Attacking people who disclose when they’ve used AI won’t bring the future you desire, it will only encourage people to not be forthcoming about their sources. I could have come here and dropped some paraphrased version of Gemini’s response and passed it off as my own, and no one would’ve been the wiser. I use AI sparingly, and when I do, I make sure to be clear about it.
I get that AI isn’t a favored technology around here, but with the enshittification of search engines, niche answers aren’t easy to surface anymore. It seems OP has done as much searching as he was willing to do, and I figured pursuing traditional search results on my own wouldn’t yield anything better, so I searched with AI.
You can call it slop if you want to, but that doesn’t mean it’s necessarily wrong. An LLM is a tool that groups words together by probability from training data. If that data includes documentation or forum posts that discuss the location of the flasher relay module in a '96 Sonoma, there’s a decent chance the LLM is going to spit out a worthwhile answer.
What are the stakes?
Job impact? OP clearly wasn’t going to bring this to a mechanic.
Environmental? An AI search has less environmental impact than a burning vehicle from a rigged bandaid fix.
Safety? This particular AI query hasn’t produced any dangerous suggestions. It’s exploratory.
Waste of time and effort? The LLM suggested a place that takes very little effort to verify for someone who has the truck in question. I don’t, so I can’t. If the information is correct, it gives OP the opportunity to implement a proper fix and undo the fire hazard he’s rigged up. If it’s not, no one’s in a worse position than they were before.
I use AI sparingly, and when I do, I make sure to be clear about it.
How about instead, you just don’t use it at all.
And especially don’t inject AI slop into a conversation with other people when nobody asked you to. You’re not adding anything of value to the discussion. If you have nothing of your own to say, it’s okay to say nothing at all. Go ahead, try it. Say nothing at all. Easy, isn’t it?
You’re not adding anything of value to the discussion.
If that’s your complete, honest take, I don’t see eye-to-eye with it. It was correct information, posted before anyone else came here with the same information. That’s the definition of adding to the discussion. That I sourced that information from an LLM and that I couldn’t independently verify on my end without access to the truck in question is irrelevant. Fuck AI for other reasons, but sometimes it’s useful.
“My confirmation bias confirms that it worked out.”
This is no different than writing, “the first ten Google results for Sonoma flasher are…” It’s the most basic, non-contributing information that anyone would get with even cursory investigation. It’s fake-work that doesn’t contribute anything, but gives you the warm-and-fuzzies for “helping”, while most likely muddying the waters with misinformation.
Dude, look at the vote totals on your posts here. Nobody sees eye-to-eye with you on it. Nobody wants it.
And it’s up to the people you’re ‘helping’ to determine if your slop is useful or not. Pretty obviously, nobody here thinks it’s useful.
If you don’t have anything worthwhile to say without AI, then just don’t say anything at all. If OP wanted AI slop answers, they could have just googled it and got google’s AI slop directly. In all likelihood, they already did. Who would ask a question like this without googling it first? And google always throws this shit in your face every time you search for something.
Next time you want to add AI slop to a conversation, just don’t. Nobody asked you to. Nobody wants you to. We’re here to talk to real people. We don’t want to talk to AI.
Then don’t post it.
Just don’t.
Nobody wants your AI slop.
Guess what? Despite the fact that I loathe AI, it turns out that’s exactly where the OG flasher is installed.
I’ll fix it properly in the next day or three. I got the right part, just didn’t know where it went…
Attacking people who disclose when they’ve used AI won’t bring the future you desire, it will only encourage people to not be forthcoming about their sources. I could have come here and dropped some paraphrased version of Gemini’s response and passed it off as my own, and no one would’ve been the wiser. I use AI sparingly, and when I do, I make sure to be clear about it.
I get that AI isn’t a favored technology around here, but with the enshittification of search engines, niche answers aren’t easy to surface anymore. It seems OP has done as much searching as he was willing to do, and I figured pursuing traditional search results on my own wouldn’t yield anything better, so I searched with AI.
You can call it slop if you want to, but that doesn’t mean it’s necessarily wrong. An LLM is a tool that groups words together by probability from training data. If that data includes documentation or forum posts that discuss the location of the flasher relay module in a '96 Sonoma, there’s a decent chance the LLM is going to spit out a worthwhile answer.
What are the stakes?
Job impact? OP clearly wasn’t going to bring this to a mechanic.
Environmental? An AI search has less environmental impact than a burning vehicle from a rigged bandaid fix.
Safety? This particular AI query hasn’t produced any dangerous suggestions. It’s exploratory.
Waste of time and effort? The LLM suggested a place that takes very little effort to verify for someone who has the truck in question. I don’t, so I can’t. If the information is correct, it gives OP the opportunity to implement a proper fix and undo the fire hazard he’s rigged up. If it’s not, no one’s in a worse position than they were before.
How about instead, you just don’t use it at all.
And especially don’t inject AI slop into a conversation with other people when nobody asked you to. You’re not adding anything of value to the discussion. If you have nothing of your own to say, it’s okay to say nothing at all. Go ahead, try it. Say nothing at all. Easy, isn’t it?
If that’s your complete, honest take, I don’t see eye-to-eye with it. It was correct information, posted before anyone else came here with the same information. That’s the definition of adding to the discussion. That I sourced that information from an LLM and that I couldn’t independently verify on my end without access to the truck in question is irrelevant. Fuck AI for other reasons, but sometimes it’s useful.
“My confirmation bias confirms that it worked out.”
This is no different than writing, “the first ten Google results for Sonoma flasher are…” It’s the most basic, non-contributing information that anyone would get with even cursory investigation. It’s fake-work that doesn’t contribute anything, but gives you the warm-and-fuzzies for “helping”, while most likely muddying the waters with misinformation.
Dude, look at the vote totals on your posts here. Nobody sees eye-to-eye with you on it. Nobody wants it.
And it’s up to the people you’re ‘helping’ to determine if your slop is useful or not. Pretty obviously, nobody here thinks it’s useful.
If you don’t have anything worthwhile to say without AI, then just don’t say anything at all. If OP wanted AI slop answers, they could have just googled it and got google’s AI slop directly. In all likelihood, they already did. Who would ask a question like this without googling it first? And google always throws this shit in your face every time you search for something.
Next time you want to add AI slop to a conversation, just don’t. Nobody asked you to. Nobody wants you to. We’re here to talk to real people. We don’t want to talk to AI.
Slop
I gotta give you this much though, you apparently noticed that I installed this bodge fix with an easy Undo button.
All I gotta do is pull the bodge out and plug a 15 amp fuse back in, if I ever manage to find the original flasher module anyways.
This works perfectly for now. Easy Undo…