例构建碎片化的定制代理。

智能体AnthropicMCP/工具产品更新
推荐理由：Agent Skills 把「给 Agent 喂知识」从手写 prompt 变成了可复用的文件夹协议，做 Claude Code 或 Agent 产品的人现在就该动手试，这比 MCP 更轻量也更贴近日常开发。

10月13日10月13日周一

星期一 · 1 条
08:00
Answer.AI 官方研发博客（RSS）精选
AI 评分 60/100
Kerem Turgutlu 用 SolveIt 将 Karpathy 的 tokenizer 视频教程做成书章

Answer.AI 的 Kerem Turgutlu 撰文讲解如何用 SolveIt 把 Andrej Karpathy 超两小时的 tokenizers 视频教程转成带可运行代码、超链接和图片的书章。

智能体教程/实践编码
推荐理由：作者完整复盘两阶段对话工作流，把超两小时视频转成可运行代码的书章，方法可直接迁移到任意长视频转写场景。

10月9日10月9日周四

星期四 · 1 条
23:49
Jeff Dean@JeffDean精选
Gemini 与 Google Cloud 推出组织上下文数据及 Agent 系统新功能Today we're rolling out a whole slew of new capabilities for using contextual data relevant to you and your organization and for building and using agent-based systems on top of Gemini and @googlecloud!Learn more at:https://cloud.google.com/blog/products/ai-machine-learning/introducing-gemini-enterpriseUsing Gemini and agents to extract information or accomplish things based on data from your organization makes lots of things easier. For example:(1) "Based on past meeting notes, what are the project's pending action items"译Google 为 Gemini 和 Google Cloud 推出一系列新功能，支持基于组织上下文数据构建 Agent 系统，实现企业数据信息提取与任务自动化，例如根据历史会议记录自动整理项目待办事项。
智能体Google检索增强产品更新
推荐理由：Google推出Gemini企业级Agent构建功能，支持基于组织私有数据的智能任务处理

10月7日10月7日周二

星期二 · 1 条
21:19
Demis Hassabis@demishassabis精选
CodeMender 早期成果发布：可自动修复关键软件漏洞的 AI agentExcited to share early results about CodeMender, our new AI agent that automatically fixes critical software vulnerabilities. AI could be a huge boost for developer productivity and security. Amazing work from the team - congrats!译CodeMender 是一款可自动修复关键软件漏洞的 AI agent，早期成果显示其有望大幅提升开发者生产力与安全性。
智能体DeepMindGoogle产品更新
推荐理由：DeepMind CEO 发布自动修复代码漏洞的 AI agent CodeMender 早期成果

