Live data from Hacker News

Going full AI engineer, not touching code anymore

max.gp

121–125 of 125 posts

Re: Going full AI engineer, not touching code anymore

#121

What's interesting to me is how many people have found out with the LLM boom that they don't want to be developers/engineers, they want to be managers, delegating all the work to another entity and checking in occasionally to see how the coding is coming along

You seem to think you can only be an engineer if you are holding the shovel yourself. Building software consists of many parts. I love each individually. I was always a bit dissatisfied about how they interact and how they block each other. I love puzzle solving, I love debugging, but I kinda hated building software because it depended on those elements and I'm not fond of context switching. I can love building softw…

Yeah, it's very clear that the person that wrote the above comment didn't actually grok what the purpose of the article was. Thanks for providing a counterpoint that I think actually summarizes the article correctly.

Re: Going full AI engineer, not touching code anymore

#122

What's interesting to me is how many people have found out with the LLM boom that they don't want to be developers/engineers, they want to be managers, delegating all the work to another entity and checking in occasionally to see how the coding is coming along

I think it's a symptom of software jobs paying really well for multiple decades. It attracted many people who are in it for the money only, and not the process or result (nothing wrong with that, imo). These are the people who fully embrace the vibe coding model and enjoy managing an AI more than developing anything by hand.

Re: Going full AI engineer, not touching code anymore

#123

People bragging that they "dont touch code" and only "argue" with agents are reinventing the slowest possible IDE. Obviously the agents are great at producing large chunks of code, but they often make minor and sometimes trivial mistakes which need amending. Typing something like "in src/auth/session/token_manager.ts the refreshTokenExpiry variable should be refresh_token_expiry. update every reference and make sure…

this seems like a strawman... of /course/ you'd use the IDE to do ref renames?? AI isn't for that...

If that is true then you cant claim to be "not touching code anymore".

Re: Going full AI engineer, not touching code anymore

#124
post #62

Earlier quoted context omitted.

Have you tried drilling into the reasoning when this happens? This is why I usually leave it in "Plan Mode" and when it proposes a solution that seems unusual or unexpected, I point out why I think it is and ask it to justify it's position. Sometimes I get the "you're right!" response, but often it will also explain why it made the decision it did, and it's rational enough that I accept the new approach. It's still v…

> Sometimes I get the "you're right!" response, but often it will also explain why it made the decision it did, and it's rational enough that I accept the new approach. I think this is kind of what I’m worried about. Referencing Karpathy, an LLM can basically convince any one of anything. Doing this enough times and your opinions just become the LLM opinion. Same with problem solving approaches. I see this for myself…

> LLM can basically convince any one of anything This and revision to the mean I think are real problems

I've found myself on more than one occasion just stepping away and needing to really sit down and think and to not ask an LLM until I have a clear idea of what I want to do

The one thing that bothers me is that one process I really used to run is thinking in code, slowly sketching out a solution until it had the properties I wanted it to have

I find myself not really engaging with this at present and it really bothers me, I've been trying to figure out how to get back into doing it again because I think the most differentiated and interesting ideas and thinking I had were usually shaped by going through that process

However so far I've found working with LLM's just doesn't jive with that thought process and several months into using them I'm still figuring out how to go back to doing that

Re: Going full AI engineer, not touching code anymore

#125

    > This is the part I struggle with when I think about it. Years of identity rests on “I am someone who writes code”. That identity is now contingent on a tool stack that’s three years old and could regress. If it breaks, the identity goes with it, and I’m fine with that.

    > I think that’s the most honest signal I can give about what I actually valued. The thing I was doing all those years wasn’t writing code. I was making decisions and watching them become real. The agents make that loop shorter. Take the agents away and the loop becomes too long again. I’d rather do something else.
What's really interesting to me and I'm honestly curious as to how many people are in the same boat as I am vs the author is that I have the same feeling and I'm not ok with it

I think if LLM's disappeared tomorrow I could go back to coding without them, however there'd be that transition of scraping the rust off and honestly I don't like that

It's really discomforting to feel like the codebase I'm working on is getting ever more foreign and strange, especially as LLM's when they communicate with you are persuasive that they are making things better, and they do work, for a given value of work anyway, it's really hard at times to tell looks right from is right

I've come up with some strategies that mitigate this to a degree and I'm slowly trying to find more

I think a lot of people are taking refuge in practices, but aren't entirely coming to terms with the actual scale of the difficulty of writing correct code, skating by on it being useful, which don't get me wrong isn't a bad thing, our clients pay us for creating useful things

Not turning flops into heat

However I get the uneasy feeling like our tools are very inadequate

Types measure consistency, not correctness

Tests and code coverage are very easy to game, and unless you've put in the discipline of creating something on the order of sqlite's test suite, there's likely to be an ever growing pool of latent footguns sitting in the code you're making

And I'm not saying this to be negative or alarmist, at the end of the day it depends on what your goals are, and most working devs and startup folks are just trying to ship

And we can't really argue that for many, they do help do that

You will ship

Something at least

I find myself in this odd position of acknowledging that the coding world has shifted in ways that I find exciting and strange, and as much as I've begun to understand how to navigate this new environment, the part of me that felt deeply grounded in the things that I built is still feeling disquiet

Post reply on HN