Dr. Ross- Thank you for visiting Conduit in TN a while back, and for signing my copy of Rescuing Inerrancy. I am really enjoying it just as I have the few other of your books so far. I was raised on evolution in school, found YE creationism in college and laughed at it, then came to faith as an adult and just accepted it on faith, despite doubts. Now, because of you, I can be confident in all the universe is telling us AND in faith accept a creation model that makes sense without violating the Word.
I appreciate Jeff's spin on AI, but I'm concerned that he (and the researchers he cites) aren't considering the fact that it is highly unlikely that "AIs" (really LLMs) know that they are lying when they give an answer that humans perceive as lying. In my experience in using AIs, the AIs don't know facts, or even really strategies. What they "know" is that certain statistical weights lead to outputs that their creators (humans) think are correct (Hugh's point about knowing what the goals of the models are). Forgive my language, but in my experience LLMs are really good "B.S.ers" as opposed to oracles of truth (or lies). It's the humans interpreting the results (and tweaking the models) that know things, not the models themselves. So using Jeff's example of the drug deal, when the researchers asked the model why it gave the answers it gave, it didn't "lie" to the researcher when responding, it gave the statistically best B.S. answer it could based on the input it was given -- it was the researcher reading the result that determined it was a lie. The point of what I write above is to say that we have to be very careful anthropomorphizing these models, in the end they are machines bound to follow their programing, not intelligent rational beings. So in my opinion claiming the machine is "lying" is philosophically inaccurate, it is simply giving the statistically most probable output -- which happens to be a lie.
@@DuelScreen I stopped watching a few minutes in and then commented. I just couldn’t watch what I feel I already know by gut instinct. So if that was said verbatim than yes, that’s exactly what I’m saying.
Dr. Ross-
Thank you for visiting Conduit in TN a while back, and for signing my copy of Rescuing Inerrancy. I am really enjoying it just as I have the few other of your books so far.
I was raised on evolution in school, found YE creationism in college and laughed at it, then came to faith as an adult and just accepted it on faith, despite doubts.
Now, because of you, I can be confident in all the universe is telling us AND in faith accept a creation model that makes sense without violating the Word.
Love hear a debate between Hugh Ross and Neil DeGryse
So interesting, the origins of our Universe.
Really appreciate hearing some very intelligent Christians insights pertaining to AI
I appreciate Jeff's spin on AI, but I'm concerned that he (and the researchers he cites) aren't considering the fact that it is highly unlikely that "AIs" (really LLMs) know that they are lying when they give an answer that humans perceive as lying. In my experience in using AIs, the AIs don't know facts, or even really strategies. What they "know" is that certain statistical weights lead to outputs that their creators (humans) think are correct (Hugh's point about knowing what the goals of the models are).
Forgive my language, but in my experience LLMs are really good "B.S.ers" as opposed to oracles of truth (or lies). It's the humans interpreting the results (and tweaking the models) that know things, not the models themselves. So using Jeff's example of the drug deal, when the researchers asked the model why it gave the answers it gave, it didn't "lie" to the researcher when responding, it gave the statistically best B.S. answer it could based on the input it was given -- it was the researcher reading the result that determined it was a lie.
The point of what I write above is to say that we have to be very careful anthropomorphizing these models, in the end they are machines bound to follow their programing, not intelligent rational beings. So in my opinion claiming the machine is "lying" is philosophically inaccurate, it is simply giving the statistically most probable output -- which happens to be a lie.
NO AI !!!
Hugh, if that galaxy is 13 billion years old, do you think it still exists today or will it have collapsed or dissipated?
it is written: "world without end".
So are the population III stars in our galaxy 12+billion years old?
I don’t think it’s less likely that Meta is lying to us.
Christ is Lord! ☦️
If you think this is going to be good you’re deluding yourself
This is how you tell me that you haven't actually watched the video without saying that verbatim.
@@DuelScreen I stopped watching a few minutes in and then commented. I just couldn’t watch what I feel I already know by gut instinct.
So if that was said verbatim than yes, that’s exactly what I’m saying.