conv.

All stories
AIQuiet 5d · day 6

Garry Tan spends up to $200K yearly on AI agent tokens at Y Combinator

Y Combinator's CEO reveals his heavy reliance on AI agents for work and personal tasks, including access to decades of personal data.

What to know

  • Y Combinator CEO Garry Tan spends $100K–$200K annually on AI agent tokens and grants agents access to 20 years of personal data including therapy notes.
  • Y Combinator operates an internal AI agent called Stray Light with full institutional context on founders and applications, redesigning company operations.
  • Tan acknowledges AI agent costs are prohibitively expensive and advocates for user-controlled personal AI rather than Big Tech–controlled systems.
  • Early reactions online range from skepticism to alarm about the implications of such extensive personal data access and AI autonomy.
some voices

Tan's level of AI agent use and personal data access is troubling or dystopian.

  • “Gary Tan has lost his mind. this sounds fucking awful.”

    peter@thepit.social · Mastodon ↗

“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”

Garry Tan, Y Combinator CEO · TechCrunch / Walter Isaacson's podcast The Builders ↗

Garry Tan Y Combinator CEOWalter Isaacson Podcast host

Garry Tan spends up to $200K yearly on AI agent tokens at Y Combinator
techcrunch.com

How it unfolded 1 development · click the chart to see its coverage articlesposts

Peak 9 pieces in two hours at Nov 4, 11 AM; 38 pieces over 6 days (3 articles · 14 posts · 21 comments) Nov 2, 7 PM — 6 pieces · 2 articles · 4 posts — Mastodon 5, Newswires 1Nov 2, 9 PM — quietNov 2, 11 PM — 1 piece · 1 post — Hacker News 1Nov 3, 1 AM — 1 piece · 1 post — Mastodon 1Nov 3, 3 AM — quietNov 3, 5 AM — quietNov 3, 7 AM — quietNov 3, 9 AM — quietNov 3, 11 AM — quietNov 3, 1 PM — 2 pieces · 2 posts — Bluesky 2Nov 3, 3 PM — quietNov 3, 5 PM — quietNov 3, 7 PM — 1 piece · 1 post — Bluesky 1Nov 3, 9 PM — quietNov 3, 11 PM — quietNov 4, 1 AM — quietNov 4, 3 AM — quietNov 4, 5 AM — quietNov 4, 7 AM — quietNov 4, 9 AM — 2 pieces · 1 article · 1 post — Hacker News 1, Newswires 1Nov 4, 11 AM — 9 pieces · 9 comments — Hacker News 9Nov 4, 1 PM — quietNov 4, 3 PM — quietNov 4, 5 PM — quietNov 4, 7 PM — 1 piece · 1 post — Bluesky 1Nov 4, 9 PM — 4 pieces · 2 posts · 2 comments — Hacker News 2, Bluesky 1, Mastodon 1Nov 4, 11 PM — quietNov 5, 1 AM — 2 pieces · 2 comments — Hacker News 2Nov 5, 3 AM — 2 pieces · 2 comments — Hacker News 2Nov 5, 5 AM — quietNov 5, 7 AM — 2 pieces · 2 comments — Hacker News 2Nov 5, 9 AM — quietNov 5, 11 AM — 1 piece · 1 comment — Hacker News 1Nov 5, 1 PM — 1 piece · 1 comment — Hacker News 1Nov 5, 3 PM — quietNov 5, 5 PM — quietNov 5, 7 PM — quietNov 5, 9 PM — quietNov 5, 11 PM — quietNov 6, 1 AM — quietNov 6, 3 AM — quietNov 6, 5 AM — 1 piece · 1 comment — Hacker News 1Nov 6, 7 AM — quietNov 6, 9 AM — quietNov 6, 11 AM — quietNov 6, 1 PM — quietNov 6, 3 PM — 1 piece · 1 post — Bluesky 1Nov 6, 5 PM — quietNov 6, 7 PM — 1 piece · 1 comment — Hacker News 1Nov 6, 9 PM — quietNov 6, 11 PM — quietYesterday, 1 AM — quietYesterday, 3 AM — quietYesterday, 5 AM — quietYesterday, 7 AM — quietYesterday, 9 AM — quietYesterday, 11 AM — quietYesterday, 1 PM — quietYesterday, 3 PM — quietYesterday, 5 PM — quietYesterday, 7 PM — quietYesterday, 9 PM — quietYesterday, 11 PM — quietToday, 1 AM — quietToday, 3 AM — quietToday, 5 AM — quietToday, 7 AM — quietToday, 9 AM — quietToday, 11 AM — quietToday, 1 PM — quiet 1
Nov 3Nov 4Nov 5Nov 6yesterdaynow · 3:04 PM ET
  1. 1

    Tan warns AI agent costs are prohibitively high and advocates user-controlled systems

    Tan acknowledged that AI agent costs are too expensive for mainstream adoption and described his spending as a privilege unavailable to most. He expressed concern about Big Tech-controlled agents and advocated for personal AI agents that users control and program themselves.

    “It's too expensive right now… I know the amount of money he spends on AI agents is 'a crazy privilege that not everyone can do.'”
    — Garry Tan
    1. first by Mastodon, 5d ago · also TechCrunch, HN Frontpage

      2 more headlines
    • richardtuffin.bsky.social

      We're so screwed... —-- Anthropic warns that it’s very close to RSI capability, aka self-building AI

      richardtuffin.bsky.socialBluesky5d ago2▲view on Bluesky ↗
    2 more of the top 3 · 29 posts in this stretch
    • TechCrunch@mstdn.social

      Y Combinator Garry Tan has, in the past, jokingly admitted that he's suffered from "cyber psychosis” and barely sleeps sometimes 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. https://…

      TechCrunch@mstdn.socialMastodon5d ago1▲view on Mastodon ↗
    • Headline aside, I agree with him wholeheartedly that the future of AI is personal agents, and that costs have to come down for that to happen.Not cloud costs, though; consumer costs. These agents must run on-device for privacy protections, because anything less will be a global non-starter. That means we need models that excel on phones, tablets…

      stego-techHacker News4d agoview on Hacker News ↗
    all of them →
  2. background

    Y Combinator deploys internal AI agent called Stray Light for operations — Tan revealed that Y Combinator uses an internal AI agent named Stray Light with a personality and company-wide context, including information on every founder backed, all emails received, and complete applications. This agent is reportedly rebuilding how Y Combinator operates.

  3. background

    Tan describes AI agents as omnipresent assistants with broad data access — Tan characterized his AI agent as a comprehensive personal assistant that reads all his email, Dropbox, Apple Notes, and even therapy notes spanning 20 years. He compares the experience to having a Jiminy Cricket on your shoulder that helps with day-to-day tasks and can anticipate needs.

  4. background

    Garry Tan reveals $100K–$200K annual AI agent spending on Isaacson podcast — Y Combinator CEO Garry Tan disclosed on Walter Isaacson's podcast The Builders that he spends between $100,000 and $200,000 per year on AI tokens. He uses agents for work projects, coding tasks, and personal life management, including access to his email, Dropbox, Apple Notes, and therapy notes from the past 20 years.

What people are saying 21 voices from 3 sites · best of 29 · verbatim