I think The Future Doesn’t Suck has it mostly right. While I think the current crop of AI has made it well passed “works good enough to actually bend the efficiency curve”, I don’t think the result of this will be reduced headcount. Like with most things, simple supply and demand models are still at work here. When it becomes less expensive to deliver something, demand goes up. The fact that I can write more code, whether it’s delivering features or improving bugs, means that the expectation from users of software products will be higher. They’ll expect more features, at higher quality, delivered faster than ever before. And they’ll want that because if I don’t offer it, someone else will.
While I’d like a four day work week, making a white collar worker more efficient at building zero marginal cost products is not going to get me there. Customers will choose software that has engineers who work five days a week but produce 6+ days worth of improvement over those that work 4 days and get 5 days of improvement.
This is not literally always true or endlessly true, but it’s true enough and the accessibility of AI is enough that in competitive markets I expect this to be the case.
I can do my job more efficiently and effectively, but so can everyone I’m competing against. So I’ll have to keep chasing them. Doing a good job still means delivering a better product, but the bar has been raised.
I don’t think I’m going to work less– I think I’m going to ship better products. What I’m actually hopeful for is that my job will be more fun. The jury is still out on that one.