↑
logo

Ask HN: Was I proven correct SWE is dying?

Posted by blindprogrammer |an hour ago |2 comments

kooi an hour ago

It's dead only if you want it to be dead. It's not that complicated.

I.e: You're a carpenter who is very good at using a hammer and doing timber framing. Then, nail guns arrive and your crew starts doing stick framing.

A few options:

1. You're old enough to retire 2. You cry, don't re-skill, and be grumpy 3. You re-skill and use your know how to be even more effective

Remember the E in SWE: Engineering. This involves way more than coding, button pressing. You're interfacing with clients, you're designing interfaces, you're debugging, designing systems, coming up with new projects, guiding LLMs.

If you're not doing E, when you're not a SWE. Maybe just a software developer.

serhei an hour ago

SWE as a job category might be dead, in terms of unrecognizable changes relative to the norm. One can probably look forward to the "are you frontend or backend?" question dying out, for one thing.

Software 'engineering' is only beginning because (almost) no one has actually applied AI to making 'well-engineered' code rather than 'eh, it breaks infrequently-enough for business purposes'. The methods for producing 'perfect' software by hand are known. If LLMs help produce software faster, they should be able to help in producing 'perfect' software faster, rather than advocates attempting the 'you don't want bug-free code' Jedi-mind-trick. (Alternatively, we would need to develop much more detailed collective wisdom about where LLMs are and aren't appropriate.)