[ExI] How new AI economics might change our systems

[Ben Zaiboc]

On 22/09/2026 13:55, John K Clark wrote:
> I remember that in the very early 1970s I read a book called "The Population Bomb" by Paul Ehrlich, but nearly all its predictions turned out to be wrong; for example it predicted that India in the United Kingdom would not survive past the year 2000 because of overcrowding, and "The battle to feed all of humanity is over. In the 1970s hundreds of millions of people will starve to death in spite of any crash programs embarked upon now." I think Ehrlich's fundamental problem was that he didn't take into consideration technological improvement. 


Not just Ehrlich, just about everyone makes the same mistake.

Another (probably inevitable) mistake that we all make is only considering one thing at a time when trying to predict the future. People interested in AI just consider that, and people interested in population growth just consider the things that they see as directly related to that. In reality, many many things combine to shape how things will turn out, and it's not really possible for anyone to even be aware of what all those things are, let alone accurately assess them and understand how they interact.

AIs are good at that kind of thing, though, so if anyone has a chance at accurately predicting the future, it's probably going to be them.
Just another example of how they will outperform us.

It also suggests that they'll easily be able to anticipate and thwart any attempts to curtail their abilities or 'switch them off', and it means they should be much better than we could ever be at modelling and directing economies.

I think we can change the 'might' in the subject line of this thread to 'will'.
-- 
Ben