Einstein was not a very good mathematician. This quote from David Hilbert sums it up well:
“Every boy in the streets of Göttingen understands more about four-dimensional geometry than Einstein. Yet, in spite of that, Einstein did the work and not the mathematicians.”
A lot of the best startups (Airbnb, Uber) started with what was actually a dumb idea. But that dumb idea was in the neighbourhood of a very good idea, and the founders were flexible enough and persistent enough to navigate the startup from the former to the latter.
Nash said "the same way", not the "the same voice":
“the ideas I had about supernatural beings came to me the same way that my mathematical ideas did. So I took them seriously.” - Sylvia Nasar, "A Beautiful Mind"
Interesting tidbit, there was a bug in the computer assisted part of the proof (now fixed):
"Many thanks to Brendan McKay and Gordon Royle for letting me know overnight that they had successfully 4-coloured my 1567er; as a result I found a bug in the part of my code that implements the relaxation described in section 5.4 and now it agrees. "
I fear this is wishful thinking. Advances in AI are progressing a lot quicker than advances in neural interfaces. So we will most probably have superhuman AI long before we have neural interfaces.
And at that point it's game over for homo sapiens.
Identifying and squashing potential threats has been part of Facebook's playbook from the very early days. It's not surprising at all that they have developed internal tools and systems to formalize the process and make it more efficient.