10月2日10月2日周四

星期四 · 1 条
01:09
Andrej Karpathy@karpathy精选
听完 Sutton 的这期播客：有趣且发人深省Finally had a chance to listen through this pod with Sutton, which was interesting and amusing.As background, Sutton's "The Bitter Lesson" has become a bit of biblical text in frontier LLM circles. Researchers routinely talk about and ask whether this or that approach or idea is sufficiently "bitter lesson pilled" (meaning arranged so that it benefits from added computation for free) as a proxy for whether it's going to work or worth even pursuing. The underlying assumption being that LLMs are of course highly "bitter lesson pilled" indeed, just look at LLM scaling laws where if you put compute on the x-axis, number go up and to the right. So it's amusing to see that Sutton, the author of the post, is not so sure that LLMs are "bitter lesson pilled" at all. They are trained on giant datasets of fundamentally human data, which is both 1) human generated and 2) finite. What do you do when you run out? How do you prevent a human bias? So there you have it, bitter lesson pilled LLM researchers taken down by the author of the bitter lesson - rough!In some sense, Dwarkesh (who represents the LLM researchers viewpoint in the pod) and Sutton are slightly speaking past each other because Sutton has a very different architecture in mind and LLMs break a lot of its principles. He calls himself a "classicist" and evokes the original concept of Alan Turing of building a "child machine" - a system capable of learning through experience by dynamically interacting with the world. There's no giant pretraining stage of imitating internet webpages. There's also no supervised finetuning, which he points out is absent in the animal kingdom (it's a subtle point but Sutton is right in the strong sense: animals may of course observe demonstrations, but their actions are not directly forced/"teleoperated" by other animals). Another important note he makes is that even if you just treat pretraining as an initialization of a prior before you finetune with reinforcement learning, Sutton sees the approach as tainted with human bias and fundamentally off course, a bit like when AlphaZero (which has never seen human games of Go) beats AlphaGo (which initializes from them). In Sutton's world view, all there is is an interaction with a world via reinforcement learning, where the reward functions are partially environment specific, but also intrinsically motivated, e.g. "fun", "curiosity", and related to the quality of the prediction in your world model. And the agent is always learning at test time by default, it's not trained once and then deployed thereafter. Overall, Sutton is a lot more interested in what we have common with the animal kingdom instead of what differentiates us. "If we understood a squirrel, we'd be almost done".As for my take...First, I should say that I think Sutton was a great guest for the pod and I like that the AI field maintains entropy of thought and that not everyone is exploiting the next local iteration LLMs. AI has gone through too many discrete transitions of the dominant approach to lose that. And I also think that his criticism of LLMs as not bitter lesson pilled is not inadequate. Frontier LLMs are now highly complex artifacts with a lot of humanness involved at all the stages - the foundation (the pretraining data) is all human text, the finetuning data is human and curated, the reinforcement learning environment mixture is tuned by human engineers. We do not in fact have an actual, single, clean, actually bitter lesson pilled, "turn the crank" algorithm that you could unleash upon the world and see it learn automatically from experience alone.Does such an algorithm even exist? Finding it would of course be a huge AI breakthrough. Two "example proofs" are commonly offered to argue that such a thing is possible. The first example is the success of AlphaZero learning to play Go completely from scratch with no human supervision whatsoever. But the game of Go is clearly such a simple, closed, environment that it's difficult to see the analogous formulation in the messiness of reality. I love Go, but algorithmically and categorically, it is essentially a harder version of tic tac toe. The second example is that of animals, like squirrels. And here, personally, I am also quite hesitant whether it's appropriate because animals arise by a very different computational process and via different constraints than what we have practically available to us in the industry. Animal brains are nowhere near the blank slate they appear to be at birth. First, a lot of what is commonly attributed to "learning" is imo a lot more "maturation". And second, even that which clearly is "learning" and not maturation is a lot more "finetuning" on top of something clearly powerful and preexisting. Example. A baby zebra is born and within a few dozen minutes it can run around the savannah and follow its mother. This is a highly complex sensory-motor task and there is no way in my mind that this is achieved from scratch, tabula rasa. The brains of animals and the billions of parameters within have a powerful initialization encoded in the ATCGs of their DNA, trained via the "outer loop" optimization in the course of evolution. If the baby zebra spasmed its muscles around at random as a reinforcement learning policy would have you do at initialization, it wouldn't get very far at all. Similarly, our AIs now also have neural networks with billions of parameters. These parameters need their own rich, high information density supervision signal. We are not going to re-run evolution. But we do have mountains of internet documents. Yes it is basically supervised learning that is ~absent in the animal kingdom. But it is a way to practically gather enough soft constraints over billions of parameters, to try to get to a point where you're not starting from scratch. TLDR: Pretraining is our crappy evolution. It is one candidate solution to the cold start problem, to be followed later by finetuning on tasks that look more correct, e.g. within the reinforcement learning framework, as state of the art frontier LLM labs now do pervasively.I still think it is worth to be inspired by animals. I think there are multiple powerful ideas that LLM agents are algorithmically missing that can still be adapted from animal intelligence. And I still think the bitter lesson is correct, but I see it more as something platonic to pursue, not necessarily to reach, in our real world and practically speaking. And I say both of these with double digit percent uncertainty and cheer the work of those who disagree, especially those a lot more ambitious bitter lesson wise.So that brings us to where we are. Stated plainly, today's frontier LLM research is not about building animals. It is about summoning ghosts. You can think of ghosts as a fundamentally different kind of point in the space of possible intelligences. They are muddled by humanity. Thoroughly engineered by it. They are these imperfect replicas, a kind of statistical distillation of humanity's documents with some sprinkle on top. They are not platonically bitter lesson pilled, but they are perhaps "practically" bitter lesson pilled, at least compared to a lot of what came before. It seems possibly to me that over time, we can further finetune our ghosts more and more in the direction of animals; That it's not so much a fundamental incompatibility but a matter of initialization in the intelligence space. But it's also quite possible that they diverge even further and end up permanently different, un-animal-like, but still incredibly helpful and properly world-altering. It's possible that ghosts:animals :: planes:birds.Anyway, in summary, overall and actionably, I think this pod is solid "real talk" from Sutton to the frontier LLM researchers, who might be gear shifted a little too much in the exploit mode. Probably we are still not sufficiently bitter lesson pilled and there is a very good chance of more powerful ideas and paradigms, other than exhaustive benchbuilding and benchmaxxing. And animals might be a good source of inspiration. Intrinsic motivation, fun, curiosity, empowerment, multi-agent self-play, culture. Use your imagination.译Sutton（《The Bitter Lesson》作者）在播客中质疑 LLM 并非真正的"苦涩的教训"产物--它们依赖有限的人类数据且充满偏见。他主张 AI 应像动物一样通过 RL 与世界动态交互，而非模仿人类文本。作者认同 LLM 确实充斥人工干预，但认为预训练是应对冷启动的实用"进化替代方案"，纯 RL 在现实世界难以行得通。

