As an aside, the FAQ at https://bitkeys.work/faq.php seems a bit...wrong and alarmist?

He's implying that you need 2^80 (~1e24) tries to get a 50% chance of "colliding with one of those keys" (using the birthday paradox). Which seems pretty bad given the current network hashrate of ~1e20 H/s.

But... I think you'd only be colliding with one of your other zillion collision attempts, not one of the relatively miniscule ~1e8 addresses actually used. Unless my math is wrong. *shrug*

(1eX = 1*10^X)

More info: https://download.wpsoftware.net/bitcoin-birthday.pdf