> As soon as we get proto-AGI we will probably get advanced BCI's and full dive vr.

I think that's quite the leap. "full dive" vr would require us to essentially have completely mastered neural encoding both from a theoretical perspective (we're not even close) and from a practical perspective (we're not even close to being close).

It's important to realize that decoding neural signals (brain computer interface) is profoundly different than encoding neural signals (full dive vr). We're currently getting better at neural decoding, such as turning brain signals into limb movements and controlling virtual keyboard and pong paddles.

As far as I know we have virtually no success in artificially simulating sensory stimuli input. We aren't able to plug something into your brain and make you feel touch sensations, or make you clearly see images. We're not even remotely close to that. If we get even a small fraction of the way there the first thing we'd do is create prosthetic eyes, ears, skin, etc. Even the best tech today with cochlear implants requires an intact nerve to take the signal into the brain to get interpreted; we are not at the stage where we can directly encode auditory stimuli into the brain.