Make a camera and software that gets a pretty good elementary definition of the most basic image needed for driving a car. Put five or six of these around the car exterior. As long as the image elements can be ordered, that is all you need. Just connect the bots, and feed them these elements and connect the bots together, then drive the car. These hierarchical bots will organize the elements into minimally redundant. recursive groups which must remain adiabatic to drive safely. That is all, no brainer with the TOE. The elements must have a tangent bundle, they need a divide algorithm that generates their indices.
That half built semantic processor to the right. It navigates graphs using the match function. All it needs is a match function with some 6 bits of accuracy, and its all over. The web become intelligent to a recursive set of fifth order compact polynomials, polynomials with coefficient 1.. When one of these banker bots has reached it limit on getting a rational approximation of Phi, it just dublicates itself, takes a seat as a member bank, then launches it duplicate. The new bot goes exploring and semantic phrases and recursive bots take member seats and update thier estimate of phi, and so one. All they are doing is fining their maximum divergence while remaining connected. And as long as the precision of the 'match' function is incentivizes humans, then its all over. We will have the singularity.
No comments:
Post a Comment