"AI is a smoke screen for what they really want to do, control who works where. If you are familiar with the way that algorithms work, you will understand that AI today is simply an inefficient automatic labeling machine made quick through the use of large computation centers. It will never approach human intelligence and they know this. They just need an excuse that allows them to control whatever it is they will claim the 'AI' will manage. Probably government."
I can't speak to the technical aspects, but as far as the agenda behind AI, I think he is exactly right. The real purpose is to centralize control. The rulers of this world are willing to spend billions of dollars to fund this kind of technology because what is billions of dollars when at the end of it you get total control? That's also what is behind the whole Internet of Things business. Even now, with all the control they have, you cannot shut down someone's refrigerator at the touch of a button. You would have to physically take it away, which is much more difficult. But if the "Internet of Things" goes the way they want, it will indeed be possible.
In other words, all the talk about whether self-aware or independently learning machines are possible is just to fool the masses or to motivate workers and make them think they are working on some grand project. But those funding the whole business don't really care about the philosophical aspects and probably don't believe in them anyway. It's just a tool for control.
So, the question is not, what will an AI do, but what will the people in control of it do? Pay attention to the man behind the curtain.
Another Wmbriggs post discusses some incidents where AI tools don't behave in the way the makers want them to. There are several examples of this kind of thing, but ultimately, this is not inherent to the technology; it just means that the makers have not figured out how to get it to say what they want. I have full confidence that given enough time and effort, they will figure out how to create a system which will given them the exact results they desire and even to change as acceptable terminology changes. Because that was the goal in the first place.
Thanks for Duhem pointer. If NLR = No Longer Reading. Otherwise, oops.
ReplyDeleteYes, it does. There's a lot of good stuff in that book. I'd be curious to know what you think of the essay when you finish reading it.
ReplyDeleteI doubt I am a person whose opinion is worth soliciting, being sparsely read and, via negativa, a born again positivist due to having had my head kicked in with fancy pants sciences of late 20th/ 21st C. Science cannot be below the level of positivism for, when it comes to a dust up, observation trumps explanation…. I came to Mach (and I believe, the name Duhem) via Julian Barbour’s history of classical mechanics which basically ends with Mach. The ideas of Mach (and Leibniz) appeal to me, although they are authors I have more heard about. I have never read Duhem. At Barbour’s prompting I started on Mach’s history of science but never stayed the course. I keep meaning to look at his analysis of sensations as a way in. As the 19th C recedes the more valuable is it’s mind. Maybe you provided the cattle prod I needed.… The Duhem extract surprised me because I associate Mach with *ruthless* positivism. The equation of will with pressure and Duhem’s subsequent comment arrested me. Now I wonder if I have Mach all wrong. Anyway thanks again.
ReplyDeleteBy fancy pants science do you mean things like string theory? I would not disagree with that characterization. Looking at the science of the present day compared to the past, even scientists who were great theoreticians or mathematicians were still deeply involved at the concrete level. What they described were aspects of nature that we could directly experience. Now, it's moved a lot more into abstract notions.
ReplyDeleteI am greatly impressed with the intelligence and dedication of string theorists, but on the other hand, one gets the impression that they have pressed that kind of thinking as far as it will go and that it's not the way to make further progress.
Where I would disagree with positivism is that I see the positivist project as unecessarily restricting thinking. I don't think that we really can axiomatize scientific thinking and only restrict ourselves to allowable modes of investigation because we don't know precisely what kinds of thinking can give us insight about nature or how those that do work exactly. And I don't think that this is something human beings ever will know.
Though what I do think we should do is to pick some phenomena that is concrete and also observable and try to understand that as well as possible. And then proceed from the concrete into the unknown. Let nature tell you about itself, so to speak.
I was thinking more of the sciences of the artificial (h/t Herbert Simon) and the complex that had bubbling away after the war and blossomed into popular consciousness, I guess, with the advent of ubiquitous computing. When someone can’t or won’t learn the necessary language [mathematics, me] such disciplines are effectively on par with literary criticism. An example, and a shoddy one in its latter days, was the online salon Edge. But I was referring specifically to personal circumstance. I am - per want of horsepower - condemned to something like positivism and unnecessary restrictions *are the attraction.*
DeleteNo quibble with the gist of what you write.
Here's another quote by Mach from the review which I found insightful:
ReplyDelete"Purely mechanical phenomena do not exist. The production of mutual acceleration in masses is, to all appearances, a purely dynamical phenomenon. But with these dynamical results are always associated thermal, magnetic, electrical, and chemical phenomena, and the former are always modified in proportion as the latter are asserted. On the other hand, thermal, magnetic, electrical, and chemical conditions can also produce motions. Purely mechanical phenomena, accordingly, are abstractions, made, either intentionally or from necessity, for facilitating our comprehension of things. The same is true of the other classes of physical phenomena. Every event belongs, in a strict sense, to all the departments of physics, the latter being separated only by an artificial classification, which is partly conventional, partly physiological, and partly historical.
The view that makes mechanics the basis of the remaining branches of physics, and explains all physical phenomena by mechanical ideas, is in our judgment a prejudice. Knowledge which is historically first, is not necessarily the foundation of all that is subsequently gained. As more and more facts are discovered and classified, entirely new ideas of general scope can be formed. We have no means of knowing, as yet, which of the physical phenomena go deepest, whether the mechanical phenomena are perhaps not the most superficial of all, or whether all do not go equally deep. Even in mechanics we no longer regard the oldest law, the laws of the lever, as the foundation of all the other principles."
Your work is done. I’m reading the whole thing. There is a lesser gem regarding elastic and inelastic bodies. All the fuss about the abstraction *rigid body* revealed itself in full perplexity. The further in, the more it is proving a similar experience to reading Clifford on the foundations of mathematics; he taught me what I actually learned in *sums*. Thanks again. Now I really must re-lurk.
DeleteAfter the Mach essay, I would recommend "Some Reflections on the Subject of Experimental Physics" in the same book. It touches on the issue about how we should think about physics and contains some good descriptions of how exactly to interpret experiments.
Delete