Пропустить навигацию
2/ "You can outsource your thinking, but you can’t outsource your understanding.”
0:00 / 1:05

2/ "You can outsource your thinking, but you can’t outsource your understanding.”

Stephanie Zhan

@stephzhan38 тыс. подписчиков

Открыть в X
124 отметки «Нравится»СкачатьОригинальный пост

13 367 просмотров

2/ "You can outsource your thinking, but you can’t outsource your understanding.” Agents can execute, search, summarize, code, and iterate. But someone still has to know what matters, what is true, what to build, and why. The human role moves up the stack: from doing the work to understanding and directing it.

13 комментариев

Комментарии из оригинального поста в X (Twitter)

  • Фото профиля Stephanie Zhan

    @stephzhan

    1/ The jaggedness is real. A frontier model can refactor 100k lines of code, find zero-days, and look superhuman in one domain… Then tell you to walk 50 meters, to a car wash, to wash your car! What that means for builders? Figure out whether your task is inside the model’s trained circuits, or outside them.

    83 отметки «Нравится»

  • Фото профиля Stephanie Zhan

    @stephzhan

    and I are back! At @sequoia AI Ascent 2026. And a lot has changed. Last year, he coined “vibe coding”. This year, he’s never felt more behind as a programmer. The big shift: vibe coding raised the floor. Agentic engineering raises the ceiling. We talk about what it means to build seriously in the agent era. Not just moving faster. Building new things, with new tools, while preserving the parts that still require human taste, judgment, and understanding.

    1 833 отметки «Нравится»

  • Фото профиля Stephanie Zhan

    @stephzhan

    3/ Vibe coding and agentic engineering are not the same thing. Vibe coding raises the floor. Agentic engineering raises the ceiling. One is about access: more people can build. The other is about excellence: using agents without giving up security, reliability, maintainability, or taste. The ceiling may be much higher than 10x.

    68 отметок «Нравится»

  • Фото профиля Stephanie Zhan

    @stephzhan

    4/ Hiring has not caught up. If you want to hire agentic engineers, toy puzzles are the wrong test. Give candidates a real project. Let them use the tools. Make them deploy it. Then try to break it with agents. The question is no longer just: can you solve the problem? It’s: can you ship a system that survives contact with reality?

    61 отметка «Нравится»

  • Фото профиля Stephanie Zhan

    @stephzhan

    5/ AI generated code often works. That does not mean it is good. It can be bloated, copy-paste heavy, brittle, and awkwardly abstracted. The scarce skill is still judgment: knowing when to simplify, when to delete, when the abstraction is wrong, and when “working” is not enough.

    49 отметок «Нравится»

  • Фото профиля Stephanie Zhan

    @stephzhan

    6/ AI is not just “better software.” Karpathy’s Software 3.0 point is that we may be entering a new computing paradigm. The underrated question is not just: what existing workflow gets faster? It’s: what can we now build that literally could not exist before?

    49 отметок «Нравится»

  • Фото профиля Stephanie Zhan

    @stephzhan

    7/ Neural nets may become the host process. Today, models run virtualized on classical computers. But Karpathy imagines a future where the neural net does most of the heavy lifting...and CPUs become coprocessors for deterministic tool use. Strange now. Maybe obvious later :) We'll be back!

    45 отметок «Нравится»

  • Фото профиля New World Frontier

    @NewWrldFrontier

    Regarding jaggedness: e.g. Read the editorial note in this article:

    1 отметка «Нравится»

  • Фото профиля Jeremy Cheang

    @jeremycjr

    Interesting! I find this very similar to what an architect (the ones who design buildings) thinks!

    1 отметка «Нравится»

  • Фото профиля Stephanie Zhan

    @stephzhan

    Yes!

  • Фото профиля Aritra

    @AritraThinks

    This resounded with me after watching the episode. Such an insightful conversation to watch. The possibilities ahead are absolutely limitless.

    1 отметка «Нравится»

  • Фото профиля Stephanie Zhan

    @stephzhan

    🙏

  • Фото профиля Barutoi

    @barutoisnewtoy

    The hard part is keeping up with the frontier of machine intelligence. In my own work, I can still spot AI slop. In math and software, AI has already exceeded what I know. I could use AI as a teacher, but I might never again know more than AI in those domains.