I very much fear that the spontaneous ontology of those who wish to set themselves up as the makers or re-creators of the world know nothing of the beings who inhabit it, only lists of characteristics. If the nanobiotechnological dream were ever to come true, what still today we call love would become incomprehensible. –Dupuy Cybernetics is Antihuman
So here is the crux in the essays by Dupuy and Domingos that I have been reading, there may be an digital clone of us (a digital double), or there may be AI with the appearance of consciousnesses, but we should not mistake these creations as being ‘human’; we should not give them anthropomorphic mental states.
Dupuy runs through the metaphysical assumptions behind different version of cybernetics. The question is who is at the center of the universe, is it God? is it Man? Is it a tool (or technology)? Can man be reduced to a tool, can man as a tool maker become equivalent to God?
The mark of sophistication for the tool has moved from the watch (and the watchmaker) to the dynamic system (and the systems developer). It is a movement from a predictable deterministic universe to an unpredictable non-deterministic world, where the mark of freedom somehow becomes confused with unpredictability. Perhaps we should reformulate the Kantian distinction between the natural laws and ethical laws. No longer is freedom granted because we are all sovereign in the kingdoms of ends, but because we are opaque to ourselves and unable to predict how we will act. This is in contrast to the immutable laws of nature which we can predict.
Perhaps this means that we have an unconscious, that there is something in our self that is not our I (our Ego). And perhaps this is our spirit, but perhaps this is just a hidden part of ourselves. Can this be programmed into a computer or an AI? Even if we try to make an AI to evolve a ‘human’ consciousness, is there something that perhaps we are missing that we do not even know we are missing, as Domingos suggests?
Moving to Domingos, he runs through the 5 types of strategies pursued in AI, and the types of problems they are best at solving. We can break apart the Dupuy paper, and ask what if there is no ‘cybernetics’ but a multiplicity of ways to generate ‘cyberneticity’. It is the age old conflict between Parmenides and Heraclitus. Is my artificial consciousness one or many? Or perhaps I just contain multitudes.
But this is all just background to the real discussion of what we will use the AIs for, virtual selves. What is the ontological status of our virtual selves? What is the metaphysics of their virtual world. Domingos assures us that we should not feel bad for AI doubles, which endure various virtual trials on our behalf before suggesting to us the recommended course of action.
I mean, should we feel bad for the virtual bank accounts that lose money during monte carlo simulations of trading systems? But these virtual actions make me think about Nietzsche and the eternal return. The Eternal return is what gives our life weight (perhaps meaning), if our lives occur once they are too light or too meaningless. Do our virtual selves experience the weight of the eternal return, while our real selves only live once?
‘All truth is crooked, time itself is a circle.’ So is this even a problem that can be solved through optimization (through computation and modeling) – that is the problem of living. In modeling our optimum lives, what are we optimizing for? As Dupuy recounts, in his story about the Holocaust survivor, a cloned wife of the Holocaust survivor does not satisfy the man’s longing for his wife and when he recognizes this then he is able to let her go. Dupuy suggests this is because most perfect simulation fails to capture something. This is not a great leap of the imagination, but it does touch on something I found in both essays.
In reading both these essays I was struck by how the human is associated with the MIND. I envision a phenomenological computer game with a gigantic brain, or gigantic brains, interacting. Where are the limbs and the eyes? They no longer seem to matter. How does this particular perception of ourselves or of what it means to be human contribute further to the AIs that we make?
Some further reflections I took down:
Man vs God: A nod to the metaphysics implicit in cybernetics.
Intelligent Design: How do we reconcile our AI design with evolution and the deist notion of a “designer”
Freedom and Capitalism: Freedom like capitalism cannibalizes itself. It is never satisfied, nor has it ever accomplished its goal, which is itself a cybernetic loop.
Living and nonliving: The reduction of consciousness to AI, nano bots, crispr and genetic engineering the idea of what is life or non life is blurred as everything has the potential to be life or to be integrated into one’s life as one becomes a cyborg.