A simple logical argument for why AI is possible

Premises:
1 - Biological brains operate through the interaction of neurons communicating
2 - Neurons communicate through electrical impulses and neurotransmitters
3 - The exact informational content of a biological mind is stored within the structure of the connections between neurons
4 - Sensory input and motor output both take the form of electrical impulses with measurable voltages
5 - Single neurons and clusters of neurons can be mathematically modelled in computer software
6 - Biological brains can be mapped to produce a connectome
7 - If a mathematical model of a component of a system is accurate enough, the entire system can be modelled by linking together the simulated components
8 - There is no element of human cognition handled by any process outside of the interaction of neurons in the brain
9 - An accurate model of an entire human brain will result in human cognition and reproduce the same intelligent behaviours as in the original biological brain
10 - Running such a model can be done so long as sufficient memory is available (see https://en.wikipedia.org/wiki/Space%E2%80%93time_tradeoff if you're unfamiliar with this basic aspect of computer science)

Conclusions:
1 - A collection of mathematical models of neurons built from a scan of a human brain will result in human-level intelligence
2 - Such a model can theoretically be built using present day computers if sufficient storage space is made available - making it run in realtime or faster is an optimisation problem

Of course simulating an entire human brain is a horrendously inefficient way to obtain intelligence and shortcuts can be taken - for example simulating whole networks instead of single neurons, or using an alternative cognitive architecture.

The real obstacle to implementing whole brain emulation is primarily one of resources: current hardware is not really practical to use. Advances in neuromorphics and in advanced data storage may eliminate these issues. Practically and physical possibility are two very different things however, and it is entirely reasonable that one day mankind may have the ability to practically implement a model of a whole brain.

The key point is that it is NOT impossible to have intelligence in a computer system - it is just a problem that requires hard work as in all other human endeavours and it would be foolish to claim the pursuit of the idea to be pointless based on the assumption that it will never come to fruition - that is a true self-fulfilling prophecy.