Mastodawn

"FOSS maintainers are unpaid and don't have the resources to manage large open source projects, they _have_ to use AI to get anything done."

I've seen this take a couple of times in recent weeks, and I've got two things to say about it:

1. We've been able to manage large open source projects for decades in the past, without LLMs.

2. Ever thought about how there may be a connection between "FOSS is unpaid" and capitalism producing trillion-dollar AI companies and more and more billionaires?

Show thread

zivi Mar 13

@scy “They have to introduce unmaintainable unreliable slop into their codebase to maintain their codebase” is what that logic resolves to lmao

Show thread

zivi Mar 13

@scy and as mentioned elsewhere in the thread, LLMs don’t make developers more productive, they induce a sort of psychosis that makes them think they’re more productive when it’s the other way around and actual development skills are being lost

Show thread

i’m Mi Mar 13

@zaire @scy what if they need the psychosis against executive dysfunction, maybe this is a bit like medicating their ADHD with nicotine

maybe society does do everything it can to not give people meth
no matter the cost

Show thread

zivi Mar 13

@mi @scy i’m pretty sure whatever you’re referring to is unrelated. the average sloperator is too normie to have executive dysfunction, and LLM psychosis isnt that kinda psychosis

abuse of mind-altering substances is a health concern probably dont do it

Show thread

i’m Mi

@zaire @scy what if executive dysfunction scales, so to say, and happens for a variety of reasons, including even the most basic lack of confidence in the subject? what if ADHD is more prevalent in one way or another, and we conservatively threshold our belief of what is productive and appropriate use of ADHD medication — just because we need it to overlap with the threshold for when we call it ADHD? (and I, of course, meant appropriate dosages rather than abuse)

I will not speak specifically on defining the LLM psychosis because I miss too many aspects of it when I try; maybe if someday I'll read any research.