Comment Re:My only question: does it work at Google-scale? (Score 1) 91
the finite number of minigames they set up with their finite number of items in them, rendering the whole thing pretty useless.
There might not be a benefit to that outcome, but a "good" CAPTCHA system does have a good outcome when it's broken.
I was talking to the guy who started reCAPTCHA many years ago, and his idea was that the OCR work they were farming out was too tough for algorithms to beat. As long as bots could not do better than humans, reCAPTCHA would be offering a valuable service. As soon as the bots were as good as the humans, accurate OCR had been solved, and reCAPTCHA had made that happen, so it was also a win, and he'd have to come up with another CAPTCHA.
I tend to shy away from helping Google StreetSpy on people, and use the audio CAPCHA when available now, but more people are doing the street number thing, which could still be used for good (if we trust Google). And if the bots solve that, maybe their algorithms could be applied to ambulance services, or whatever.
I'm not sure that the TFA's proposals "solve two problems" the way that great engineering solutions universally do. But there are certainly worthy ones out there.