Comment on (LLM) A language model built for the public good
thedruid@lemmy.world 1 week agoThat’s not research. That’s simply surfacing tidbits it found on the net the happen to be true
.I’ve asked many questions of many llms in my chosen areas of interest and modest expertise , seeking more than basic knowledge( which it often surprisingly lacks ) it always has at least one error. Often so subtle it goes on noticed until it’s too late.
Plebcouncilman@sh.itjust.works 1 week ago
So what you’re saying is that it’s good for research, because you can’t research what you don’t know about.
It’s good for giving starting points which is exactly what I meant.
Next time I’ll write a dissertation with hyper specifics because it seems it’s necessary every time LLMs are involved as there’s always someone looking to nitpick the statements.
thedruid@lemmy.world 1 week ago
No you rude fuck.
If i ask a simple question about a subject, let’s say foraging as a I do that a lot. And it’s wrong, it’s friggin wrong.
I’ll ask about a specific plant. Full disclosure this is one of my things. 40 years at it. Ok? No big stretch to think I know a thing or two.
So I ask about let’s say, Japanese barberry. An invasive plant that is hated by many and rightly so at times. The question is , is it edible?
The answer given was no. The truth is the opposite. It is edible. Hell there’s recipes online for barberry jam. Now don’t go just eating them though. Smart to test one or two leaves to see if an individual is allergic. That’s not part of the answer, that’s foraging 101. But I digress. The a.i was wrong and then argued about it until I pulled up all of the evidence. The a.i then admitted it was wrong, but who cares? It’s not alive. Winning an argument with a.i is like beating oneself at poker.
Another example I’ll ask about intervals in music ( guitar teacher as my main profession now, as my passion for 48 years). It got the major scale intervals wrong.
I asked ask if yogurt can replace eggs as a binding agent to one of them (can’t remember which, apologies ) and it said no. That’s a friggin home ec tip that’s been around for at least a century
People who give dissertations don’t brag about it. Especially to make a point in a thread. It only makes one seem like a person who isn’t confident in what they’re saying so they drop a line that they feel will impress others. It doesn’t
Others experience is as important, vital and real as yours, regarding the answers given by a.i , but you’ll brush it off because you feel that some how you have more insight than others. You dont. You just have more time to pour through a.i’s mistakes to massage it to getting something close to what you want. That shows an abundance of time available. Which means you aren’t doing the things I’m talking about.
Or it means this is something you do for your job and it works for those specific needs . Which is fine, but your needs are not the world’s. My need have been poorly met by that tool you spouse. Much like a rake won’t help a guy digging a hole, a.i is the wrong tool for most jobs.
Which means your opinion of my evaluation of a. I results is skewed because you don’t value others experience, no matter how intelligent you are. And that is a sign of ignorance.
I wish you a good day
sykaster@feddit.nl 6 days ago
I use perplexity as my main go to if i want to use an LLM, since they have access to a wide scale of models. It was correct in the cases you mentioned. It’s a tool focused on correctness of information and I’ve had it hallucinate a lot less than other tools.
Give it a shot if you’re looking for one that focuses on correctness of information. It searches the Web and then feeds the results into the model you choose.
You can also tell it to only use academic papers, social discussions, or SEC filings.