Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I feel my job in the future will be more secure than ever. Tons and tons of AI generated garbage code (trained on more and more existing garbage code) that the „developer“ will at a certain point no longer be able to maintain or fix. Not even speaking about trusting the output. Feels similar to all the outsourced development to cheap suppliers that inevitably collapse or create horrible maintenance overhead.


Do you feel you would be able to “maintain and fix” LLM-generated 100-megabyte source code blobs? And if you could, do you think it would be a job you’d want to do?


I spent 2000-2015ish as a Perl programmer, which meant I spent most of that time working for a small number of e-commerce and SaaS companies who had large codebases that had mostly been written around 2005 with some hair-raising tech practices. Nursing a nasty codebase to health bit by bit can be very rewarding.


As long as you don't have pressure to just ship X Y Z as you are doing it, in a team of 20 so any nursing you do is quickly undone by the team and its pressure to ship some half thought out feature.


> Perl programmer

Perl programmers are as always ahead of the curve. Writing code that looks like LLM slop before LLMs!


No, writing a replacement. Either part by part or a whole new software system.


Secure, sure. But this sounds like a crappy job.


Not so crappy if the pain is big enough for the customer. Once a certain threshold is reached they are often very open to changes and giving freedom to develop a proper solution.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: