The claim of the article is that expectations increase at a faster rate than the work becomes easier.

It probably feels that way to someone who got comfortable doing their job the old way wishes the world could go back to when they had it all comfortably figured out.

If this does happen, it’s a second order effect from reducing the barrier to entry. Now that anyone can make some code with enough prompts, it’s no longer impressive to be able to make simple apps. You have to bring something different and better to the table. Good developers still do this on top of what AI can do by itself, but I also think there were a lot of relatively easy programming jobs where you didn’t have to be particularly good at anything. You had to know how to follow the rules, get easy tickets from the tracker, make enough changes that you could call it done, and then go on through the dog and pony show of getting the PR approved and getting those story points to count for your standup report. Repeat 5 days per week.

Every mid-size and larger company I worked for had a collection of developers like this: Not really trustworthy or skilled enough to handle big tickets, but you could assign all the loose ends and easy tickets to them. They’d get done eventually while the other devs could focus on bigger efforts.

Now, it’s hard to justify having a lot of those easy ticket solver developers around. An experienced developer can take 10 easy tickets, use their codebase knowledge to write targeted prompts with Opus or Fable, quickly review the code, and have all the easy work done in an afternoon. Then back to the hard tickets.

If you were one of those developers coasting on the easy work, this revolution probably feels like it increased expectations unfairly. If you were one of those devs doing the hard work, you have another tool to use but you’re still grinding away at the hard problems with an extra tool to make some of the parts easier.

The expectation is always that employees will do more for less. Employees are considered a cost, while their productive value is considered an entitlement, so the economics always push for cost reduction.

In practice businesses above a certain size accrete layers of make-work and bureaucracy, which is how you end up with "It takes three weeks to change a button" friction.

AI will make it easier to change the button, but it won't necessarily - yet - cut through the processes around it.

> Now, it’s hard to justify having a lot of those easy ticket solver developers around.

This is such a flawed, but common take. Are you seriously going to claim you were never a junior dev that nobody trusted with "hard tickets"?

Note that I am not making a statement about what your skills were like when you started getting paid for dev work.

I'm saying that the "easy ticket solver" devs are an illusion. When they really fail to perform, you should ask yourself whether it's because many years of bad management undermined the motivation they may have once had. Yes it would suck to be one of those unlucky people, and I'm not saying they shouldn't get fired, but you still need to recognize the bias there. I won't comment on why they weren't given opportunities. There are so many.

Choosing dev work while lacking the necessary skills over other career paths that often pay more, whether back then or even now, is so irrational that it breaks your premise.