I must admit, I was curious to see, whether an advanced LLM can do what I am trying to do; esp. with similar strict and explicit pointers and instructions.

It turns out, the models cannot even accept the length of the request. Breaking down the task into smaller bits makes it run into rabbit holes and repeated undoings of previous decisions. It is annoying and probably to an experienced, well-versed developer an atrocity to observe.

I can't definitely tell, for me, however, it is not worth it.

Fuck those who replace humans for AI for profit

I mean, that's part of the tragedy of this whole thing: At this point, they're not even replacing people with Ai for profit, they're just replacing them for revenue.

In my experience with LLMs you still need to know how to code even if you’re just using it to write the code for you.

If you don’t know how to solve a problem in code you can’t effectively prompt an LLM to write that code for you unless it’s something very simple.

At that point though you’ve done most of the work so why not just write the code? 🤣

> Fuck those who replace humans for AI for profit

I don't think this has actually happened yet (in the software industry). All of the events that have happened recently are actually a continuation of the post-covid layoffs. It's just that companies now use AI as an attempt to look good, rather than saying they're laying off people because they have budget problems

Sensitive content Reveal/hide