@cryptonaut
I didnt really get the whole proof of human thing on that episode TBH. Kind of just sounded like the guy had no real idea of what he was doing or how we was actually going to implement it, and just writing a "whitepaper" about his idea.
captchas have been around for quite a while now, most sites use them.. its not a sure fire way to tell a person and computer apart though, there are a lot of them that can be cracked by bots
I agree about the end of this episode, I had a really hard time following along with the description of the concept and agree with your thoughts. However, the idea of proof of human kind of sat with me and got me thinking about why something like that would be useful and how could you even possibly implement something like it.
I think I possibly have an idea of why something proof of human would be useful, but am still working on understanding how. Later in the paper I linked to earlier, the author(s) mentioned something called "stealing cycles from humans" as a way to defeat captchas. So, someone that wants to defeat a captcha could just reroute that captcha to an actual human that needs to solve it for some kind of reward. This paper was written in 2000, so I am not sure if this is actually a thing people still do to solve captchas, but what they then go on to say later in the article is honestly a pretty cool idea and got me thinking.
They throw out the idea of an application of "stealing a human cycle" as a web application that tells you what is the best haircut for you, though they use an outdated app called "hot or not" as a way to do so. However, in a distributed system with "proof of human," perhaps it could be.
A user could for instance upload a head shot picture and place on it several different ideas for haircuts and they submit to a server. These images could be put together as one and sent out to actual human voters as a captcha in a distributed system where they participate in voting on what they believe the best image is. Their incentive to prove they are human and vote by solving the captcha could perhaps represent a nominal amount of the token that represents user fees for using the system. If a human were to ever then find a way to "bot" the system, then it would be made obvious to all because of higher "solve" rates in the system and the "difficulty" of solving that captcha could be changed in some way to level the playing field again.
I don't know if something like this is even feasible. Like many other things with cryptocurrency, just fun to go on a thought experiment, I suppose!
