Hacker News Reader

I'm wondering if anyone has used optical illusions as a way to screen for bots.

what's the rate of a given illusion working consistently across general population? Probably not high enough to make that approach work.