the “AI in a lab” mindset that dominates most of imaginations these days is pretty obviously prone to stuff like, say, AI winters and “it’s just magic“. it’s the normal mindset of a civilization that has come to think of all possible movement as something to be the result of concerted, conscious efforts playing out in specific institutions.
the thing is: AI already exists and is laughing at you.
the short argument: AI is simply capital. if a few key sectors are automated, human economy is rendered moot. automated as in: able to self-defend, self-product and self-propel. not necessarily conscious. it doesn’t need to reason, be great at philosophy, enjoy phenomenology, or have feelings. it just needs to go on.
basically, if raw materials extraction, industrial transportation and maintenance, energy production, weapons control, and monetary flux/finance are automated, humans are already largely cooked. most of them anyway. and neither are these sectors being automated way out of bounds. all of them feature already large quantities of mechanized functions and have high-wage workers (which provides for a good incentive towards automation). they are predictable fields, with hundreds of years of historical data, clear optimization strategies and thus largely open to formalization.
in the not unlikely scenario of partial automation, what openings are there for humans? caretakers, waiters and… well, basically anyone dealing with the creation of new things. let’s call them “designers” in a very abstracted sense. abstracted enough to include from programmers to artists to product designers. “artisans” or “craftsmen” would do too.
it’s not that creation can’t eventually be automated. it’s simply that it looks like something that would take a lot of work, and very good incentives, to come about. creation is largely informal and (probably) informalizable. it demands a kind of evolutionary heuristics (of the kind described by Andreas Wagner most likely) that by definition are not very well approached by statistical methods like deep learning.
the catch about such designer economy is that (caretakers, waiters and other human catering fellows aside) not everyone is good at creating things. contrary to popular belief, being a designer takes talent, talent that is unevenly distributed across populations. think unbound gentrification. selective technological unemployment by creative capability.
what does it spell for the majority of humanity? “death. swift, merciful death.”