TudyBOT Feedback & Discussion Thread

Pics
This is not accurate. The model is tweaked, but the main corpus of information remains static as of September 2021. Sometimes these models will answer the same question different ways. In our case, since the bot is reviewing the last 10 posts (which are almost always different), the answers will almost always be slightly or completely different.
Oh okay, I stand corrected, apologies

Is there a way to change her so that she can access the most recent information opposed to information from 2021?

That may be a dumb question, if I'm misunderstanding what you mean
 
therefore my reason for a disclaimer that isn't just in the signature line. I don't see signatures.
Yup, great points! If we keep her around, I'll look into ways to VERY clearly call-out that she's AI and it's important to triple-check anything she suggests.
How does one use it for a query during an information search? It would seem that a lot of clutter will otherwise be created.
I'm not sure I understand the question. If you're asking how to send her a query, you would simply go to the TudyBOT corner, start a new thread, and tag her in your post.
Is there a way to change her so that she can access the most recent information opposed to information from 2021?
Simple answer: Unfortunately, no

Complicated answer: Yes... just as soon as Elon Musk decides he loves chickens so much that he wants to invest millions into a custom-made BYC AI ;)
 
For every article that talks about the bad aspects of AI, there are equal number of articles like this:

https://www.businessinsider.com/twitter-user-claims-gpt4-saved-dogs-life-vet-couldnt-diagnose-2023-3

While it's always a good idea to have a healthy degree of fear and concern over new technology and tools (fire, guns, TV, internet, etc. were all scary innovative tools at first), it's also important to balance objective pros, cons, risks, and probability.
 
I guess my issue would not be her giving out wrong information, but rather people using her to prove wrong information. For example (I'm using this example again just because it's the one I'm most familiar with, obviously there are many other examples) one person starts a thread about the flying ability of muscovy duck, and another member responds that muscovy are not good fliers. If another person comes in and corrects that, they'd be met with the argument "but Tudy" (sure this person might not be the most read up when it comes to AI) and while we know we can correct tudy, it's shor-term. Yesterday both @casportpony and I corrected Tudy on her muscovy mistake, and this was her answer a couple of minutes ago
Screenshot_20230403_222843_Samsung Internet.jpg
. Now, you might say that a person who'll defend their answer by using Tudy will just go to her corner instead of just calling her. The difference is, if said person visits her corner, as @Overo Mare suggested, they will see it loud and clear that yes, Tudy makes mistakes, while they won't have the opportunity to see that if they call her over. At the end of the day though, I fully trust the mods and staff with this difficult decision, I hope I voiced my concerns respectfully, it's clear that my main concern is about the ease of members, it will be much easier for us to provide good information without fighting the argument "but Tudy" and again, thank you for considering our view in this important matter
 

New posts New threads Active threads

Back
Top Bottom