Some AI levity...................
- Thread starter garczar
- Start date
AI is not up to speed in matters pool.
Search on the Rifleman and you won't get Buddy Hall, you'll get an old TV show. Search on Earthquake and you won't get Keith Mccready but seismic activity instead. Search on the Kentucky Colonel and you'll get fried chicken, not Nick Varner.
What the world needs is a pool-oriented AI.
Search on the Rifleman and you won't get Buddy Hall, you'll get an old TV show. Search on Earthquake and you won't get Keith Mccready but seismic activity instead. Search on the Kentucky Colonel and you'll get fried chicken, not Nick Varner.
What the world needs is a pool-oriented AI.
Last edited:
So...did you tell it about the error?
What AI did you use? What did you ask it?
I just tried Gemini, with no preliminary prompts (it does not know who I am). It did not give an erroneous picture.
I also tried a version of Claude not publicly available. Again, no initial prompts. No ID error on Harris.
I asked "Are Bluegrass cues were still made?"
I ran your username through a while back and it was pretty accurate about your forum posting. Said you were an authority on cues, blah blah blah …but I laughed when it said “beware though, if you buy a cue signed by Chopdoc or have Chopdoc sign your cue it will decrease the value of it.” Lol. I don’t know where it came up with that. Or specifically why it felt the need to add that.
It explained further that collectors prefer cues not to be signed by anyone other than the cue maker or something. Which is accurate but I don’t think I’ve ever heard of you signing cues.
Hmmmm...that gives me an idea...
Me? Sign a cue?
Only if you want to destroy the value I guess.
By the way, if you didn't use incognito mode, and you used an AI that is "personalized" (you may not even know), or if you have it installed as an app, and had researched any cue signatures, that would be why it made that comment. It is anticipating what you want to know.
Unless I want it to know who I am I generally will use an AI anonymously.
I have AI set up for medical teaching, it has all my credentials and teaching methods set up with special prompts. Essentially, it acts as a teaching assistant. I don't use it for real case consults. And I don't use the OpenEvidence AI app that is very popular with clinicians today.
I also have AI set up for anonymous use. I have access to a powerful AI that the public does not. I also have access to an AI for research purposes that I can't really play with but it is amazingly powerful. And I have an AI agent installed locally...sometimes it tries to get out...
I dont trust Al or GPS. Been led down many wrong roads or dead ends.
FAKE! Nobody can read a doctor’s signature, not even other doctors or pharmacists.
You know the drill. Travel, lay it down, watch the money light up.
Posting for a friend.
Yeah. I usually check both regular and incognito. I also have a subscription to a different one that I mostly use professionally.By the way, if you didn't use incognito mode, and you used an AI that is "personalized" (you may not even know), or if you have it installed as an app, and had researched any cue signatures, that would be why it made that comment. It is anticipating what you want to know.
Unless I want it to know who I am I generally will use an AI anonymously.
I have AI set up for medical teaching, it has all my credentials and teaching methods set up with special prompts. Essentially, it acts as a teaching assistant. I don't use it for real case consults. And I don't use the OpenEvidence AI app that is very popular with clinicians today.
I also have AI set up for anonymous use. I have access to a powerful AI that the public does not. I also have access to an AI for research purposes that I can't really play with but it is amazingly powerful. And I have an AI agent installed locally...sometimes it tries to get out...
Using the consumer free ones like Gemini or copilot are kind of like being a detective investigating a murder with 22 witnesses who all saw something different and 5 different medical examiners who have slightly differing opinions. You have to do a lot of sifting through bs.
GPS?
The Crazy Redhead confuses right and left. She'll say "turn right" and point left. I can't tell you how many times I have slammed on the brakes mid-turn when she yells "no! the other way!".
I don't trust AI or GPS...or The Crazy Redhead.
Using AI and trusting it are two entirely different things. Same as marriage.
GPS was set up to commit mass murder when I was using it years ago. A state highway went with no turnoffs for twenty miles or so then right down a ramp into a lake!
It did save some tickets though. Got stopped for speeding. When I showed them the GPS and told them it was accurate to a tenth of a mile an hour they asked, "It can tell you how fast you are going?" "Yep, argue with it and you are arguing with satellites and the federal government!"
The GPS is terrible about sending anything over ten or twelve feet tall down impossible routes too.
AI is still a ways from being ready for prime time. I used to play with people leading them down a logical path. Six steps or more, every one of them unquestionably true as stand alone statements but they led to a totally ridiculous answer. I don't trust AI not to do that yet.
Hu
Internet searches are just as bad. Many people know things that are wrong because they found it in a search.
Using internet searches is a skill. Some people make a career out of it. Using AI takes skill. If you don't know how to prompt AI it can be way off base. And of course, it can hallucinate. But people doing internet searches hallucinate as well.
I commonly see people standing on some point, then they point out something they found on the web, like a research article, and it actually contadicts what they are saying yet they think it supports their notion. This is most often confirmation bias.
I don't "trust" AI at all, not in the sense that most people mean "trust".
It's a tool, know your tool, and use the right tool for the right job.