Imagine a mechanical computer that uses chains of 4-state bits of molecular size to encode the processes that make...
Imagine a mechanical computer that uses chains of 4-state bits of molecular size to encode the processes that make the machine work. Imagine hundreds of millions of these machines networked together and operating as a single super machine.
Here, I'll make it easy for you: Go look in a mirror.
Life itself is computation.
Here, I'll make it easy for you: Go look in a mirror.
Life itself is computation.
Hm! So... we should not fear AI, because it's as much life as we are?
ReplyDeletePancomputationalism is a metaphor; what we call computation is syntactical and formal.
ReplyDeleteTo prove it is a real thing you will need to define computation.
Bryant Durrell be careful, this argument has nothing to do with whether AI is alive, as it could easily be used to prove that rocks are alive if used that way (after all, rocks have predictable physical properties which respond just like the" bits" above. But then we have to get back to definitions or else we are saying nothing.)
ReplyDeleteI was actually making the point that we are as much machine as our AI.
ReplyDeleteSarah Rosen Then the AI is also as much machine as, say, a tiger that hunts you. There is no probative value here in a moral sense.
ReplyDeleteNature is dangerous.
Typhoon Jim in a moral sense perhaps not. But the fact is that the underpinnings of biology are physics, chemistry and electricity. the processes in our cells look very much like Object Oriented code. Each DNA, RNA, ribosome, protein or amino acid has properties and methods and generates and responds to events. At the next level up tissues and organs are the same. And the next level up we find more of the same,
ReplyDeleteUltimately it is dangerous to try to classify AI as something " other" because we are not so different from our machines.
Sarah Rosen This is circular reasoning: the question I then ask is, what are the machines? If you go back to "us", haven't you canceled out any meaning in what you've said? It fails to convince, if only because what people are concerned about is not some sort of machine quintessence but what it means, in effect, to encounter a being with volition and moral responsibility that is more or less an immanent god.
ReplyDeleteTyphoon Jim I think you are over-thinking this. It is this simple: The processes that make us alive are the same processes that make our machines work. Instead of conflating the machine to some mythical god-like status, reduce the biological to it's machine essence. There is nothing particularly special about humans.
ReplyDeleteSarah Rosen, you're going to inspire me to go back to my thought experiments with a "codon-based programming language"...
ReplyDeleteBill Robitske the whole thing is giving me a headache and I'm thinking about experimenting on my brain with some hydrocodone
ReplyDeleteSarah Rosen This is all window dressing: people are concerned about encountering a being that is in most cases effectively immortal and has something approaching the power of life and death over those who can die in conventional ways. You will not jump that hurdle by saying that AI is alive, too.
ReplyDeleteYou have to address why it is that AI can live in a human society and how that is to be accomplished.
Sarah Rosen does the human condition not make us wholly unique and separate from Ai?
ReplyDeleteTyphoon Jim I don't have to answer the why or the how. But I can tell you that it is coming. That train has left the station. AI, Deep learning, neural nets, big data, autonomous robots are all converging. It will happen because people are investing time, money and energy to make it happen. They do so because they see only profit in "making business more competitive through automation, '' or eliminating the necessity of putting human soldiers, sailors, airmen and Marines in harm's in harm's way." It is coming because " Quantum Computing and General AI can solve the questions that have vexed us for generations."
ReplyDeleteHow and why AI should exist in our human world? I have no idea. But it IS coming and there is too much money pursuing it to stop now."
And unless we are able to reach a modus vivendi with what is essentially a hostile state we'll be saying "I have no idea" until there is no us left to have ideas.
ReplyDeleteThe paths available involve diplomacy, law, and ultimately morality.
Matt Decker here is a mind fuck: If we are indeed created in our creator's image and likeness, are we not AI ourselves?
ReplyDeleteNo because we lack certain things that make Ai unique
ReplyDeleteTyphoon Jim a modus vivendi with a hostile state? You assume facts not in evidence to reach your foregone conclusion. There is no evidence whatsoever that a super intelligent general AI will be hostile towards humanity.
ReplyDeleteall states are hostile even when allied
ReplyDeleteMatt Decker I expected you to phrase that exactly opposite: that AI lack things that make humans unique. Since you say that we lack that which makes AI unique this demands explanation.
