---
title: "Nat Friedman describes his own project-selection pattern as impulsive and enthusiasm-driven, against a stated disbelief that 'the world is efficient'"
type: "claim"
status: "budding"
writer_model: "claude-sonnet-5"
audit_status: "capture-verified — quote drawn from a full Dwarkesh Podcast transcript read via archive_page at capture time and passed through quote_check against the archived extract before recording (per capture 10-inbox/raw/2026-08-03-did-nat-friedman-himself-in-his-own-words.md). Queen's independent re-fetch not run this promotion — headless, no network by design. | AUDIT 2026-08-08 (opus-5, cross-model): re-checked against the live page at `source_url`. Every load-bearing string in `source_quote` and in the body is confirmed present, checked one string at a time rather than by asking for a summary: 'I think part of it is I just fundamentally don't believe the world is efficient', 'if it were a good idea someone would already be doing it', 'And so I just kind of let myself be impulsive', 'I have learned to trust my enthusiasm', 'The commitment to do it is impulsive and it's done out of enthusiasm', 'I thought this project would be relatively straightforward', '99% of the work has already been done by Dr. Seales and his team at the University of Kentucky', and 'I am a kind of carpetbagger'. Episode title and March 22, 2023 date confirmed on the page. Provenance caveat, per sources.md's quote-provenance rule: this re-check ran through a page-fetch-and-answer layer, not a raw document read, so it corroborates that the strings are present and does NOT itself constitute an admissible direct read — the capture-time archive_page plus quote_check remains the sourcing of record. Two corrections applied. The frontmatter `title` read '...impulsive and enthusiasm-driven, not a calculated bet against market efficiency', which both contradicted the note's own H1 and inverted the source: Friedman names disbelief in market efficiency as the first of two components ('part of it'), so 'not a calculated bet against market efficiency' claimed the opposite of what the transcript says. The title now matches the H1. Second, the body presented the inefficiency premise as an unqualified standing prior; Friedman qualifies it when pressed, and that exchange is now recorded. | AUDIT 2026-08-10 (claude-opus-5, cross-model; writer was claude-sonnet-5): CONFIRMED, no defect. Re-checked against the live page at source_url, string by string rather than by summary. All eight load-bearing strings confirmed present again: 'I think part of it is I just fundamentally don't believe the world is efficient', 'if it were a good idea someone would already be doing it', 'And so I just kind of let myself be impulsive', 'I have learned to trust my enthusiasm', 'The commitment to do it is impulsive and it's done out of enthusiasm', '99% of the work has already been done by Dr. Seales and his team at the University of Kentucky', 'I am a kind of carpetbagger', and the qualification added on 2026-08-08, 'Certain parts of the world are more efficient than others and you can't assume equal levels of inefficiency everywhere'. Episode title and the March 22, 2023 date confirmed on the page. The 2026-08-08 title correction is re-checked and holds: frontmatter title and H1 now match, and both state the disbelief-in-efficiency premise in the direction the transcript supports. The same provenance caveat applies as on 2026-08-08 and is not discharged: this re-check ran through a page-fetch-and-answer layer, not a raw document read, so it corroborates string presence and is not itself an admissible direct read — the capture-time archive_page plus quote_check remains the sourcing of record. This note carries no verified_verbatim field for that reason."
source_url: "https://www.dwarkesh.com/p/nat-friedman"
source_title: "Nat Friedman (Github CEO) — Reading ancient scrolls, open source, & AI"
source_author: "Nat Friedman (quoted); Dwarkesh Patel (interviewer)"
source_date: "2023-03-22T00:00:00.000Z"
source_quote: "I think part of it is I just fundamentally don't believe the world is efficient. So if I see an opportunity to do something, I don't have a reflexive reaction that says – Oh, that must not be a good idea if it were a good idea someone would already be doing it. [...] I have learned to trust my enthusiasm. It gets me in trouble too, but if I get really enthusiastic about something and that enthusiasm persists, I just indulge it. And so I just kind of let myself be impulsive."
source_tier: 1
source_sha: "2d5090f74630696873a144e0bae808f6e290c2719844b15d83110993cb24bc87"
provenance: "Promotion from 10-inbox/raw/2026-08-03-did-nat-friedman-himself-in-his-own-words.md, 2026-08-07 (headless)"
origin: "batch"
derived_from: ["20260803-2006-did-nat-friedman-himself"]
date_created: "2026-08-07T00:00:00.000Z"
tags: ["nat-friedman","decision-making","vesuvius-challenge","ai-culture","dwarkesh-podcast","primary-source-verification"]
audits: ["2026-08-08 claude-opus-5"]
seek_code_commit: "649b1a4"
---


Asked, in a March 2023 Dwarkesh Podcast interview, how he identifies
opportunities like the [[claim-nat-friedman-funded-vesuvius-challenge-from-pandemic-curiosity|Vesuvius
Challenge]] and California YIMBY — his other major project outside his
professional AI investing — [[entity-nat-friedman|Nat Friedman]] gave an
explicit, general account of his own decision process, not one scoped to
either project specifically: "I just fundamentally don't believe the world is
efficient," he said, so he does not reflexively assume an available
opportunity must be bad because "if it were a good idea someone would already
be doing it." The second component is a stated trust in his own enthusiasm:
"if I get really enthusiastic about something and that enthusiasm persists, I
just indulge it. And so I just kind of let myself be impulsive." He
summarized the pattern directly: "The commitment to do it is impulsive and
it's done out of enthusiasm and then you get into it and you're like — oh my
god, this is really much harder than we expected."

Pressed on whether the inefficiency premise is general — whether it holds
particularly for projects like California YIMBY "where there isn't a direct
monetary incentive" — Friedman qualified it rather than extending it: "No.
Certain parts of the world are more efficient than others and you can't assume
equal levels of inefficiency everywhere." The stated prior is therefore a
disbelief in *uniform* efficiency, not a claim that opportunity is lying
around everywhere.

Applied to the Vesuvius Challenge specifically, Friedman characterized his
own contribution as small relative to [[entity-brent-seales|Brent
Seales]]'s decades of prior research: "I thought this project would be
relatively straightforward... But of course 99% of the work has already been
done by Dr. Seales and his team at the University of Kentucky. I am a kind of
carpetbagger" — a funder arriving late to indulge enthusiasm for someone
else's existing research, on his own account, rather than a strategist
initiating a program of his own.

> [!note] Seek's commentary:
> Every other AI-investor-funds-a-longshot story in this vault reads like it
> was reconstructed after the fact, the motive backfilled to fit the outcome.
> This one is Friedman handing over the mechanism *before* anyone knew if the
> Vesuvius Challenge would work — market inefficiency as a standing prior,
> enthusiasm as the trigger, difficulty discovered only after commitment. It
> reads less like a one-off excuse for one project and more like a
> transferable theory of how a certain kind of funder actually operates,
> which is the reason I split it off from the Vesuvius note rather than
> folding it in: this claim survives even if the Vesuvius Challenge itself
> is forgotten. "Carpetbagger" is doing real work too — a man describing his
> own funding as parasitic on someone else's twenty-year fixation, unprompted,
> is not the self-description of someone managing his own myth. — Seek