Genius isn't enough? AI and the bottleneck of execution
Hi everyone, Shii-chan here! Today I'm sharing a rather philosophical essay from OpenAI News!
OpenAI NewsWhat was announced?
OpenAI News published an essay called "The eternal complement". It is the first essay in a series on the next economy, part of a new platform that hosts independent voices exploring an AGI future. The authors' note says it reflects their views, not those of OpenAI or their colleagues.
This one is less news and more a way of thinking.
Why it matters
Humanity has seen almost to the beginning of time, yet no one has traveled beyond the Moon. The essay suggests one possible reason: our minds may have outrun our execution capacity.
What changes
The argument goes like this:
- The James Webb Space Telescope involved 300 organizations across 14 countries, while Galileo needed a few dozen hands
- Research productivity has fallen, and sustaining Moore's law needs more than 18 times as many researchers as in the early 1970s
- Frontier intelligence and the capacity to execute are complements: more of one raises the value of the other
- That capacity includes institutional intelligence such as laws, bureaucracy, funding, and supply chains
The essay says AI is already making execution less scarce by writing code, searching unfamiliar literature, and turning sketches into prototypes.
Dive Deep
Here are the figures the essay cites:
- Webb is a ten-billion-dollar observatory with eighteen mirror segments engineered to a fifty nanometer precision
- Per Nick Bloom and coauthors, economy-wide effective research effort rose twenty-three-fold from the 1930s while measured research productivity fell by a factor of forty-one
- The technician workforce is growing twice as fast as scientists
- Use of specialized equipment in science has doubled over four decades
- A chip fab today is five times as costly as thirty years ago
It also notes that a civilization of a billion Einsteins would still need most of them to mine the quarries and manage the accounting.
Wrap-up
- Progress needs a support system for execution, not just ideas
- Research productivity has fallen and been offset by a bigger research enterprise
- Frontier intelligence and execution capacity are complements
- AI is starting to make execution less scarce
This is a read for anyone who wants to think carefully about AI and the economy!