Do any of these jobs survive in the “no humans have any other jobs than this” environment? E.g., marketing research is not valuable when AIs do ~all economic consumption.
I doubt that these jobs will keep the unemployment rate anywhere near the historical average, but I am confident that these jobs will survive.
I have two counterpoints to your claim about economic consumption:
In this case, the absolute measure of humans’ consumption matters more than the relative measure. Unless the overwhelming majority of humans will live in far worse circumstances than they do today, billions of humans is a large-enough customer base to warrant marketing research.
On a per-capita basis, AIs consume far less than humans do. That’s precisely what makes them cheaper to employ. Even if AI agents will outnumber humans by a factor of 1,000, I doubt that their total consumption would exceed humans’. (AIs might purchase things to increase their own productivity, but that’s capital investment, not consumption. If they really were consuming in vast amounts—e.g. collecting paperclips just for the heck of it—then I’d question whether we’ve really solved the alignment problem.)
Do any of these jobs survive in the “no humans have any other jobs than this” environment? E.g., marketing research is not valuable when AIs do ~all economic consumption.
I doubt that these jobs will keep the unemployment rate anywhere near the historical average, but I am confident that these jobs will survive.
I have two counterpoints to your claim about economic consumption:
In this case, the absolute measure of humans’ consumption matters more than the relative measure. Unless the overwhelming majority of humans will live in far worse circumstances than they do today, billions of humans is a large-enough customer base to warrant marketing research.
On a per-capita basis, AIs consume far less than humans do. That’s precisely what makes them cheaper to employ. Even if AI agents will outnumber humans by a factor of 1,000, I doubt that their total consumption would exceed humans’. (AIs might purchase things to increase their own productivity, but that’s capital investment, not consumption. If they really were consuming in vast amounts—e.g. collecting paperclips just for the heck of it—then I’d question whether we’ve really solved the alignment problem.)