Extreme skew

Over 100 billion humans lived on the earth since they have arrived hundred thousand years ago. However, most of the fundamental knowledge they have gained from fire to quantum mechanics over this time period came from an incomprehensibly few number of people, perhaps as few as 100. This is a hard constraint for humans to advance knowledge as it depends very much on the arrival of a unique individual, perhaps in a few generations. Increase in abstract and fundamental knowledge does not seem to depend on anything else except the presence of that special individual who propels humanity across a discontinuity. Forecasting of knowledge progression in such a regime that follows a smooth stochastic evolution with an extremely small probability of a jump, is a futile exercise. 

Why is this the case? The variation in intelligence across humans appear very small and their physical abilities as good as perfect cloning. But a single person out of one billion appears to possess special powers, albeit all measurable characteristics of that individual are within expected norms. If the probability of this error solely depends on quantity and not on time, then, there could be as many as 7 such individuals who could be present in today's world. This does not seem to be the case as there has been little advancement in fundamental knowledge for close to hundred years.

Systems that exhibit time based errors are similar to electrical systems such as computers and not mechanical systems that can move. The hundred errors seen over 100 thousand years appear evenly spaced in time and show no acceleration in the presence of much larger number of contemporary experiments. If these errors were part of a natural process such as evolution, humanity would have produced a dozen such fundamental leaps in a single generation by now. However, all indications are that there is no difference in the arrival rates of the genius who changes the human mind.

If the universe were a rules based simulation or driven from a random initial state, then larger experiments will unambiguously lead to larger number of errors. It appears to be programmed to regime shift only with time, implying a game with possibly prescribed and controlled end outcomes.