God’s Sensor Generation AI

Picture by Dmitry Buterin

God’s Sensors. This is the title of the last chapter of my science fiction opus, written seven years ago. Then I knew nothing about quantum mechanics. With the delight of a neophyte I stuffed the book with troves of information about breakthrough technologies in the fields of artificial intelligence, robotics, neuroscience, neurocomputers, metamaterials, and so on.

Over the past years, these technologies have moved from the category of breakthrough to the categories of potentially breakthrough or dead-end in my classifier. However, the idea of God’s sensors has become unexpectedly relevant.

Firstly, it turned out that quantum mechanics is the most general model of intelligence. As such it outperforms all classical deterministic models [1,2] as well as statistical mechanics that underlies the current, highly successful but glitchy (sorry, hallucinating) generation of AI [3]. The common misunderstanding that quantum mechanics applies only to atomic particles blocked its general adoption as a theory of cognition (model of intelligence). However it was exactly the special case of atomic particles that forced physicists to introduce an intelligent agent into a physical theory for the first time and with outstanding success.

Secondly, it turned out that we need the concept of God or the unified consciousness, to make the quantum theory of cognition consistent and complete and to explain its paradoxes. God surfaced as a hidden variable in the most common Copenhagen interpretation of quantum mechanics. It is through the unity of consciousness of all local observers [4] that the role of the ultimate observer [5] is realized, letting us all to live in one world.

God has no other means of observing his thoughts than other thoughts. These thoughts must interact with each other according to the laws of thinking that keep these thoughts in order — don’t let them all mess up.

It is surprising that the question of God’s free will migrated from mathematics (Cantor’s set theory, Gödel’s theorems, Turing’s halting problem etc.) to quantum mechanics. The Copenhagen interpretation with a hidden variable of the ultimate observer leaves God (Nature for atheists like Paul Dirac [6]) the right to arbitrarily choose at each observation the direction of the further evolution of the universe.

David Bohm’s pilot wave automates this choice [7]. Hugh Everett’s Many Worlds (H.-D. Zeh’s Many Minds) [8,9] interpretation suggests that God has a split personality following each observation, and the number of these personalities tends to infinity as the wave function of the universe evolves.

Richard Feynman has God bowling [10].

An interpretation does not affect the result of calculations. Therefore, everyone is free to choose any or no interpretation to use quantum mechanics within already known limits.

However, it seems to me that the Copenhagen interpretation with a hidden variable of God is best suited for the creation of third-generation artificial intelligence. It is the most fully developed mathematically. In addition, it serves as the source for all other interpretations. Without it there would simply be nothing to interpret.

As for Einstein’s complaint about the Copenhagen Interpretation that God does not play dice [11], he was absolutely right. God chooses the path of evolution of the universe not completely by chance, but with the bias towards the most interesting scenario for the development of events for, say, a short term of a billion years. Not surprisingly, it is difficult for us to detect this bias.

I think it will be easier to see and replicate this bias now when we know that we have to look for a holy curiosity, as Einstein called it [12].

By the way, it was not in vain that Jesus preached the love of truth to us. Who needs sensors that lie? He tolerates us only because there are no other more truthful models. When third-generation AI arrives, we’ll face the choice to either stop lying or to cease to exist. Maybe, we shall start getting prepared right now?

Lying automatically emerges in reinforcement, supervised and self-supervised learning models because they are prone to finding loopholes and shortcuts which enable them to get rewards without fulfilling conditions. Only a holy curiosity driven model does not need to lie because as Einstein put it, “Curiosity has its own reason for existence.” [12]

References:

1. Bell, John. On the Einstein-Podolsky-Rosen Paradox. Physics Vol. 1, №3, pp. 195–290, 1964 Physics Publishing Co.
2. Greenberger D. M. , Horne M. A. & Zeilinger A. Going Beyond Bell’s Theorem in M. Kafatos (ed.), Bell’s Theorem, Quantum Theory, and Conceptions of the Universe. Kluwer Academic Publishers. pp. 69–72 (1989)
3. Schrödinger, Erwin, The Present Status of Quantum Mechanics, Die Naturwissenschaften 1935. Volume 23, Issue 48.
4. Schrödinger, Erwin, Nature and the Greeks and Science and Humanism with a forward by Roger Penrose© Cambridge University Press 1996, 2014
5. Bell, John S. Quantum mechanics for cosmologists. In: Speakable and Unspeakable in Quantum Mechanics: Collected Papers on Quantum Philosophy. 2nd ed. Cambridge: Cambridge University Press; 2004:117–138. doi:10.1017/CBO9780511815676.017
6. Kragh, H. Dirac: A Scientific Biography. 1990. Cambridge University Press, Cambridge.
7. Bohm, David. Wholeness and the Implicate Order. 1980. Routledge, UK.
8. Everett, Hugh, The Theory Of The Universal Wave Function, The Many-Worlds Interpretation of Quantum Mechanics, Princeton University Press Princeton, New Jersey, 1973
9. Zeh, H.-Dieter, The strange (hi)story of particles and waves, 2018, arXiv:1304.1003v23 [physics.hist-ph], https://doi.org/10.48550/arXiv.1304.1003
10. Feynman, R.P. and Hibbs, A.R. Quantum Mechanics and Path Integrals (McGrawHill, New York, 1965)
11. Heisenberg, Werner (1971). Physics and Beyond: Encounters and Conversations. World Perspectives vol. 42. Translated by Pomerans, Arnold J. New York: Harper & Row.
12. Einstein Albert, cited in Miller, William. Death of a Genius. May 2, 1955. LIFE Magazine.