When I talk to AI researchers about the possibility of an artificial general intelligence, or AGI, I often ask them if they think that it is possible. I generally get an answer to the effect of, “Yes, I don't think there's anything magic going on in the brain.”
Sometimes people seem to think that question boils down to that, either it’s mechanistic, or it’s magic. And if it’s mechanistic, it’s something that we can eventually replicate. The challenge with that is that intelligence may not come from the connections between the neurons, like we think it does now, but from within the neuron itself.
There's something known as the Planck Scale, which is the smallest discrete unit of space that's possible. Think of it as the resolution of the universe. What one grain of salt is to the universe, one Planck unit is to a grain of salt. So, it's incredibly small. There is more than one theory that suggests that what is going on in human intelligence occurs down at that level. In which case, while it may be, in some remote distant future, possible to generate an AGI, it's prohibitively difficult right now.