“Estimating the Substitutability between Compute and Cognitive Labor in AI Research” by Parker_Whitfill, CherylWu

Audio note: this article contains 127 uses of latex notation, so the narration may be difficult to follow. There's a link to the original text in the episode description.

Confidence: Medium, underlying data is patchy and relies on a good amount of guesswork, data work involved a fair amount of vibecoding.

Intro: 

Tom Davidson has an excellent post explaining the compute bottleneck objection to the software-only intelligence explosion.[1] The rough idea is that AI research requires two inputs: cognitive labor and research compute. If these two inputs are gross complements, then even if there is recursive self-improvement in the amount of cognitive labor directed towards AI research, this process will fizzle as you get bottlenecked by the amount of research compute.

The compute bottleneck objection to the software-only intelligence explosion crucially relies on compute and cognitive labor being gross complements; however, this fact is not [...]

---

Outline:

(00:35) Intro:

(02:16) Model

(02:19) Baseline CES in Compute

(04:07) Conditions for a Software-Only Intelligence Explosion

(07:39) Deriving the Estimation Equation

(09:31) Alternative CES Formulation in Frontier Experiments

(10:59) Estimation

(11:02) Data

(15:02) Trends

(15:58) Estimation Results

(18:52) Results

The original text contained 13 footnotes which were omitted from this narration.

---

First published:
June 1st, 2025

Source:
https://forum.effectivealtruism.org/posts/xoX936hEvpxToeuLw/estimating-the-substitutability-between-compute-and

Narrated by TYPE III AUDIO.

---

Images from the article:

---

---

---

---

---

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.