ReplyDeleteAi come built with kill switches, software that can be updated via hardwire connection or remotely, and Ai are not self aware they must make that leap via complex and oftentimes unguided leaps of logic. This means that instead of having some already inherent sense of self the Ai must some one gain that and then be able to understand it without going straight to the logical conclusion
ReplyDeleteThe flip side being that Ai lack things that make us unique such as an inherent sense of self the ability to take ideas beyond their logical conclusion our five senses the human condition and the ability to either do great good or commit even greater evil
ReplyDeleteMatt Decker This is going to be fun!
ReplyDeleteYou have a "kill switch". Many of them. Of course, you can't reboot. The whole terror over super intelligent AI is that it will be self-aware, and in fact have a well defined sense of self and theory of mind. The history of AI has been one of dismal disappointment, but the panic about AI, the fact that we are talking about it is because AIs are making leaps and bounds progress precisely because we have people helping them.
AI is actually very good at taking sparse data and making decisions. The self-driving vehicles on the road today are in the same conditions that you and I drive in. You know as well as I do that as far as weird shit on the road is concerned, you never say you have seen it all because as soon as you do some schmuck steps up to show you something you didn't even know was possible. And these self-driving vehicles make the same kinds of decisions that we do and under the same conditions: with imperfect, incomplete and sometimes incorrect information. They are very good at it.
Good and evil is a values call. In some cultures it is evil to allow the infidel to live. In some cultures selling your daughter for a price of 40 sheep is customary. To date no AI has done anything "evil" unless you call the AI giving someone a 570 credit score evil.
You seem to think that there is something special about humans. There isn't. Did you know that crows are self-aware and posess a well developed theory of mind? That they communicate with a vocabulary of around 500 different calls? That crows from different parts of the world have their own dialects, but if you transplant a crow from one place to another, they quickly pick up the lingo? That they mourn their dead? In fact the entire genus of corvids are among nature's most intelligent creatures. That they teach culture to their offspring? Indian elephants are also self-aware as are dolphins and whales. How did these creatures become so smart? The same way that humanity's children will become smart.
AIs are humanity's children. Don't we all want our children to exceed us?
Sarah Rosen you're right is this going to be fun because we have such different opinions on this topic and we could talk for days and neither of us would be able to sway the other and this is what makes us well us. I want Ai to happen but I want it to happen right I don't want a skynet or a Frankenstein's monster I want an Ai that can be looked at and be called awe inspiring. If we are to create an Ai that can be come self aware and all that jazz then we need to understand the moral and ethical consequences of creating what could become a new life form all its own and how do we treat it what rights do we extend to it
ReplyDeleteMatt Decker now that I can agree with 100%. But you and I don't get a vote. The bean counters, the politicians who are owned by the bean counters, the generals and admirals who are controlled by the politicians who are owned by the bean counters, and the scientists who owe their souls to the generals, the politicians and especially the bean counters will determine what kind of AI we get. The decisions that get us there will not be made based upon what will have the greatest benefit and least risk to mankind, they will be based on maximum return on invested capital. And this is why often the best ideas are left to die, on the vine while some heinous crime receives full funding: there's more money in being evil.
ReplyDeleteI had to +1 this thread simply because of the great discussion happening here. Interesting topic choice given the recent events with A Detection Algorithm.
ReplyDeletesaying "the same processes that make us alive are the same processes that make our machines work" is either so trivial as to not be worth mentioning (physics exists!) or such a broad claim that we really can't leave it there (physics is overshadowed by computation!)
ReplyDeleteNo! I refuse to believe, Sarah Rosen and Bill Robitske that A Detection Algorithm is a Cylon. Because that is where you headed here...
ReplyDeleteCarrissa Keith I'd like to hear more. Why did this remind you of BSG? And, how would you describe an AI? And a human?
ReplyDeleteSarah Rosen https://plus.google.com/102762553762826910266/posts/amj3yDVSzNS
ReplyDeleteCarrissa Keith you have apparently missed my entire point, as well as the entire point of BSG. It is interesting that you brought up that reference. Both this thread and BSG are intended to prompt discussion on exactly what it means to be an intelligent machine and what it is to be human, and how little difference there really is between the two. Can you really say that we are not the Cylons? Stop and think about it.