Dwarkesh Patel: .@RichardSSutton, father of reinforcement learning, doesn’t think LLMs are bitter-lesson-pilled. My steel man of Richard...

智能体大佬观点数据/训练
推荐理由：Karpathy解读Sutton对LLM的批判，提出「召唤幽灵」vs「构建动物」的深刻比喻，反思AI发展路径

10月1日10月1日周三

星期三 · 2 条
08:00
OpenRouter：Announcements（RSS）精选
AI 评分 58/100
OpenRouter 推出每月100万免费BYOK请求

OpenRouter 向每位客户每月提供100万次“Bring Your Own Key”（BYOK）请求，完全免费。

智能体产品更新部署/工程
推荐理由：这个公告虽然旧了，但 OpenRouter 的 BYOK 免费额度至今仍在，对想低成本折腾模型的个人开发者来说，每月 100 万次请求比很多付费计划都香。
02:52
Anthropic@AnthropicAI精选
Anthropic 工程博客新文：开发者熟知 prompt engineering，但要让 AI agents 发挥最大价值，你需要 context engineeringNew on the Anthropic Engineering Blog: Most developers have heard of prompt engineering. But to get the most out of AI agents, you need context engineering.We explain how it works: https://www.anthropic.com/engineering/effective-context-engineering-for-ai-agents译Anthropic 工程博客发文解释 context engineering。与 prompt engineering 不同，context engineering 通过优化上下文帮助 AI agents 发挥最大效能，文章详解其工作原理。
智能体Anthropic教程/实践
推荐理由：Anthropic提出「上下文工程」方法论，指导开发者优化AI Agent设计

9月30日9月30日周二

星期二 · 3 条
04:20
02:52
Ethan Mollick：One Useful Thing（RSS）精选
真正的 AI Agents 与真正的工作

探讨 AI Agents 在真实工作场景中的定位，指出其核心挑战在于对抗"无限PPT"的形式主义陷阱。强调真正的智能体应当服务于以人为本的实质性工作，而非制造更多文档流程或官僚化产出。

