Hacker News Reader

Comment by timr | Hacker News Reader

I was just thinking about this the other day (after spending an extended session talking to an LLM about bugs in its code), and I realized that when I was just starting out, I enjoyed writing code, but now the fun part is actually fixing bugs.

Maybe I'm weird, but chasing down bugs is like solving a puzzle. Writing green-field code is maybe a little bit enjoyable, but especially in areas I know well, it's mostly boring now. I'd rather do just about anything than write another iteration of a web form or connect some javascript widget to some other javascript widget in the framework flavor of the week. To some extent, then, working with LLMs has restored some of the fun of coding because it takes care of the tedious part, and I get to solve the interesting problems.

View on HN

I'm with you. I love solving puzzles to make something go. In the past that involved writing code, but it's not the code writing that I love, it's the problem solving and building. And I get plenty of that now, in a post-LLM world.

I spend all day fixing bugs and that's why they pay me -- because for most people it's not an enjoyable task. I'm not denying your experience but I will tell you I think you're an outlier. For most people, fixing bugs they didn't create is the worst part of the job.

Part of the fun is also figuring out the “best” way to achieve a thing. LLMs don’t often propose the best way and will happily propose convoluted ways. Clean approaches are still hard to come up with, but LLMs certainly help implementing them once thought up

Yeah, I agree with that. I still enjoy doing the conceptual and design work that once upon a time would have been the role of an "architect", and that's still required with LLMs.

I think for me, the fun comes from preventing bugs; being able to draw on my experience to foresee common classes of bugs, and being able to prevent them through smart code architecture, making it easier for future contributors/readers to avoid walking into those traps. I'm hoping I'll keep being able to do that.

Code reviews are fun when you can trust the competence of the coder.

Maybe we have changed or the industry changed, but last 5 years I’ve rarely met anyone who took the code reviews seriously. Any feedback, or glaring issues, are swept under “we will address this later”, “we work in MVP way(read:deploy barely working code without any accountability to production”, or just outright take it as a character challenge. I’ve lead a team of junior-level engineers who would fight with you via the product owner/scrum master because of the “priorities” and “deadlines”. And then, you’d be on the call solving the same issues you’ve raised in pull request earlier. Maybe I’m getting old, but, taking accountability or pride in one’s work, is getting only worse with Vibe coding and ai generated garbage

Well, from personal experience, code reviews at the FAANG I worked at were taken quite seriously. (But we were building OS frameworks so it mattered more.)

I find both fun. Writing a web form is still kinda fun even after 20 years. I guess it’s like how some people still play the same video games after years and years while others want something new.

Fixing bugs in my own code is fine. In other's code its less fun.

Why?

That’s the wrong mentality. You and your team own all the code. A bugs peer is your bug too.

Not OP but I’ll give my perspective. I have no problem fixing someone else’s bugs up to a point, but I also don’t want to be the one cleaning up someone else’s mess who didn’t take time to do proper testing but is getting rated higher than me for clearing more tasks last month because they cut corners

Bugs are awesome. I picked my current job at Google because it involved lots of bug fixing and crash investigation.

I hate fixing bugs but I enjoy big refactoring, AI-assisted or not - sometimes you just gotta do it all

100% agreed.

I solve a problem, let the AI mull on the next bit, solve another problem etc.

Reply on news.ycombinator.com