Y Combinator 首席执行官 Garry Tan 深度重度使用 AI Agent：每年 Token 花费高达 20 万美元

Y Combinator 首席执行官 Garry Tan 此前曾戏称自己患上了“赛博精神病（cyber psychosis）”，有时会因为与自己的 AI Agent 协同工作过于兴奋而几乎难以入眠。在沃尔特·艾萨克森（Walter Isaacson）即将于周二上线的新一期播客 The Builders 中，Tan 为他的这份狂热标出了具体的价格标签。

“我每年在 Token 上的花销大概在 10 万到 20 万美元左右，”他告诉艾萨克森。他将 Agent 用于无数工作项目和编程任务，并在 GitHub 上免费分享了他的一套包含 23 项 Claude Code 技能的工具库——gstack。他表示，这是为了帮助其他人了解 Agent 到底能做到什么。

一方面，他像许多人使用 AI 一样使用自己的 Agent：帮他调研艾萨克森的播客，并为可能被问到的问题做准备；另一方面，他的 Agent 也在深度协助管理他的日常生活。

“它读过了我所有的电子邮件和 Dropbox 里的所有内容。它还读了我所有的 Apple Notes，甚至包括过去 20 年的心理咨询记录，”他说。

他将使用 AI Agent 描述为在“你日常生活中所做的几乎任何事情上获得协助，并且让它以更具自主性的方式去完成。可以把它想象成时刻停留在你肩膀上的吉米尼蟋蟀（Jiminy Cricket）。就像是，对于你几乎想做的任何事情，它都了解你，随时在旁协助。它甚至会主动跑在前面提议：‘这个怎么样？’，或者帮你洞见未来、防患未然。”

作为早期 OpenClaw 尝鲜者兼如今的重度资深用户，Tan 表示他总结出的一条核心心得就是：不要把 AI Agent 当成传统软件对待。他指出，有些人过于担心“可复现性和安全性，以至于他们实际上牺牲掉了与 AI Agent 协作中最美妙的部分——它在本质上可以与人类非常相似。”

例如，Y Combinator 本身也在使用一个正在“重塑 YC 内部运作模式”的 Agent。他说道：“例如，我们现在拥有一个名为 Stray Light 的内部 AI Agent，它兼具某种个性与‘公司大脑’。它实际上掌握着我们曾经支持过的每一位创始人、每一个曾给我们发过邮件的人，以及所有项目申请材料的完整上下文。”

作为掌舵硅谷知名创业加速器的领军人物，Tan 显然是一位狂热的技术拥趸。他认为，只要解决一个关键瓶颈，个人 AI Agent 终将成为主流：那就是成本必须降下来。

“它现在太贵了，”他坦言，并补充说自己很清楚每年在 AI Agent 上投入的资金是“一种极其疯狂的特权，绝非每个人都能承受得起”。

但由科技巨头掌控的免费 Agent 或许也不是终极答案。在谈及他希望科技行业构建的未来，以及 Meta 等社交媒体巨头对大众产生的影响时，他表示：“那是一家控制着算法的庞大企业，而在某种程度上，我们都希望他们不要拥有这种控制权，”他说，“因此，你可以设想一个由你自己全权掌控、由你自己来编程的个人 AI。”

Y Combinator CEO Garry Tan has, in the past, jokingly admitted that he’s suffered from “cyber psychosis” and sometimes barely sleeps because he’s so excited to be working with his AI agents. On a new episode of Walter Isaacson’s podcast The Builders, set to publish on Tuesday, Tan put a price tag on his passion.

“I spend about maybe $100[,000] to $200,000 a year on tokens,” he told Isaacson. He uses agents for countless work projects and coding tasks and has freely shared his set of 23 Claude Code skills, called gstack, on GitHub — an effort to help teach others what agents can do, he said.

On the one hand, he uses his agent the way many people use AI: to help him research Isaacson’s podcast and prep for the kinds of questions he might be asked. On the other, his agent also helps him run his life.

“It’s read all of my email and all of my Dropbox. It’s read, like, all of my Apple Notes, even my therapy notes going back 20 years,” he said.

He describes using AI agents as getting help with “almost anything that you do in your day-to-day life, and for it to do it somewhat more autonomously. Think of it as a Jiminy Cricket on your shoulder at all times. Like, almost anything you want to do, it knows you, and it’s there to help. And it will actually sort of run ahead and say, ‘What about this?’ Or it’ll help you see around corners.”

As an early OpenClaw adopter who is now a power user, Tan says one thing he’s learned is not to treat AI agents like traditional software. Some people are so worried about “repeatability and safety that they actually trade off all of the beautiful parts of working with an AI agent, which basically can be so similar to a human,” he said.

For instance, Y Combinator is also using an agent that is “rebuilding the way we operate at YC,” he said. “We now have an internal AI agent, for instance, named Stray Light that kind of has a personality, and then it has a company brain. Like it actually has the total context of every founder we’ve ever backed, anyone who’s ever emailed us, like all the applications.”

Tan, who is, naturally, an avid lover of technology (as anyone running Silicon Valley’s famed startup accelerator would be), believes personal AI agents will eventually become mainstream, if one major problem is solved: Costs have got to come down.

“It’s too expensive right now,” he said, adding that he knows the amount of money he spends on AI agents is “a crazy privilege that not everyone can do.”

But freebie agents controlled by Big Tech maybe aren’t the answer, either. When discussing the future he’d like tech to build, and how Meta as a social media company has impacted us, he said, “This is a huge company that controls the algorithm, and we all wish that they didn’t at some level,” he said. “So you could imagine a personal AI that you control, that you program.”