[2muchcoffeeman]

This is such a weird comment to me. These are trivial functions that could be faster to code than describe. I’d hope people aren’t relying on AI for functions as simple as clamp.

I mean, people have been writing these simple functions over and over for decades when they just needed one or two things that importing a library wasn’t needed. I wasn’t aware there was a gap to be filled.

[wavemode]

Clamp is a simple enough function that if you simply typed its name and arguments, copilot would autocomplete its body. No prompting needed, in that case.

[lenkite]

Man, people are prompting and copying LLM output for min/max functions. I saw my junior do this in a meeting - it was apparently too much work to lookup the standard library. When most people work with a LLM, the thinking part of their brain gets switched off for some reason.

[hawk_]

Not clamp per se, but there are enough tiny snippets which can have off by one or other issues not relevant to when you think about the problem description.

[qazxcvbnmlp]

I agree:

Senior engineer: "you shouldn't have to use llm to do simple things" Their boss: "hey I need you to write this thing in Go for performance reasons" Senior engineer: "you shouldn't make me learn new languages, js is plenty performant for this task just install a new server. why are they making us use all these new tools? everything is new again, blah blah excuses excuses"

vs.

Boss: "hey can you write this in go" Vibe Coder: "write this loop to ignore the first semicolon in this response", and then proceeds to move on with life.

It seems like when people evaluate the merits of what a code generation ai agent should and shouldn't do, they leave out a lot of their implicit assumptions about how project should work and their own theory of mind around coding.

[2muchcoffeeman]

Sure. Then all the common tools that you need should be bundled into a library. I should not be importing 1000 string manipulation libs with a single function each.

[whatevaa]

Like lodash?