The Eternal Sloptember | Hacker News Reader

The Eternal Sloptember

geohot.github.io

21 pointstherepanic7 comments

I've observed a strong correlation between the level of fundamental knowledge of computers (Linus, James Gosling, OP etc) and the level of AI psychosis. Guess which direction it is.

I increasingly agree with the spirit of this. I'm not even convinced that they'll never be able to "program" I just find that these tools are to midwits what flame is to a moth and destroy the systems-thinking that drives good engineering (as opposed to merely coding)

You have to be good enough to know what to ask for but not good enough to ask yourself what is needed next, or to say no to it - for example, you can fairly easily ask an agent to write small changes and so on, almost no one does.

Management are also seem the most prone to psychosis (IMO at least) e.g. if you are a once-technical manager it probably feels great to kind of being doing programming again with a minion who doesn't eat lunch?

I am yet to meet more than probably 2 vibe coders who uses their newfound "productivity" to try _new_ things as opposed to just pump out correct-looking slop - the problem here isn't just that there is now more code, but also that for some strange reason people seem to choose to automate the thinking as well as the coding.

I am yet to see any AI driven productivity around me. They can definitely turn a days work into a lot less time, this is magical, but I'm yet to see any "real" projects be delivered in a timeframe that's particularly different to what they would've taken previously.

When people say “you’re holding it wrong” it tells me they can’t even conceive of a better way of doing it.

The models produce the same slop for everybody, you don’t have a special way of doing it you lack taste and an opinion on your problem domain from lack of research and studying prior work.

Is Apple pushing their engineers to use AI? I thought that was just Microsoft.

On an individual level you still have control over your slop, but just imagining what it must be like in a megacorp where your colleagues are contributing to it, it must be a nightmare.

You're using it wrong.

I have tried all the different models, different harnesses, different prompts. It’s not this. The people who say this would probably say the same thing about slot machines, you see, you have to bet 5 lines after you get a cherry no wonder you aren’t winning!

What i meant was: your experience might not be representative of the coding models capabilities. At least based on the poor examples in the post, you’re just scratching the surface of learning how to extract value from them, and, as i understand, you gave up and shared your experience. Given that you fully understand how an LLM works(and you do) it might also be that you’re using it outside training data. Not all code is equally represented(quality/quantity) in the training data, like there’s not the same amount of latin text as opposed to english text there(it’s all gravy to an LLM). You also suggested that manual might be faster and better quality - in MY experience, this is just you (using search through prompts), using it wrong.

Edit: added gravy pun

Reply on news.ycombinator.com