Law of Large Numbers states that sum of a large amount of i.i.d variables approaches its mathematical expectation. Roughly speaking, "big samples reliably reveal properties of population".
It doesn't state that "everything can happen in large samples".
Thanks. Memory is more fragile than thought, wrong folder. Updated.
The other rationality quotes thread operates under the rule:
Lately it seems that every MIRI or CFAR employee is excempt from being quoted.
As there are still interesting quotes that happen on LessWrong, Overcoming Bias, HPMoR and MIRI/CFAR employee in general, I think it makes sense to open this thread to provide a place for those quotes.