  # Did some calculation refactoring for ASI

Note that this isn’t an algorithm if I understand exactly what I algorithm is, but I came up with a calculation for the amount of neurons needed for approximately a chimp level of intelligence. ( Note I actually think chimp are really intelligent, but topic for another day. )

T = Total number of training examples
D = Total number of matching methods to use training examples.
A = All possible numbers of newly generated training examples created by the algorithm ( I found a way to get the mechanism to generate its own training examples ).
I = Number of instances duplicated for redundancy.
P = The total growth of AI within the century to the level of a sleeping monkey.
S = Represents super-intelligence.

This can be summarized as: N = ( T + D ) ^ A for what I’m currently able to program.

So super AI might be S = ( ( ( T + D ) ^ A ) ^ I ) ^ P.

I say T and D is the same value, because you must have the same amount of methods as one is able to train, or you get errors.

Not terribly sure how this will effect my own programming as an individual.

But call this the Sleeping Monkey Calculation: How many sleeping chimps does it take to be equivalent to a weak super-intelligence.

1 Like