Uhm, this is probably incorrect as an analogy, but, do we humans actually need those 75 billion neurons on our brains?

I mean, there are lots of people who have lost a brain hemisphere for different reasons, and yet, they live happy lives.

However, what they lose is flexibility. This means they have a hard time when faced to new situations and have difficulties adapting to them.

I can't be certain, but it's possible that the number of parameters in large language models can account for their flexibility. That is why you can throw anything to chatGPT and it will answer, within the scope given by its restrictions.

I'm not sure either if enlarging the number of parameters will give us emergent properties or if it will only slow down data processing. Blue whales have immense brains, but they aren't necessarily smarter than us. And this is because a larger brain means larger distances for neurons to connect, slower response times and increased energetic expenditure.

I could be wrong, though. Electronic brains don't have the same limitations of physical brains, so maybe increasing their size won't affect their output.