EXCLLENT POST!!Borrowing from the theory of Ricardian comparative advantage, in theory there can still be productive exchanges, even if your trading partner is better than you at everything you do.
However, this theory relies on there being limited capacity for production. You allocate your time / resources to the thing you're best at, and the trading partner allocates their time to whatever they are best at, and you trade, and both come out ahead vs. not trading.
However, limited capacity doesn't really apply to AI. One piece of software can run 10 billion instances, in theory.
So if AI gets better at our jobs than we are, there isn't really any way back in, unless we come up with something that is 1) worthwhile and 2) can't be done by AI.
And "can't be done by AI" is a list that seems to shrink faster than the list of worthwhile activities grows.
This also leaves aside the possibility of "true" AI, something that approximates a thinking machine on par with or beyond human intelligence. (Doesn't need to actually think, just needs to behave as if it does.) Such a thing doesn't seem too far off anymore. Many of the "computers will never be able to..." hurdles have been cleared in the past few years alone.
Once a machine is smarter than you for any given task you might both perform... are you even employable IN THEORY? I say no. This is a significant break with economic orthodoxy, but again, workers are built into the math for traditional economics, because the idea of a distinction between man and machine is inherent to the concepts of labor and capital.
To the extent that distinction weakens or even goes away (just for the purposes of employment, let's not get all Blade Runner here) then people are simply no longer required for production, and will not be paid for it. Which actually destroys our entire concept of an economy... it's predicated on investors paying people who are performing productive work.
Once workers are out of the picture, one final round of investment can happen before money itself becomes conceptually invalid.
Vonnegut's Player Piano coming in hot...
I had to scream.
I find these advances in AI, terrifying. The sheer amount of availability is no laughing matter. Aren't we playing with a fire we cannot extinguish? Haven't we crossed the Rubicon? A point of no return? The question needs to be asked : For what purpose?
A very concerned Human being.
Peace.