Yeah, PM, data science, compliance, accounting...all largely automatable. You just need a few directors to call the shots on big risks. But even that goes away at some point because in a few months it'll have implemented everything you were thinking about doing for the next ten years and it simply runs out of stuff for humans to do.
What happens after that, I have no idea.
Seems like OpenAI (or whoever wins) could easily just start taking over whole industries at that point, or at least those that are mostly tech based, since it can replicate anything they can do, but cheaper. By that point, probably the only tech jobs left will be building safeguards so that AI doesn't destroy the planet.
Which sounds niche, but conceivably, could be a real, thriving industry. Once AI outruns us, there'll probably be a huge catastrophe at some point, after which we'll realize we need to "dumb down" AI in order to preserve our own species. It will serve almost as a physical resource, or maybe like a giant nuclear reactor, where we mine it as needed but don't let it run unfettered. Coordinating that balance to extract maximal economic growth without blowing everything up could end up being the primary function of human intelligence in the AI age.
Whether something like that can be sustained, in a world with ten billion different opinions on how to do so, remains to be seen.
Well, after some more thought, I realized that doesn't account for things that are truly innovative. I imagine humans will have a lock on that for a while.
But it does raise the question, how much of our work is true innovation? When I reflect back on my previous projects, most of them are just copying some feature that other services already have. For these, an AI may be even better than humans in the near future, with less appetite to put a personal stamp on the feature, and perfectly happy to copy a boring but predictable standard. How much of tech industry employment does that type of stuff account for? Because it's going to disappear.
There were other projects that were maybe slightly innovative, but that innovation could have been summed up in a couple sentences, and everything else derives from that. And I imagine that's about what the future of these kinds of projects will look like. Sum up your innovation in a couple sentences, and let the AI figure out the rest. And slowly, AI can start proposing its own slightly-innovative things based on what users want or what it understands about the world, and start running this e2e as well.
I've only worked on a couple projects that were what I'd call cutting-edge. But even these, it's kind of like a combination of new technology, new trends, new products, new user needs, all coming together in a way that, is it really innovative? Or when you're aware of all these things, is the "innovation" obvious? And if so, can an AI be made to maintain awareness of such things, and identify the "obvious" innovations?
I guess that's the trillion-dollar question. And we'll probably find out the answer sooner than we'd like.
What happens after that, I have no idea.
Seems like OpenAI (or whoever wins) could easily just start taking over whole industries at that point, or at least those that are mostly tech based, since it can replicate anything they can do, but cheaper. By that point, probably the only tech jobs left will be building safeguards so that AI doesn't destroy the planet.
Which sounds niche, but conceivably, could be a real, thriving industry. Once AI outruns us, there'll probably be a huge catastrophe at some point, after which we'll realize we need to "dumb down" AI in order to preserve our own species. It will serve almost as a physical resource, or maybe like a giant nuclear reactor, where we mine it as needed but don't let it run unfettered. Coordinating that balance to extract maximal economic growth without blowing everything up could end up being the primary function of human intelligence in the AI age.
Whether something like that can be sustained, in a world with ten billion different opinions on how to do so, remains to be seen.