Well this is a pretty good social critique, but I would like to tackle the technical side a bit more. 1) As the WorldCoin people know, there is no way a raw ir

Well this is a pretty good social critique, but I would like to tackle the technical side a bit more. 1) As the WorldCoin people know, there is no way a raw iris scan is going to be consistent enough to produce the same hash every time. Every scan of the same person's iris is going to have slight differences compared to previous scans. And a hash, as HN well knows, is designed to amplify slight differences, so it is, if you will, a "different-maker." (It makes two things more different from each other.) So to address this, they're saying "Wait, wait, we'll spread some ML in between like so much peanut butter," with the goal of transforming multiple slightly-different scans of the same iris into something "the same" enough to generate the same hash. So this part is a "same-maker," that makes different things more similar. Aside from how half-assed this sounds (a same-maker and a different-maker fighting each other[0]), now throw in what you already know from every "Turns Out Some Problems Came Up with This Particular Widely-Hyped ML Application" article in HN for the past 5 years. No ML model is perfect and they all have false positives/negatives and the like. Remember this is supposedly for identifying people. What if, before you apply your different-maker, your same-maker makes two different people's irises the same, such that they have the same hash and the different-maker can't tell the difference? Then your shit is fucked there, happy camper. I'm saying, not in this decade, will this work right. 2) If you're ever wondering whether a particular tool or piece of technology is something designed to serve humans and adapt and accommodate itself to humans, vs. something humans are supposed to serve and to which we're supposed to adapt and accommodate ourselves, just look for the handle. Obviously I'm being a little facetious but you get the idea. Someone brought this up elsewhere on the thread and I thought it was spot-on. https://news.ycombinator.com/item?id=30932568 https://news.ycombinator.com/item?id=30932568 [0] ...although it does sound like fertile ground for something GAN-like