Learning curve? I was promised there wouldn’t be a learning curve
I’m getting there. I have a very specific task for our robot masters: I upload a picture of a specific real-world building and you put it in a night-time landscape with a full moon. It’s for work.
What I’m getting: I take your building and I turn it into a gigantic Disney fairy princess castle. Also full moon.
I mean, it’s making some astonishing images, but they aren’t what I need.
Finally, I got mad and typed in “that’s the wrong building, you stupid AI” and it gave me this.
Huh.
Uncle B said, “wouldn’t it be easier just to do the illustration yourself?” and I’m, like, “SHUT UP!”
Posted: April 9th, 2025 under personal.
Comments: 5
Comments
Comment from Some Vegetable
Time: April 9, 2025, 9:18 pm
I, personally, am having trouble trusting both Grok and ChatGTP. I asked it a couple questions about Formula 1 Racing – which I thought would be common enough for it to have knowledge. I asked who is likely to win the 2025 championship.
It suggested Lewis Hamilton. Fine, but it said he is (still) racing for Mercedes. I told ol’ Chatty that was wrong he’d left that team. It came back apologized and made another thoughtful prediction that he’ll win, but this time it placed him in a different team (McLaren) which is still the wrong team.
I don’t mind it being wrong as much as I worry about the confident way in which it gives the wrong answer. In this case I was asking something for which I knew the correct answer, and I knew it was wrong. How can I trust in for questions where I do not know the correct answer?
Comment from Rich Rostrom
Time: April 9, 2025, 10:14 pm
Some Vegetable: A friend asked Grok “Who is ?” Grok came back with a lot of mostly accurate information, a fair amount of misinformation, and then said he’d died last year.
The ability of AI to generate plausible falsehoods is terrifying.
Ms. Weasel: “this” looks like a steampunk version of Magritte’s “Castle in the Pyrenees”.
Comment from Durnedyankee
Time: April 9, 2025, 11:15 pm
Now I want to know what all that plumbing is about on the wrong building.
Things should have a purpose!
Comment from QuasiModo
Time: April 10, 2025, 12:03 am
Nothing good is going to come to humans from AI, you should avoid it.
Comment from S. Weasel
Time: April 10, 2025, 3:40 pm
I figure the images are harmless, Quasi. I’m very suspicious of text information, though – I’ve been given bad advice, too.
Write a comment
Beware: more than one link in a comment is apt to earn you a trip to the spam filter, where you will remain -- cold, frightened and alone -- until I remember to clean the trap. But, hey, without Akismet, we'd be up to our asses in...well, ass porn, mostly.<< carry me back to ol' virginny