I’m getting AI-hate fatigue, like there’s a lot of valid reasons to hate AI based on how it’s been implemented and how it’s fucking with markets regarding PC components. But it’s gotten also to a point where, just the minor things in where AI actually has a use in, it gets scrutinized anyways because it’s the norm and cool thing to hate it.
I’m completely cool with that kind of dislike, plus the general annoyance with “AI slop,” but the knee-jerk hatred for LLM’s is where I draw the line.
In reality, they’re sometimes a waste of time, and can indeed contribute to AI slop if used that way, but I’ve also found them to be hugely useful on various tasks, probably with something like a ~70% success rate on the issues I ask for help with. Whatever else, they’re some robust tools, never seen before, and will only get better.
EDIT : haha, and let the haters hate… whatever they think they’re hating. 🙃
I treat LLMs not as god-like oracles of pure truth, but rather as sounding boards to get me to consider things that I hadn’t previously. Like, I was just doing some research on home warranties, and I got some potentially valuable information that I then independently verified.
When I don’t quite know what question to ask about the thing I want to know about, an LLM can open those pathways for me.
I most often use GPT5.2 as an enhanced search engine. It’s very good about understanding nuance in terms of the parameters I’m searching on (way, way better than Google, for example), and can pull together a sort of ‘report’ from many different sources.