Very interesting - it's useful to have such a dense summary of your thoughts on this topic.
One preliminary question that comes to mind: what lasting influence do you think horses have had on transportation, that wouldn't have occurred if they hadn't existed as an intermediate between walking and motorised cars? Maybe this would be a somewhat useful reference point for the potential ability of ems to preserve a human legacy.
Sorry for the moderate tangent but - Robin, I'm wondering how much you (and others) expect AI to rely on 'normal' software.
The extent to which humans rely on formal systems seems noteworthy - suggests that, even in spite of our brains being enormously more powerful than these simple artificial systems, we apparently needed to build them anyway. It wasn't possible to just look at each problem in isolation and human it away, through sheer ingenuity, relying on nothing but our big awesome brains.
This seems to suggest to me that, even if our brains do just implement a couple of simple algorithms, and therefore a clump of simulated neurons in any configuration will in fact spontaneously go foom, it will in the process of taking over the world still need to build lots of cheap, context-specific, brittle systems to work though.
Should we be happy to let self driving cars relieve us of the burden or regret becoming passengers? Should we be happy no conscious being is consigned to that fate or regret the loss of potential friend or an enduring existence for ourself?
Horse needs were so close to human needs that we didn't change that much to accommodate them.
Very interesting - it's useful to have such a dense summary of your thoughts on this topic.
One preliminary question that comes to mind: what lasting influence do you think horses have had on transportation, that wouldn't have occurred if they hadn't existed as an intermediate between walking and motorised cars? Maybe this would be a somewhat useful reference point for the potential ability of ems to preserve a human legacy.
An em IS an upload.
Yes, I expect ems to use lots of normal software, and more over time, just as we humans have.
See my added to this post.
Sorry for the moderate tangent but - Robin, I'm wondering how much you (and others) expect AI to rely on 'normal' software.
The extent to which humans rely on formal systems seems noteworthy - suggests that, even in spite of our brains being enormously more powerful than these simple artificial systems, we apparently needed to build them anyway. It wasn't possible to just look at each problem in isolation and human it away, through sheer ingenuity, relying on nothing but our big awesome brains.
This seems to suggest to me that, even if our brains do just implement a couple of simple algorithms, and therefore a clump of simulated neurons in any configuration will in fact spontaneously go foom, it will in the process of taking over the world still need to build lots of cheap, context-specific, brittle systems to work though.
Should we be happy to let self driving cars relieve us of the burden or regret becoming passengers? Should we be happy no conscious being is consigned to that fate or regret the loss of potential friend or an enduring existence for ourself?
I do not see how you get emulation without upload. So emulation seems very hard to me. How else can you know enough detail to do the emulation?
What about non-ordinary software? What about brain-rooted external cognitive extensions?
Agree with this if those were the only two possibilities. But they're not.