Adding to this: it's not just that the apprenticeship ladder is gone—it's that nobody wants to deal with juniors who spit out AI code they don't really understand.

In the past, a junior would write bad code and you'd work with them to make it better. Now I just assume they're taking my feedback and feeding it right back to the LLM. Ends up taking more of my time than if I'd done it myself. The whole mentorship thing breaks down when you're basically collaborating with a model through a proxy.

I think highly motivated juniors who actually want to learn are still valuable. But it's hard to get past "why bother mentoring when I could just use AI directly?"

I don't have answers here. Just thinking maybe we're not seeing the end of software engineering for those of us already in it—but the door might be closing for anyone trying to come up behind us.

It's not helping that in the last 10 years a culture of job-hopping has taken over the tech industry. Average tenure at tech companies is often ~2 years and after that people job hop to increase compensation.

It's clear why people do it (more pay) but it sets up bad incentives for the companies. Why would a company invest money in growing the technical skill set of an employee, just to have them leave as soon as they can get a better offer?

I hope juniors will figure out how to use AI to do larger tasks that are still annoying for seniors to do, while seniors take on larger tasks still. I think it's just seniors are learning this stuff faster at the moment and adapting it faster to current work, but as all that changes I would guess juniors reclaim some value back.

That said, you hit on something I've been feeling, the thing these models are best at by far is stuff that wasn't worth doing before.

I grew up in the 70s. The hand wringing then was calculators. No one was going to be able to do math anymore! And then wrist watches with calculators came out. Everyone is going to cheat on exams, oh no!

Everything turned out fine. Turns out you don't really need to be able to perform long division by hand. Sure, you should still understand the algorithm at some level, esp. if you work in STEM, but otherwise, not so much.

There were losses. I recall my AP physics professors was one of the old school types (retired from industry to teach). He could find the answer to essentially any problem to about 1-2 digits of precision in his head nearly instantly. Sometimes he'd have to reach for his slide rule for harder things or to get a few more digits. Ain't no one that can do that now (for reasonable values of "no one"). And, it is a loss, in that he could catch errors nearly instantly. Good skill to have. A better skill is to be able to set up a problem for finite element analysis, write kernels for operations, find an analytic solution using Mathematica (we don't need to do integrals by hand anymore for the mot part), unleash R to validate your statistics, and so on. The latter are more valuable than the former, and so we willingly pay the cost. Our ability to crank out integrals isn't what it was, but our ability to crank out better jet engines, efficient cars, computer vision models has exploded. Worth the trade off.

Recently watched an Alan Guth interview, and he made a throwaway comment, paraphrased: "I proved X in this book, well, Mathematica proved...". The point being that the proof was multiple pages per step, and while he could keep track of all the sub/superscripts and perform the Einstein sums on all the tensors correctly, why??? I'd rather he use his brain to think up new solutions to problems, not manipulate GR equations by hand.

I'm ignoring AGI/singularity type events, just opining about the current tooling.

Yah, the transition will be bumpy. But we will learn the skills we need for the new tools, and the old skills just won't matter as much. When they do, yah, it'll be a bit more painful, but so what, we gained so much efficiency we can afford the losses.

> AI means that those 'easy' tasks can be automated away, so there's less immediate value in hiring a new grad.

Not disagreeing that this is happening in the industry but it still feels like a missed opportunity to not hire juniors. Not only do you have the upcoming skill gap as you mention, but someone needs to instruct AI to do these menial/easy tasks. Perhaps it's only my opinion but I think it would be prudent to instead see this as just having junior engineers who can get more menial tasks done, instead of expecting to add it to the senior dev workflow at zero cost to output.

“automate it away” ironically still requires a human in the chain to determine what to automate, how, and to maintain that automation. Whether it be derived from an ai or a systemd script or an Antikythera mechanism. Now if you leave that to seniors you just ate a big chunk of their day playing shephard to a dozen plus “automated” pipelines while they still have stuff to do outside the weeds. Now you need more seniors and pretty soon they want triple what you could pay a junior and I don’t think they are 3x more prolific if the junior is managed efficiently quite frankly.

Anyone reccomend an analysis, article or book or video, of this effect on the blue collar industry decades ago?

This assumes there will still be a demand for software developers in 5 years. I believe we'll be out of jobs much sooner than that.

We might need a lot of young adults for war in the near future, according to some.

> I feel the effects of this are going to take a while to be felt (5 years?);

Who knows if we'll even need senior devs in 5 years. We'll see what happens. I think the role of software development will change so much those years of technical experience as a senior won't be so relevant but that's just my 5 cents.