Dear Max,
I'm a technophile passionate about the AI discussion, "the most important conversation of our time" as you say, am having fun reading Life 3.0: Being Human in the Age of Artificial Intelligence, but why didn't you mention that strong objection raised by Roger Penrose ?
You share a lot of "Holy Shit" moment about AI's prowess, but barely touch the substance of interesting objections to the very possibility of Artificial General Intelligence.
You said: " physicists know that a brain consists of quarks and electrons arranged to act as a powerfuk computer and that there's no law of physics preventing us from building even more intellugent quarks blob"
I was wondering why you didn't mention at all the objection below, and what you think about it.
- human-level intelligence might not be replicated with AI as human intelligence isn't necessarily computable
- No one has proved the computability of the human mind
- There are many good reasons to think that a human mind is not computable
- Not all physics is computable to start with
- As a simple example, there is no way to simulate true randomness with a computer program, without external input. You can do a good approximation, but however hard you try, it will always have some patterning to it. Yet we have examples of pure randomness in physics, e.g. in radioactive decay. So it appears that we can't simulate all of physics in a computer program.
- There are many problems in maths that are non computable
- It's possible that a human mind is another example of non computable physics; a much more difficult example than randomness.
- In quantum mechanics, the most complex system we can model exactly is the hydrogen atom. In the Newtonian theory of gravity, we have exact ("analytic") models for the 2 body problem but normally have to use approximations when we get to three or more bodies mutually interacting.
- With models of the brain similarly, we can't model every atom, nowhere near. The idea is to model it at a higher level, as simplified nodes in a neural network in place of neurons. It's an approximation, has to be.
- But why assume the functioning of the brain has to be computable? Perhaps this process of simplification is the step where we lose whatever it is that permits humans to understand truth?
- The idea of how the brain works used by many researchers is that the brain is made up of neurons which are each basically quite simple logical units. The idea then is that if we can build a neural net that resembles the brain, it would approach artificial intelligence. Then more tweaking could get us to the goal.
- Well, there are reasons for supposing the brain is at least several orders of magnitude more complex than this approach suggests. Because - even an amoeba can make decisions and has a fair degree of basic intelligence. It can distinguish food, escape predators, seek out more habitable conditions etc. If you modeled the behaviour of an amoeba with a neural net you would need thousands of neurons. But it doesn't have any!! It's a unicellular organism.
- So, surely our brain's neurons are more complex than just simple logical units. Otherwise how can a being with a single amoeba type cell for a brain compete with another being that has thousands of neurons as conventionally understood ?
- These simplified neuron models are valuable and lead to insights into brains and computer vision. But it's just the idea that it captures all of it. It may only captures some of the things neurons do. By the amoeba analogy, they may be dealing with a comparatively crude "macro layer" of how the brain functions.
- Roger Penrose argues that there are quantum processes happening in the brain, that we get quantum coherence in the brain - not just within the cell but spanning many cells. And eventually when the amount of mass involved in the coherent quantum state reaches the Planck mass that it then collapses due to gravitational effects.
- Many people said in the past that Roger Penrose was crazy, but the possibility of quantum processes in biology has since been confirmed.
- For instance, photosynthesis is more efficient than can be explained without quantum coherence and entanglement. Enzymes may use quantum tunneling over long distances. Magnetoreception in birds navigating by the Earth's magnetic field uses quantum entanglement to put an electron in two states at once. It's also found to be an important factor in how we are able to smell things.
- Roger Penrose's idea is that this actually happens on a very large scale in our brains. And then the collapse is an example of a non computable phenomenon.
Last, I was glad to read you mention an AI creating movies in the prelude of your book, I happened to have written an article about that very capacity a few weeks before that you may find entertaining as well :) Before 20 years, AI will be able to generate an episode of Game of Thrones instantly
Regards,
twitter.com/thomasjestin