Format: I choose the question format to allow other people to provide their own answer. I write my answer as a post (ie. in the description of the question) given I think it’s the quality of a post. This is a norm I would personally like to see adopted.
Confidence: speculative; good probability that others could update my model significantly | Created: 2019-01-18 | Updated: 2021-01-12 (major update; basically written then)
Context: I wrote this in the context of the cryonics signup sequence to complement the article “2: Neurocryopreservation vs whole-body preservation” which I’ve reviewed pre-publication.
When considering the question of neuro-only vs whole-body preservation, I evaluate the following 3 aspects:
a) Part of identity preserved in the body
b) Difficulty of creating and connecting a body to a brain / mind
c) Practical considerations or preserving a body: facility to move to and out of the facility, and cost, as well as impact on the preservation of the brain
This post will be about (b). For (a) and (c), check out part 2 of the cryonics signup sequence.
Reviving a neuropatient as opposed to a whole-body patient requires 2 extra steps: building a body and connecting it to the mind.
Note: It’s also possible to preserve a whole-body while still focusing on preserving the brain first, possibly even while still having a neurosperation.
On building a body
We know building a body is doable, as nature already does it. While we don’t have a particularly advanced understanding of the human genome, we know it encodes the instructions to build a human body. As our understanding of the genome increases, along with our ability to edit it, we should be able to adjust the genome to create a body without a brain, and to have the body grow to an adult size within the womb. Of course, the womb would need to be artificial.
Another plausible path is that we build a virtual body embedded in a virtual world. To some, it might seem like this would cheapen the experience. But ultimately, our brain interfaces the world only through sensors — so whether the sensors’ input are coming from the photons and atoms of nature or from a computer would produce indistinguishable experiences. And you could still share that virtual environment with your loved ones.
On connecting a mind to a body
An advanced AI might be able to determine which nerves served which purpose, and be able to reconnect it to a body with nanotechnology, or interpret its output to propagate them to a virtual body.
Although it seems like the way our brain encodes information differ from human to human, in which case the only way to learn how your brain commands map to muscle contractions might be through training — meaning you’d have to imagine doing various movements and let the AI know which movements you’re thinking of so that it learns. The AI could also learn from visible mistakes you make (ex.: dropping a glass).
Maybe recording that information prior to deanimation, by having brain scanning + video recording while doing various movements with your body would reduce the learning curve post-revival. Existing technology seems to be able to distinguish between up to ~120 commands (source: Brain-reading — Wikipedia). Part of this limitation is probably software, but to a large extent I suspect that it’s hardware — so no matter how good AI gets, they will only be able to get so much information from reading old (ie. from 2021) brain scannings.
On the complexity of reviving neuropatients
To reanimate a neuropatient, the brain needs to be healed, a body needs to be constructed, and both need to be connected.
I think building a body is by far the most trivial of those 3. Connecting a new body seems like it should be rather trivial if we allow for a learning period, although potentially annoying. Without a learning period, I don’t know if it would be easier or harder than repairing a brain.
A common line of thinking is that given creating and connecting a body to a brain should be much easier than repairing a brain, this means that by the time we can reanimate whole-body patients, we should also be able to reanimate neuropatients.