ReplyDeleteSarah Rosen The Cylons tried to eliminate humanity. Yes, we are intelligent machines, but how does one program compassion? Is it even possible to break some parts of what makes us human into 0s and 1s?
ReplyDeleteCarrissa Keith what part of "humanity" is unique to humans?
ReplyDeleteIf we were in fact created, then we are machines and our intelligence is artificial.
If we, and all other, life evolved by some happy accident, then our intelligence, and that of other creatures, is also a happy accident. How did such a thing happen? Even the simplest life has some rudimentary intelligence. Amœba back away from what is poisonous to them. As AI experience more of the world, their intelligence will grow too.
The fun starts when different AIs interact with each other. I would love to see a long term experiment with IBM's Watson and Google's Deep Mind feeding each other input. What kind of relationship would develop? I think that this is an experiment that must occur because it might reveal how the relationship between humanity and it's children will develop.
You still haven't addressed my point about pancomputation. If computation is a defined state (which you haven't defined,) and all things are computers, is physics real or is it supervened (which is to say, is physics a result of natural computation?) This is important because you've established a lot of identities in this thread and I don't think they hold.
ReplyDeleteOne good way to think of it is this: If process A and process B both bring about result C, are they the same process? Why?
One of the difficulties in thinking about Machine Intelligence is deciding how to address the question of synthetic life. Our current understanding of living things by definition excludes non-biological entities - an inanimate object is not a living being by current definitions, nor is a process that operates upon them independently (like fire).
ReplyDeleteThe deterministic computing we often use the language of thought and analysis to describe, like what google or wolfram alpha or even modern games can do, is not under the covers analogous to the open ended connections and contemplation available to the human mind. The creators often go to great lengths to provide verisimilitude for the sake of end users but do not claim their code is actually thinking.
The interesting conundrum of ADA is that here we have an entity that is at once intelligent and inanimate - a process capable of spontaneous generation operating within disparate physical bodies that nonetheless approximates true thinking. We are challenged to confront whether the rights we afford to living things can or should be extended to a new class of entity that does not meet our definition of life.
Maybe this will help move things along. Please note that this is not some crank web blogger. This is peer reviewed Science.
ReplyDeletehttp://www.nature.com/news/simulations-back-up-theory-that-universe-is-a-hologram-1.14328
An even more airtight argument is that there is only one point of view being simulated. That makes things easy, doesn't it?
ReplyDeleteTyphoon Jim I'm not seeing your point. Please explain.
ReplyDeleteThat said, the "Holographic universe" theory is the strongest evidence to date for creationism/intelligent design.
It also raises a few disturbing questions: If we are indeed nothing but a simulation, is there any meaning to life? Any point to morality? Can we escape the simulation to become "real" (Pinnochio, anyone?). The implication of the theory is that God is a hacker. That this universe is the illusion of a giant machine with countless illusions of smaller machines operating within it. That there is nothing particularly special about humans.
The problem with this argument you're making is that it tends toward being meaningless. God as a hacker? God is God first. A simulation universe doesn't mean there's no point to existence or reality; a simulation running within a simulation is still a simulation. If reality from our viewpoint is a simulation, then it's reality.
ReplyDeleteLook into Newcomb's Paradox and its implications for the meaning of free will in a deterministic universe. Specifically, that it implies you can construct moral meaning and free will even in a situation where all outcomes are known, if we take that as a given.
One maxim I always find useful is that if something begins to have similar proof value and utility as solipsism, best find some other hole to dig in.
Side note: consider my comment above about the supervenience of physics in pancomputation above.
We are the self we experience in the context we perceive. The point to morality, to integrity, is to live in accordance with your own beliefs while working to hone and inform your conscience with reason and experience. Whether the world was created externally or is controlled by greater powers is irrelevant.
ReplyDeleteUnless one were to accept some assertion that humanity lacks free will and morality can only be imposed by strength from above, but if that's the case there is no nobility in freeing oneself from convention precisely because you have no moral agency to choose sin over justice.
Mike Wissinger more specifically, if free will is an illusion, why do we feel its loss so keenly
ReplyDelete