智能体大佬观点
推荐理由：Ethan Mollick 深度解析 AI Agent 在实际工作中的应用与人类价值重塑
00:55
Claude@claudeai精选
推出 Claude Sonnet 4.5--全球最佳编程模型Introducing Claude Sonnet 4.5—the best coding model in the world.It's the strongest model for building complex agents. It's the best model at using computers. And it shows substantial gains on tests of reasoning and math.译Anthropic 发布 Claude Sonnet 4.5，称其为全球最佳编程模型。该模型在构建复杂智能体与计算机使用方面表现最强，推理和数学测试成绩也有显著提升。
智能体Anthropic模型发布编码
推荐理由：Claude Sonnet 4.5 发布，官方称最强编程与 Agent 能力

9月29日9月29日周一

星期一 · 4 条
21:40
08:00
Claude Platform：开发者版本说明（RSS）精选
AI 评分 75/100
Claude Sonnet 4.5 发布，面向复杂智能体与编程任务

Anthropic 发布 Claude Sonnet 4.5，定位为复杂智能体和编程场景的最佳模型，拥有最高综合智能水平。同时推出 Amazon Bedrock 和 Vertex AI 的全球端点定价、新的 model_context_window_exceeded 停止原因、记忆工具（beta）以及上下文编辑功能（beta），后者可在接近 token 限制时自动清理旧工具结果与调用。

智能体Anthropic模型发布编码
推荐理由：这是Anthropic放出的Sonnet 4.5，瞄准编码和复杂Agent场景，Memory工具和上下文编辑也一起开了，对靠API搞产品的开发者算大版本更新，但普通用户不必自己折腾。
00:00
Cognition 模型 / Devin 博客（网页）精选
AI 评分 65/100
Cognition 为 Claude Sonnet 4.5 重构 Devin：新版本快 2 倍、评测提升 12%

Cognition 发布基于 Claude Sonnet 4.5 重构的新版 Devin，速度提升 2 倍，Junior Developer Evals 提升 12%，已在 Agent Preview 上线，旧版 Devin 仍可用。

智能体产品更新编码
推荐理由：来自一线 Agent 团队的适配复盘，给出了 Sonnet 4.5 的具体新行为和对应的工程应对，可迁移到类似智能体架构。
00:00
Anthropic：Engineering（事故复盘 + 工程实践 · 网页）精选
AI 评分 78/100
为AI智能体实施有效的上下文工程

随着AI应用从单次提示转向构建长期运行的智能体，焦点正从“提示工程”演进为“上下文工程”。后者旨在为大型语言模型优化有限的上下文窗口内的全部信息，包括指令、工具、外部数据和对话历史。其核心挑战在于模型存在“注意力预算”限制和“上下文腐化”现象——随着上下文增长，模型回忆信息的准确性会下降。因此，上下文工程要求精心编排高价值信息，以有限的资源最大化产出期望结果，这已成为构建高性能、可操控智能体的关键。

智能体AnthropicMCP/工具教程/实践
推荐理由：Anthropic 亲自下场定义 context engineering 这个新范式，把 prompt engineering 之后的工程方法论讲透了。做 Agent 的人如果还在死磕 prompt，这篇会让你重新审视整个技术栈。

9月25日9月25日周四

星期四 · 1 条
22:28

9月23日9月23日周二

星期二 · 1 条
21:41
Sam Altman：Blog（RSS）精选
充裕的智能

计划打造每周可新增1吉瓦算力的AI基础设施工厂，通过芯片、电力到机器人的全栈创新，支撑治愈癌症、全球个性化教育等宏大应用。项目将主要落地美国，未来数月公布合作伙伴，年底披露融资方案。

智能体OpenAI大佬观点部署/工程
推荐理由：Sam Altman阐述OpenAI基础设施扩张愿景，计划每周新建1吉瓦算力
MCP 与工具调用AI 编码推理能力