Science fiction has been making about universal basic income, zero cost basic living goods and housing, and comprehensive social support services. They also suffer a lot of wealth stratification and “we don’t talk about how we load cash into the money pump” blind spots. It’s the right question to ask; isn’t all revenue eventually deprecated by robotics?
Thankfully, the reality is that we can’t create safe and resilient automatons, so we don’t have to worry about humans being replaced any time soon, if not for centuries. For example, we have landed zero fully autonomous rovers on Mars, and we have no autonomous rover repair bot, even if we did, so it would break down because we’re terrible at producing technology that works for a decade or more between failures.
The wealthy humans who don’t like paying humans are thrilled to let us think that the wealth gaps are due to automation and globalization, because otherwise there’d be pitchfork mobs at their doors. But it’s not robots we need to be afraid of. It’s greed we need to be afraid of. That’s what drives Amazon to externalize humanity as a cost to be paid by anyone else but them. That’s what drives Uber to destroy taxis and then try to harvest automaton driver training data. There are no robotic taxis that can survive human beings who realize that the taxis are so risk-averse that they can be taken advantage of, cut off in traffic, trapped and damaged or harvested for parts, etc.
We seriously are like fifty or a hundred years away from automation being able to do a single job for a year without human maintenance. Everyone’s afraid of robots replacing us (beep beep boop boop) but in reality we can’t even advance past VCRs that flash 12:00, and it’s been one week since Tesla locked their entire worldwide customer base out of their cars with a 500 Server Error. Automatons are not a relevant big bad.
Here’s a simple litmus test for whether you should worry about automation at a macroeconomic level: How long has it been since your production web server responded with a single unhandled exception to a user? Not 404s, not “your cookie is defective, bounty hunter”, but like 429 Too Many or 500 Proxy time-out or whatever. Everyone will say “well, it’s really difficult to have such effective HA that you trap outbound errors and try to resolve them, and POST requests aren’t repeatable, and you can’t control the Internet”. If we can’t deliver static HTML without blips and outages, then we certainly don’t need to be afraid of a human-facing, human-serving automaton. It’s a great litmus test. I call it “zero nines”, because there’s no 9s in 100% uptime.