CharlieKey takeaways from your favorite podcasts
← 19 November 2026AI & I

Notion’s Founder on the One Thing Your AI Is Missing

Ivan argues AI coding revives the 1960s dream of software as a medium everyone can write, but durable value lies in data, context and human understanding, not UI or procedural logic.

Watch on YouTube ↗

Charlie's summary

  1. Vibe coding revives the 1960s vision of software as a medium everyone can read and write

    Ivan says early computing pioneers saw software like reading and writing: something everyone should be able to do. The late-70s/80s Jobs and Gates era froze software into applications, separating makers from users. Now anyone can make their own software, which Ivan sees as the original hippie generation vision coming to life. It is also why Notion started: to bring computing power to people without forcing them to learn to code.

    ▶ 1:29
  2. Just because you can vibe code a tool doesn't mean you should

    Ivan separates three cases. People should have the power to modify the tools they use. But for things like a calendar, most people want someone to craft a well-thought-out solution, as with a chair or a door: “you might not want to vibe coding a bridge or a building.” Security-heavy enterprise software costs a lot to maintain, so he doesn't recommend self-building it, though you can if you want. He admits this fits Notion's business model.

    ▶ 5:32
  3. Data and context outlast UI and procedural logic

    Ivan says agents can now write the workflow code and vibe code any UI, so those no longer hold the value software did for 15 years. What lasts is your data, context and semantics, which are specific to your business and outside any model. Notion sees its job as keeper of that context: opening it to different agents and providing guard rails so a team of humans can collaborate on it.

    ▶ 9:07
  4. Moving Notion from pre-AI to post-AI was the hardest thing Ivan has done

    Ivan got very AI-pilled in late 2022 after friends in labs gave early access to a GPT-4-class model. He and his cofounder locked themselves in a hotel room to build a prototype. Colleagues saw AI as a fad, like crypto. Agents took two more years because models weren't ready. Notion's place now, he says, is where agents meet humans: a context and knowledge system for teams.

    ▶ 14:15

Give your email to unlock the rest of the key points — for every episode.

Do more with Charlie

Free — sign in with your email, no password.