(Total Views: 414)
Posted On: 10/04/2023 1:42:05 PM
Post# of 148870
Ohm, your response to my statement, "You do realize that AIs will eliminate the need for all drug trials - mouse, monkey and human? All drug trialing will be done virtually" was a surprise to me.
You say that such a thing would be "a recipe for disaster." Then you proceed to say why. Let's dissect those reasons why.
First off, how could it be a disaster when no one is going to do trialing with AI until it is ready for "Prime Time" as SamaGong says?
I'm surprised at the catagorical nature of some statements about AI trialing, such as it will "never" happen. craigakess says AI can never be smarter than we are and will only "augment" drug trials and AI will never know how a new drug molecule will act in the body.
You say a similar thing, that AI is only as good as the information it is fed and that "some medical papers are flat out wrong."
But there is no Chinese Wall between how humans learn or are "trained" and how AI learns. It takes in information, as humans do, and then correlates and corroborates that information. Non-conforming data is subjected to scrutiny and correction, especially as programmers and users are right there to point out to AI what it may get wrong. AI learns just as we do what is true and not true.
Further, AI can be set to testing things it deems to be true or postulated to be true, just as humans do. AI can do trial and error. Hell, even evolution does that. Those things that are true have "survival value" and are cached and those that do not are discarded. A machine is totally capable of that.
Drug trialing is just chemistry and is as knowable as math, though much more multi-dimensional. If AI gets chemistry wrong, it does so FOR THE VERY SAME REASONS humans would make mistakes - not complete enough information. It can seek to know what it does not already know.
There is nothing human brains can do that AI "brains" cannot do. They are faster, they never forget, they can learn from their mistakes and they can learn new things. Just like humans do. Given time, it might even develop "feeling," "see" colors and enjoy music. These are emergent properties that our ancestral minds did not have until they evolved further.
We have to recognize that there is a time element to all of this, of course. Some people say "never", some people are more flexible and say there will be a "Prime Time" some time, Respert24 says "not in our lifetime."
But AI is absolutely being used today in medicine and other fields as it learns to walk before it can run.
As it can be applied to your list of diseases Leronlimab might treat, every disease has a constellation of cells, organs, receptors and cytokines associated with it, especially CCR5, of course. Just as you have researched those constellations of characteristics, so can AI and it will see patterns and correlations and will corroborate research results about them.
As time goes on it will model biological system such as cells in petri dishes, then mice and, later, even humans and it will do testing of drugs on those models such as Leronlimab. It's all a matter of time.
You say "AI can provide possibilities but it does not contain any actual wisdom or discernment." You will be proven wrong there.
May I cite the wisdom of the "Whopper" computer in the movie "War Games," a terrific movie that everyone should be encouraged to watch.
In the movie, Global Nuclear War was simulated like a game as humans stood by helplessly. After running its simulations, the computer came up with the WISDOM to say that "The only way to win the game is not to play." Yes, a computer can figure that out! That and much more.
You say that such a thing would be "a recipe for disaster." Then you proceed to say why. Let's dissect those reasons why.
First off, how could it be a disaster when no one is going to do trialing with AI until it is ready for "Prime Time" as SamaGong says?
I'm surprised at the catagorical nature of some statements about AI trialing, such as it will "never" happen. craigakess says AI can never be smarter than we are and will only "augment" drug trials and AI will never know how a new drug molecule will act in the body.
You say a similar thing, that AI is only as good as the information it is fed and that "some medical papers are flat out wrong."
But there is no Chinese Wall between how humans learn or are "trained" and how AI learns. It takes in information, as humans do, and then correlates and corroborates that information. Non-conforming data is subjected to scrutiny and correction, especially as programmers and users are right there to point out to AI what it may get wrong. AI learns just as we do what is true and not true.
Further, AI can be set to testing things it deems to be true or postulated to be true, just as humans do. AI can do trial and error. Hell, even evolution does that. Those things that are true have "survival value" and are cached and those that do not are discarded. A machine is totally capable of that.
Drug trialing is just chemistry and is as knowable as math, though much more multi-dimensional. If AI gets chemistry wrong, it does so FOR THE VERY SAME REASONS humans would make mistakes - not complete enough information. It can seek to know what it does not already know.
There is nothing human brains can do that AI "brains" cannot do. They are faster, they never forget, they can learn from their mistakes and they can learn new things. Just like humans do. Given time, it might even develop "feeling," "see" colors and enjoy music. These are emergent properties that our ancestral minds did not have until they evolved further.
We have to recognize that there is a time element to all of this, of course. Some people say "never", some people are more flexible and say there will be a "Prime Time" some time, Respert24 says "not in our lifetime."
But AI is absolutely being used today in medicine and other fields as it learns to walk before it can run.
As it can be applied to your list of diseases Leronlimab might treat, every disease has a constellation of cells, organs, receptors and cytokines associated with it, especially CCR5, of course. Just as you have researched those constellations of characteristics, so can AI and it will see patterns and correlations and will corroborate research results about them.
As time goes on it will model biological system such as cells in petri dishes, then mice and, later, even humans and it will do testing of drugs on those models such as Leronlimab. It's all a matter of time.
You say "AI can provide possibilities but it does not contain any actual wisdom or discernment." You will be proven wrong there.
May I cite the wisdom of the "Whopper" computer in the movie "War Games," a terrific movie that everyone should be encouraged to watch.
In the movie, Global Nuclear War was simulated like a game as humans stood by helplessly. After running its simulations, the computer came up with the WISDOM to say that "The only way to win the game is not to play." Yes, a computer can figure that out! That and much more.
(5)
(0)Scroll down for more posts ▼
