David Deutsch On Artificial Intelligence
1–10 of 60 posts
Re: David Deutsch On Artificial Intelligence
#2How on earth is
Bayesianism a form of behaviorism"
? And what is it with the cult of Popper. Popperism isn't "underrated", it's considered the gold standard of epistemology despite being a vague philosophical theory full of holes. Contrast that with Solomonoff induction which has a rigorous grounding and which offers hypothesis testing and rejection as a trivial subcase.
Re: David Deutsch On Artificial Intelligence
#3Not to be nitpick but is it pretty well proven that there are other self-aware animals.
Re: David Deutsch On Artificial Intelligence
#4"only a tiny component of thinking is about prediction at all"
Here I really struggle to follow his argument. It seems, he is saying that thinking is not prediction, but how does he know that? As I see it, we don't know what thinking is, so is he then saying that thinking does not feel like prediction, and therefore it cannot be prediction? It is well known that we are only aware of our very highest level of thought. What goes on below the surface may very well be very much about prediction without us knowing about. When we think about "the world" we must have some internal representation of the configuration of objects that we think about. In my opinion this is implicitly a kind of prediction, because we create (at least mostly) configurations that are plausible, if not true (counterfactual, but possible).
It seems to me that creativity is mostly about finding good approximate solutions to hard (e.g. NP-complete) problems. How we do it is so far unknown, but it seems that it is not in principle so very special. The explanation for the calendar he speaks about is actually not very complicated inference at all, and is probably in reach for artificial systems. All he has to do is to generalize 19 as belonging to a number category, and then apply the predictive properties about how numbers work. Not all prediction is naively predicting that the world is as it was before. If change is predictable in the past, then change can also be predicted to the future. Prediction is difficult, but that does not make it less important.
Re: David Deutsch On Artificial Intelligence
#5How on earth is Bayesianism a form of behaviorism" ? And what is it with the cult of Popper. Popperism isn't "underrated", it's considered the gold standard of epistemology despite being a vague philosophical theory full of holes. Contrast that with Solomonoff induction which has a rigorous grounding and which offers hypothesis testing and rejection as a trivial subcase.
How is Bayesianism not a form of behaviorism? The essay devotes a paragraph to explain this point allthough Deutsch assumes the reader is familiar with behaviorism and probabilistic methods used i AI.
Re: David Deutsch On Artificial Intelligence
#6"only a tiny component of thinking is about prediction at all" Here I really struggle to follow his argument. It seems, he is saying that thinking is not prediction, but how does he know that? As I see it, we don't know what thinking is, so is he then saying that thinking does not feel like prediction, and therefore it cannot be prediction? It is well known that we are only aware of our very highest level of thought.…
Is there a difference between reasoning and predicting? It seems you don't think so and the author does.
Re: David Deutsch On Artificial Intelligence
#7"only a tiny component of thinking is about prediction at all" Here I really struggle to follow his argument. It seems, he is saying that thinking is not prediction, but how does he know that? As I see it, we don't know what thinking is, so is he then saying that thinking does not feel like prediction, and therefore it cannot be prediction? It is well known that we are only aware of our very highest level of thought.…
You criticize the author for making unfounded assumptions about the nature of cognitive processes and follow up with your own equally unfounded assumption about the nature of cognitive processes. I do agree with you, but it is still bad form.
Re: David Deutsch On Artificial Intelligence
#8"only a tiny component of thinking is about prediction at all" Here I really struggle to follow his argument. It seems, he is saying that thinking is not prediction, but how does he know that? As I see it, we don't know what thinking is, so is he then saying that thinking does not feel like prediction, and therefore it cannot be prediction? It is well known that we are only aware of our very highest level of thought.…
I think the predictive framework is a fairly universal one. You could always as a thought experiment feed a predictor all sensory experience humans ever had, and provided this predictor works in a manner in some sense optimal it could answer whatever question at least as well as any person.
Likewise we can already simulate fairly large biochemically accurate neural networks, and it seems a scaling in power would lead to an AGI.
The problem with the previous models is their inefficiency. It does seem to indicate a lack of understanding of fundamentals -- i.e. you can brute force essentially every well defined problem in CS, yet we don't declare them solved modulo some time for computing power progress.
Re: David Deutsch On Artificial Intelligence
#9> So, why is it still conventional wisdom that we get our theories by induction?
I thought the other conventional wisdom is that we get our theories by building mental models. I throw a ball, and I can either remember how it flew in the past and use that to predict, I can also build a mental model from first principles to find out what will happen. Maybe run a short 5 second idealized simulation in my head and I can predict what might happen. Well then I might realize that there was wind that my original model was too idealized.
Another example someone on HN (or an article posted to HN) talked about how would they be able to predict what happens to a pen dropped on the moon. And why did Astronauts stood firmly planted on the moon's surface. It was interesting how many people got that wrong. And the difference was that some seemed to have been reasoning by analogy or have encoded faulty facts (gravity happens on earth), some by past experiences (they have heavy boots), some that have the correct understanding of first principles can so to speak simulate and understand what would happen -- the pen would drop to the moon. Moon has some gravity. So do other planets. And so on.
Now most of these are still based on some principles that we hold true and are "like the past". Gravity is like yesterday, the fact that winds affects objects in flight is like yesterday, speed of light is like yesterdays. There needs to be wide set of known and predictable rules about the world and then deriving or simulating the future can be done.
There can be something in between, maybe if I built a mental model just recently it gets memoized and instead of re-running it I just remember that it is like the "past one" but with a small tweak.
Another mechanism is the "seeker of inconsistencies". Something that understands that these facts together don't make sense. So if I first thought the pencil would float away on the moon and then thought but Astronauts are planted firmly on its surface. This would, so to speak, raise an exception and say "aha, something is not consistent, pay more attention to this one".
But, that is how my brain works. Others brains might work differently. I can inspect to some degree my own thought process but I cannot get into the head of someone else and "see" how they think about things. Do they visualize, to the see words and formulas instead and plug values is. Are some completely unaware of how they think and lack much self reflexive.
I am papering over and being too simplistic here and don't necessarily disagree with the article, that just that reaction when read that paragraph.
Re: David Deutsch On Artificial Intelligence
#10"only a tiny component of thinking is about prediction at all" Here I really struggle to follow his argument. It seems, he is saying that thinking is not prediction, but how does he know that? As I see it, we don't know what thinking is, so is he then saying that thinking does not feel like prediction, and therefore it cannot be prediction? It is well known that we are only aware of our very highest level of thought.…
You criticize the author for making unfounded assumptions about the nature of cognitive processes and follow up with your own equally unfounded assumption about the nature of cognitive processes. I do agree with you, but it is still bad form.
What do you mean is unfounded? My suggestion for what creativity might be as be? I guess I was unclear, I don't mean to say I know what creativity is. What I do mean to ask, is whether creativity is something more than just solving optimization problems? To me that way of thinking about it explains many of my own experiences (a-ha moments, unknown time to completion of thinking process, thinking going on in the background etc.) using a formulation that I understand. It is free for anyone to prove it wrong.