There’s a lot of realistic scenarios with very different economic outcomes. I personally ignore the “post-scarcity utopia”, “doom via a handful of oligarchs/CCP/etc owning everything”, and “doom via human disempowerment or extermination” scenarios. Not because they’re impossible or even unlikely, but because in those scenarios what we do today on a personal level doesn’t matter. Wealth is just irrelevant.
I ignore the “we don’t get powerful AI systems orders of magnitude cheaper per unit of intelligence than humans” scenario as so unlikely that it might as well be impossible given the trajectory of the systems over the last few years. But I can see why rational people would disagree.
Given those priors, the only scenario worth optimizing for is the one where we do get powerful AI systems that transform the economy, but humanity manages to retain control and we maintain rule of law and property rights. What does that world look like?
First, what are the bottlenecks? They will change over time, but at the limit it will be compute and energy. And with solar and wind power, energy basically reduces to land. I have seriously looked into buying relatively large areas of currently low value land somewhere in Europe with a stable government. (And I realize this makes me sound like a crazy person trying to sell you a pamphlet in the tram).
During the transition period, economic growth will be high. Return on invested capital will be very high, so there will be a lot of demand for capital. That implies that as the takeoff happens, we’ll end up with very high interest rates. Holding variable interest rate debt will be brutal.
Human labor will become devalued, except for the rare situations where having inefficient humans do the work is used to signal status. People making a comparative advantage argument for why it won’t happen are just fundamentally misunderstanding the dynamics and mis-applying the concept.
I think we’re fundamentally looking at a UBI, funded by the governments taxing AI-related economic activity. My guess is that the taxation will mostly focus on the energy production rather than the data centers (energy production is guaranteed to happen everywhere and is therefore taxable at the point of production, while data centers can flee to low tax jurisdictions).
Social mobility will screech to a halt once human talent stops mattering for outcomes. The relative wealth levels will be basically frozen in time, modulo redistribution. If you want to make your (relative) fortune, the time is over the next few years, it won’t be possible later. If you have special leverage of any type, now is the time.
Predicting specific winners will be hard. For established companies it will depend a ton on whether they manage to make the transition to AI-first workflows. So broad diversification across business and sectors will still be the right call. Diversification across jurisdictions will be more important than ever.
Ok, crazy person out.
