korrents
Andrej Karpathy

Andrej Karpathy

@andrej-karpathy · 69 positions · On ppll ↗

Founding member of OpenAI and former director of AI at Tesla; creator of nanoGPT and the term "vibe coding".

Andrej Karpathy did not write this page.

We collected these quotes from things they published elsewhere, and every quote links to where it was said. They have no account here and have not endorsed this site. Quotes are word for word; the short line under each one is our own restatement, not their wording. Their own site. Is this you? Claim it or ask us to remove it. Or tell us what is wrong here.

4 of Andrej Karpathy's claims fit 4 mental models. Most often: Exponential growth, Make it real, Value moves to what stays scarce.

Models they name

Their own words name the idea.

Exponential growth

Growth that doubles looks like nothing for a long time and then like everything; intuition is linear and gets it wrong.

Used by 13 others

We are already in an intelligence explosion and have been for decades, so AI will show up as more of the same exponential rather than a break in it.

  1. Andrej Karpathy Founding member of OpenAI and former director of AI at Tesla I do, but it's business as usual because we're we're in an intelligence explosion already and have been for decades. And when you look at GDP, it's basically the GDP curve that is an exponential weighted sum over so many aspects of the industry. Everything is gradually being automated has been for hundreds of years. Andrej Karpathy — “We’re summoning ghosts, not building animals”youtube.com · 17 Oct 2025 · 1:22:50 into the videoAll korrents from this video
    I do, but it's business as usual because we're we're in an intelligence explosion already and have been for decades. And when you look at GDP, it's basically the GDP curve that is an exponential weighted sum over so many aspects of the industry. Everything is gradually being automated has been for hundreds of years.

    Watch from 1:22:50 plays here

    ↗Andrej Karpathy — “We’re summoning ghosts, not building animals”youtube.com

    17 Oct 2025 · video · 2h 26m · spoken · machine transcript

Models we see in what they say

Our reading: their claim applies the idea without naming it. The claim is theirs; filing it here is ours.

Make it real

Making a cheap, real version settles questions a description cannot.

Used by 26 others

You only really know something once you have built it — blog posts and slides leave gaps that building the thing does not.

  1. Andrej Karpathy Founding member of OpenAI and former director of AI at Tesla So, don't write blog posts, don't do slides, don't do any of that. Like, build the code, arrange it, get it to work. It's the only way to go. Otherwise, you're missing knowledge. Andrej Karpathy — “We’re summoning ghosts, not building animals”youtube.com · 17 Oct 2025 · 0:30:25 into the videoAll korrents from this video
    So, don't write blog posts, don't do slides, don't do any of that. Like, build the code, arrange it, get it to work. It's the only way to go. Otherwise, you're missing knowledge.

    Watch from 0:30:25 plays here

    ↗Andrej Karpathy — “We’re summoning ghosts, not building animals”youtube.com

    17 Oct 2025 · video · 2h 26m · spoken · machine transcript

Value moves to what stays scarce

When something becomes cheap and plentiful, the value moves to whatever is still rare next to it.

Used by 60 others

The engineer's resource constraint has moved from flops to tokens: what matters now is the token throughput you command.

  1. Andrej Karpathy Founding member of OpenAI and former director of AI at Tesla But now it's not about flops, it's about tokens. So what is your token throughput and what token throughput do you command? Skill Issue: Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AIyoutube.com · 20 Mar 2026 · 0:05:54 into the videoAll korrents from this video
    But now it's not about flops, it's about tokens. So what is your token throughput and what token throughput do you command?

When a measure becomes a target

Reward a number and people, or machines, will find ways to move the number without the thing it measured (Goodhart's law).

Used by 12 others

Process supervision by an LLM judge cannot be run for long, because a judge with billions of parameters is gameable and RL will find its cracks.

  1. Andrej Karpathy Founding member of OpenAI and former director of AI at Tesla the reason that I think this is kind of tricky is quite subtle. And it's the fact that anytime you use an LLM to assign a reward, those LLMs are giant things with billions of parameters and they're gameable. Andrej Karpathy — “We’re summoning ghosts, not building animals”youtube.com · 17 Oct 2025 · 0:46:58 into the videoAll korrents from this video
    the reason that I think this is kind of tricky is quite subtle. And it's the fact that anytime you use an LLM to assign a reward, those LLMs are giant things with billions of parameters and they're gameable.

    Watch from 0:46:58 plays here

    ↗Andrej Karpathy — “We’re summoning ghosts, not building animals”youtube.com

    17 Oct 2025 · video · 2h 26m · spoken · machine transcript