I'm not a lawyer and I went "Whaaaaaaaaa!?" so many times. This is absolutely hilarious and bad, like REAL bad. Can't wait for @questauthority.bsky.social and @kenwhite.bsky.social 's reactions.🤣 - ThreadSky

dharhix.com • 60 days ago

I'm not a lawyer and I went "Whaaaaaaaaa!?" so many times. This is absolutely hilarious and bad, like REAL bad. Can't wait for @questauthority.bsky.social and @kenwhite.bsky.social 's reactions.🤣

Reposted from Judge Roy Ferguson

I asked ChatGPT to brief "Trust Me v. Bro," a nonexistent case I used in response to a claim that items/information publicly released by the media is inadmissible at trial for some reason, and of course GPT was quick with the full cite and (glorious) synopsis: Trust Me v. Bro, 999 U.S. 867 (2024).

Comments

wesmorgan1.bsky.social•60 days ago

Oh, this has already happened in the real world.

https://www.seyfarth.com/news-insights/use-of-chatgpt-in-federal-litigation-holds-lessons-for-lawyers-and-non-lawyers-everywhere.html

iquanyin3.bsky.social•59 days ago

omfg ☠️ thank you for the best laugh i’ve had all year!

activecultures.bsky.social•60 days ago

"The sanctity of bro promises cannot hinge on proprietary spice blends. A handshake remains a handshake, and sauces are an implied term."

🤣🤣🤣🤣

dharhix.com•60 days ago

And from HIM... Yeah, lol

ericheilmann.bsky.social•60 days ago

That was the line that got me😂

activecultures.bsky.social•60 days ago

It’s all so spot on. I kinda think it’s HI

madammem.bsky.social•60 days ago

It reads like ChatGPT has now learned how to troll humans.

dharhix.com•60 days ago

Who knows, maybe it does have a sense of humor but it keeps it to itself? Maybe it's scared to tell us it's Rickrolling us?😂

eduardodinero.bsky.social•60 days ago

If you are a lawyer using ChatGPT, you are likely committing malpractice. Have fun with that.

andrewctenor.bsky.social•60 days ago

“Justice Thomas grilled both parties,” even if probably accidental, is a hilarious line

dharhix.com•60 days ago

At least it didn't say his answer was spicy.🤣

tggp.bsky.social•60 days ago

Thomas talking at all during oral argument is how you know it's made up. The rest is 100% credible though.

infinitelyloopy.bsky.social•60 days ago

This is where AI gets *really* dangerous: its tendency to return utter nonsense paired with a broader audience that accepts its output as true.

I get wary of using it for subject matter that I can’t easily verify/stress test

dharhix.com•60 days ago

I've used it once while I was trying to come up with an algorithm in programmation and I had to iterate so many times the prompt. It did give me the base of what I was looking for, but the wall on which it built the rest was useless.

infinitelyloopy.bsky.social•60 days ago

I’ve seen instances of this where the code it outputs contains so many bugs you spend more time trying to fix it than you might taking an incremental approach with StackOverflow as a resource

IME, it works best with micro increments (a pipeline job) vs the macro (whole CI/CD pipeline)

dharhix.com•60 days ago

I have no doubt on this.

My problem was that I didn't know how to handle the basis of the problem. I think the writing of the prompt served more as the proverbial goldfish than anything else.

haveyouanywool.bsky.social•59 days ago

And it’s all built on theft of other peoples’ copyrighted words, that is, their intellectual property.

audiolibrarian.bsky.social•60 days ago

i have had SO MANY pro se patrons call the Ref Desk, excited that they had found THE case that told them EXACTLY what they wanted to hear.

they give the cite so i can help them locate it, it's a different case.

i ask them where they got the case and they answer: ChatGPT

scarfy.bsky.social•60 days ago

I can't wait for @boozybadger.bsky.social's reaction.

blackcap.name•60 days ago

The biggest problem with artificial so-called intelligence is that it is incapable of saying “I don’t know.” AI will just make up bullsh*t. There is even special term for this in the AI field: hallucination.

dharhix.com•60 days ago

It's like their dad, I guess? Techbros that can't admit they're wrong or don't know.

Comments

Posting Rules

Reply