广场

OpenAI 和 Meta 联手,打通了 Agent 商业「最后一公里」

作者|桦林舞王 编辑|靖宇 Agent 能进门了,也能付钱了。 Agent 替用户「消费」的时代,终于到来了。 美国时间 2026 年 11 月 13 日,两家公司两场发布会同天发生。一边是企业级对话式智能体公司 Sierra 联合 Meta 正式发布个 人代理协议 Poppy v1.0,解决助手怎么「进门」的问题 ;另一边是 OpenAI 推出面向个人助手 dots 的 ChatGPT Wallet, 解决助手怎么「付钱」的问题 。 同一天打通进门和结账,本身就足够耐人寻味。更值得留意的,是站在两套方案背后的同一批名字。Visa 和 Mastercard 同时出现在两边的名单里,Shopify、Stripe 这些掌管电商与支付基础设施的公司,则站在 Poppy 的核心伙伴之列。 Agent 的下一场竞争,表面上看是巨头抢入口,底层的暗线却是一场围绕商业通道的提前占位。 01 个人 Agent 协议,PPP Sierra 与 Meta 联合发布个人代理协议 Poppy v1.0|图片来源:Sierra 官方博客 过去,个人助手访问企业网站只能像个隐身访客,商家看不清来者是谁,更不敢把查订单、改地址、订机票或金融预审这类敏感操作交给它。 Meta 和 Sierra 推出的 Poppy 要解决的核心就是信任 ,它让企业在接到请求时,能用密码学手段确认三件事——来敲门的是哪个助手、它代表哪位用户、到底拿了多少权限。 Poppy 协议的核心,是「门牌」、「地基」和「双向授权」。 「门牌」 。接入 Poppy 的企业,会在网站标准路径/.well-known/poppy 下放一份配置文件…

极客公园AI 工具中文

llm 0.37.3

Release: llm 0.37.3 A bug fix release for LLM , mainly to cover a problem with the most recent Homebrew package of the software. Homebrew now ships the latest SQLite, 3.54.1 , which includes a fix that broke LLM's logging code . Short version: in SQLite 3.54.0 the following SQL: select json_object('a', (select json_array(1))) Returns this: {"a":[1]} But in 3.54.1 the same query returns this: {"a":"[1]"} According to D. Richard Hipp : That was a bug fix. Per the documentation, (SELECT json_array(...)) returns SQL text, not JSON. Hence the 3.54.1 behavior is correct and 3.54.0 was a bug. The workaround looks like this: select json_object('a', json((select json_array(1)))) Tags: llm , webassemb…

[Sponsor] Ivy: The To-do List That Ends

Every to-do list runs forever. Ivy runs for a day. Each morning you turn ideas into a plan. Each evening, whatever’s left is an idea again. Nothing is overdue. Tomorrow starts empty. I’m Jürgen Schweizer , founder of Semantical. I spent years building a to-do app and never expected to build another. Then I came across the Ivy Lee Method. We built a prototype and couldn’t stop using it. Beneath the calm: Calendar and Reminders integration, focus sessions, Screen Time shields, and a daily journal. All native: SwiftUI, UIKit, Metal. Ivy is on iPhone , with Mac to follow. Let me know what you think! ★

daringfireball.net其他英文

Engineer Overhead Lines to Withstand Ice, Wind and FreezingRain

This White Paper explains how ice, wind and vegetation load overhead distribution lines at the same time. It shows engineers how to assess the complete line and compare resilience options. What you will learn about: How freezing rain, rime ice and wet snow add weight to conductors and change their shape in the wind. Why a moderate wind on a heavily iced line can govern design, and how galloping and vegetation contact create outages. How a messenger-supported spacer cable separates the electrical function of covered conductors from the main mechanical load path. How monitored severe-weather test data compares tension changes in bare, covered and spacer cable systems. What utility records from…

IEEE Spectrum其他英文

Fragments: October 21

Gergely Orosz is frustrated at those that entered his explicitly “No AI” essay competition and used AI. He goes on to say: When I start to suspect some part is AI about someone I don’t know: how can I tell how much of it is? And how much time and effort do I want to invest to figure out? I fear most people do not realize this last part: that when it takes time and effort from others to figure out what % is AI, and what % is YOU, then I will simply assume it’s probably 100% AI and 0% you So when producing anything that is meant for other humans, if you use even a little AI, you risk the assumption that it is all AI and pointless to pay attention to In general, I agree with him, and react much…

Martin Fowler开发技术英文

土豆姐最新教程,手把手教你用Grok Bot搞定一切工作流自动化

九成的自动化失败,是把流程当成了目标。Grok Bot背后那位土豆姐公开了她的重构范式:先让Bot反向采访你,再用语音把暗默知识倒出来,最后交给Dr Eggbot v2在后台造Bot。 成功订阅 Claude Code,与 Gemini、Codex、Grok 组成王炸 马斯克全家桶好多都成了摆设…..尤其是 Cursor 还是有点拉 这里面我使用频率最高的就是 Grok Bot,也建了不少 Bot 在帮我工作 一直想写一篇详细教程,没得空 刚好, Grok Bot 团队灵魂人物“土豆姐”公开了她重构工作流的工程范式,还有她开发的 Bot 架构师—— Dr Eggbot v2 写得极好 原文很长,我挑核心内容分享一下 我觉得即便不用 Grok Bot ,这些经验也很有价值 一、认知升维:别把“工序”当成你的“终极目标” 土豆姐在开篇就戳中了绝大多数打工人和小老板的软肋: 90% 的自动化失败,是因为你第一步就搞反了——把“流程”当成了“目标” 在她的体系里,任何自动化都只有两件事: 目标: 最终业务成果是什么? 流程: 达到这个结果的中间动作 土豆姐的两个生活化比喻: 1)烤苹果派: 最终那个香气扑鼻的派是 目标 ;买面粉、削苹果、配肉桂粉、控温进烤箱是 流程 。 2)半路换备胎: 车子能重新安全开走是 目标 ;千斤顶支车、扳手卸螺栓、换胎复位是 流程 。 映射到真实商业世界: 每天发一封汇总邮件、填一张 Excel、提一个工单,绝对不是你的目标! 这些全是 TM 的过程。 终极目标只有两个: 能不能为业务多赚钱?能不能帮团队少花钱/省时间? 为什么一定要先把流程扒掉、死磕目标?…

人人都是产品经理产品设计中文