x

LESSWRONG

LW

djbinder — LessWrong

djbinder

Top postsTop post

djbinder

Message

I am a researcher at Coefficient Giving where I work on biosecurity and pandemic preparedness, and also think about other transformative technologies that powerful AI systems might enable.

200

5

3

6y

djbinder

I am a researcher at Coefficient Giving where I work on biosecurity and pandemic preparedness, and also think about other transformative technologies that powerful AI systems might enable.

Top postsTop post

The AI Industrial Explosion — Part 1: Maximum growth rates with current production methods

How fast could an AI-driven economy grow? Most economists expect a few percentage points at best, comparable to previous general-purpose technologies (Acemoglu (2024)). Those closer to AI development tend to imagine something much more radical (Shulman (2023); Davidson and Hadshar (2025)). This series aims to ground growth rates in how physical production works. Once human labor is automated, the constraint on growth becomes the speed at which the economy's physical capital can reproduce itself. Government input-output tables track the full supply chain of what it takes to produce every commodity in the economy, and we can use them to compute this self-reproduction rate directly. In this post, I compute the maximum rate at which an autonomous AI economy could grow, once its production is concentrated in the sectors most important for self-replication. I take the conservative case for this calculation: full automation, but no other technological improvement. Using US input-output data, I find this economy could double in about a year, in line with other estimates that assume full automation (Hanson (2001); Trammell and Korinek (2023); Davidson and Hadshar (2025); Epoch AI (2025)). This holds up even after accounting for resource depletion and construction lags. Some output goes to consumption rather than reinvestment, which slows things down, but even moderate savings rates imply doubling times around two years. The growth rate computed here is the maximum sustainable rate for any economy using these production processes, whether it evolves from an existing economy or is built from scratch. It is also the growth rate that a rapidly growing economy will converge to in the long run, regardless of its starting composition. [1] The present-day US economy is far from the fastest-growing composition, with too many services and consumer goods relative to mining and heavy manufacturing. In Part 2 we will examine how rapidly the US could make this transition, while still a

Will whole brain emulation matter for the AI transition?

The AI Industrial Explosion — Part 2: Transition Dynamics

The AI Industrial Explosion — Part 3: Going faster

The AI Industrial Explosion — Part 3: Going faster

In Part 1, I found that a fully automated economy using today's production methods could double roughly every year. In Part 2, I modeled the transition from today's economy to that maximum-growth composition and found that energy production could double within about four years. Both parts held production methods fixed:...

The AI Industrial Explosion — Part 2: Transition Dynamics

This is Part 2 of a series on post-AGI economic growth. Part 1 established that a fully automated economy could double roughly every year using current technology. But the US economy does not currently look like a self-reproducing capital machine. It overproduces consumer goods and services relative to maximum growth,...

The AI Industrial Explosion — Part 1: Maximum growth rates with current production methods

How fast could an AI-driven economy grow? Most economists expect a few percentage points at best, comparable to previous general-purpose technologies (Acemoglu (2024)). Those closer to AI development tend to imagine something much more radical (Shulman (2023); Davidson and Hadshar (2025)). This series aims to ground growth rates in how...

Will whole brain emulation matter for the AI transition?

Summary * WBE is decades away assuming no AGI. Expert timelines for human brain emulation are mid-2040s at the earliest, with wide error bars, and generally are for "simulating the brain for a few seconds" rather than for any useful length of time. There is essentially no overlap with AGI...