The Work After Work

[Image omitted]

At some point down the road, almost all traditional work will be automated by AI and machine learning. This leaves the question — what will be left for us to work on?

It’s a weird question to think about, because it’s one that has a lot more to do with psychology and economics and less to do with business and technology. The “Great Automation” that will occur down the line has to do with a wider trend of resources getting less and less scarce as our methods for developing them improve. I’ll combine these two ideas together to help us get at least a little bit closer to the answer of the work after work.

Let’s start with the beginning. If the ML revolution goes off without a hitch, all algorithmically-based work — driving, accounting, retail, consulting (thank god), programming, etc. — goes away. This is easy stuff for ML to do once it gets a hang of it. It is the things that aren’t algorithmically based, the more abstract fields like writing, art, research, cooking, etc. — that AI probably couldn’t take over. Yes, we have AI that knows how to write. We have AI that can make unique images out of prompts and can look at genes and find out the best ways to make new drugs. But there’s two problems with this: 1) They can’t do them very well, and 2), by consequence, their ceiling on this isn’t very high. Just because GPT-3 can write an essay that vaguely makes sense doesn’t mean it’s going to write a book on the same level as Dostoevsky.

Now, where this takes a turn is in AI-assisted work. If Dostoevsky had access to a powerful writing AI that could do most of the heavy lifting for him, his work would likely be both more effective AND efficient. Not only is at least some of the AI output usable (saving him time in writing), but also the AI generates unique ideas that he may be able to go off of and use to write a better story. This gives us a good idea for how things might work.

However, we’ve only really touched art so far. And, as much as I do like fiction better than non-fiction, you can’t really make nation-state economies out of it. We trade primarily in things that make our lives more convenient. Art is nice, but it doesn’t have this flavor (which is why art sells at a heavy discount). The good news is that AI-assisting follows quite nicely into entrepreneurship. AI can work, but it can’t work by itself (sigh no we aren’t talking about general intelligence in this article, go away, I’ll make one later maybe). However, if it had a general manager of sorts — a person with a vision of one a more convenient and efficient future could look like — then perhaps the two would make a great team, while also keeping the economy afloat.

Another issue — and the more pressing short-term one at that — is what happens to the people who don’t know anything about AI. To this question I ask what happened when people didn’t learn how to code after the Internet revolution. The answer was… nothing! The tech just got more friendly and people figured it out on their own. Hell, we have NoCode trending now for goodness sake. We did just fine!

I think the worry behind workforce retraining is the rather derogatory (though perhaps not purposely so) assumption that you can’t train truck drivers to do anything but truck driving. Or replace “truck driving” with any other low-skill, easily automated work. People can pick up new skills easily if the design is friendly and especially if it correlates to their future survival. How long does it take someone to learn NoCode? A teenager probably a few hours — an old school truck driver, maybe a few months? But a few months is a far cry difference from never being able to learn something. When you’re staring at statistics and spreadsheets all day it’s easy to forget about humankind’s ability to adapt.

In conclusion: the work after work will be different, that’s for sure. But it won’t be life-threatening. Chances are it will be just like the Internet, gradual enough that we can learn the important stuff over time and transition in a perfectly reasonable manner. We might have snags — especially on an individual or group level — but the work after work will be much more exciting than trying.

[Image omitted]