Rosehip Mole asked, and Nib — askNib's tutor — drew the answer live at a whiteboard. This is the spoken transcript; enable JavaScript to watch it drawn.
Teach me this page — https://www.planned-obsolescence.org/p/could-ai-accelerate-economic-growth
Quick history: growth used to crawl, under zero point one percent a year, thousands of years ago. Today the world grows about three percent a year. Something changed.
You'd think a single new technology causes this — but economists find the opposite. Electricity, computers, the internet: each gave one bump, then growth just returned to its old rate.
So what actually raised the growth rate long-term wasn't any single invention — it was more people doing research, continuously. Population is thirty times bigger than two thousand years ago, and a much larger share of us do R&D full time.
Today about twenty million people worldwide work in research and development — that's the engine behind our three percent growth. The crux of this essay: what if AI could multiply that engine, not by one new gadget, but by flooding the world with researchers?
Here's the mechanism: training a superhuman AI takes enormous compute, but once trained, running it is cheap. So the same chips that built ONE genius AI could run billions of copies in parallel, working day and night.
Add speed: ChatGPT already writes twenty times faster than a human types. Put superhuman skill, billions of copies, and twenty-times speed together, and you get his headline estimate — over a hundred-fold jump in the world's total research effort.
Put it all together: today twenty million humans researching gives us three percent growth. If AI multiplies effective research effort over a hundred times, the argument is growth itself — not just output — could accelerate, something no past technology ever did.
The surprising claim isn't that AI is one more good gadget — it's that AI could multiply the number of researchers on Earth, which is the one thing that has ever truly changed the growth rate itself.