Ive heard this before. Im not trying to be offensive here, but it just doesn't make any sense.
It's like being concerned that planes can't fly because they don't flap their wings.
If you want to escape this state of ignorance you're gonna have to implement some neural networks, and learn about neurology.
At the very least consider that double precision floating points are extremely insanely precise. Definitely orders of magnitude more precise than real biological neurons can measure over the noise floor of activity in a brain.
The analogue/digital isn't the main distinction I think, it's a distinction of highly optimized higher order multi-modality. Our brains have Nth degree multi modality input and output that run on a few watts (calories) while nueral networks & LLMs so far take kilowatts or more. A bird can fly 10s-100s of km with flapping with a few watts, an airplane, megawatts for fixed wing flight of the same distance. Evolution was an unintentional search algorithm over hundreds of millions of years to find resilient, optimized, and efficient-ish systems. Our toys are so new on that search tree and are thus not the most complex, resilient or efficient.
I don't think that was his point. I think he just vaguely meant digital computers could never compute the same functions as analogue computers, particularly whatever the "alive function" is.
To answer your point simply, machines don't have the same energy constraints as small animals. A bird brain gets more compute per watt than my desktop pc, but humanity can (and does. often) hook up a power plant to a building sized computer.
If you believe the function a brain is computing is computeable on von neuman, efficiency isnt relevant, in so far as you build a big enough machine, and get it running the right program.
It's like being concerned that planes can't fly because they don't flap their wings.
If you want to escape this state of ignorance you're gonna have to implement some neural networks, and learn about neurology.
At the very least consider that double precision floating points are extremely insanely precise. Definitely orders of magnitude more precise than real biological neurons can measure over the noise floor of activity in a brain.