Ask HN: What is the most strategic career path for SDE in response to AI?

2 points by electrondood ↗ HN
I assume Software Engineering will be mostly automated within the next 3-5 years, or at least made completely declarative (supervising models like Copilot) rather than imperative (writing code and designing architecture).

I was surprised to find that Data Engineering/Data Science pays significantly less than Software Engineering.

What is the most strategic/lucrative career move for an experienced software engineer, given that GPTs and other technology are eating software?

5 comments

[ 2.2 ms ] story [ 25.3 ms ] thread
Learn how to use the tools to stay relevant. Humans will be a part of this process for a long, long time. Maybe using AI will be fun!
That's a great question. And one that's really hard to answer with any certainty. I mean, who's to say how much more AI is going to advance in, say, 5 years? Or 10 years?

I'm not at a "freak out and do something nutty" stage with any of this yet (especially since my day-job involves progressively less and less hands-on coding over the last few years anyway). But I'm honestly not joking when I say the thought has occurred to me to attend the local community college and get a new degree in something that seems less likely to get automated away. That could range from something "blue collar" / manual like welding or CNC machining, to just a different "white collar" field like maybe "fire science" or "public safety administration."

I don't know for sure that I'm going to do any of that, but I've gone as far as visiting the webpage for said community college and poking around at what some of the available options are and giving at least superficial thought to planning out how I might approach something like this.

Hoping the industry at large realizes it will be a much more difficult task than it seems to automate things completely. Sure, you can do 80% of it with AI automation, but how does the automation account for the new version of a programming language? How does the automation account for a new processor architecture? The AI is trained on vast amounts of data from real humans' I. If those humans go away, then there's nothing to train it on.

Even more importantly as AI takes over more jobs--and there's a full Ted Talk on this somewhere--humans must ensure that we don't break the chain of passing knowledge down to the next generation that has existed for millions of years, otherwise we may end up as a civilization of adult babies.

Hoping the industry at large realizes it will be a much more difficult task than it seems to automate things completely.

True. But things don't have to be automated "completely" for there to be significant impact in terms of employment. We're probably borderline at a place now where AI could realistically start cutting into employment for the most junior folks in software development. Now project forward another 5 years, or 10 years, or whatever. How much more is AI going to advance? I posit that none of us know the answer to that question.

So yeah, there might still be SDE's in, let's say 5 years. But will there be as many as there are today? I think that's an open question.

Tesla FSD will squeeze every drop out of A.I. for fleet management. Very much doubt Microsoft Copilot is within cooeee hearing distance to where Tesla is at. Compare the CEO track record.