I've heard of them, at least in other situations. They're somewhere between stupid and idiotic.

What this person is doing is carefully setting up his challenge so that it sounds scientific, but setting the goalposts at an absurdly high level. He's essentially demanding that people prove that evolution was able to pick a needle in a haystack, as well as demanding an explanation for why chimpanzees differ by 150 million base pairs (5% of the genome) whereas humans only differ by 3 million (0.1% of the genome). He also makes the classic mistake of demanding an explanation for why the total number of possible permutations exceeds the number of electrons in the visible universe and thus was able to 'defy' science and mathematics to produce a specific species.

Yeah. I don't think he really understands how statistical chance works, never mind evolution.

I can give a fairly concise explanation for why his reasoning is so flawed using only 20-sided dice. The chances of rolling any number on a d20 is 5%. Well, each successive roll also has only a 5% chance of coming up. Yet the chances of those two rolls coming up as they did is .05 * .05, or .0025 (0.25%). Add a third d20; still a 5% chance of getting any particular number, but it decreases the odds of that specific combination to .000125 (0.0125%). Add a fourth, and the odds decrease to .00000625 (0.000625%). Add a fifth, and the odds decrease to .0000003125 (0.00003125%). Just five d20 rolls, and you have a phenomenally small chance of that particular set of numbers having come up. Barely three-hundred-thousandths of a percent of a chance.

If I add just one more d20 roll, that decreases the odds of coming up with that particular set of numbers to a bit better than one-in-a-million. Amazing...you can get a one-in-a-million (10^{7}) chance by rolling 6d20. I wonder how many dice rolls you'd have to add to get 2^{150000000}?