First published: January 7, 2023
Last revised: April 16, 2025

One point of view shared by several philosophers and neuroscientists (e.g. Joscha Bach) is that consciousness is software. We don't experience anything that isn't virtual, therefore since we experience our consciousness, it must be virtual too. Consequently, whatever processes that run on our brain could also run on a different substrate and therefore it should also be possible to instantiate a consciousness in a computer simulation. In particular, if we were to simulate a brain exactly, the simulation indeed would be conscious.

I don't buy it. It feels like we are conflating the mechanism of consciousness with the experience of consciousness. There is a thought experiement that I like to mention when discussing this specific issue: suppose you are to carry out the computations of a simulation of the brain by hand. Sure, it would take an absurdly long time, but this is irrelevant: in principle, as a consequence of our scribbling on paper, a consciousness somehow, somewhere would feel. This, I find absurd.

To dig a little deeper on this hypothesis, what is "information processing"? At its most fundamental level it's a sequence of states of a system that either trigger each other or that are coordinated from the outside (though one could argue it's the same, just with a more complex causal chain). I will come back to this in a future post.