This guy has a great sense of humour: "How many of you have filled out a CAPTCHA? How many of you find them really, really annoying? Yeh, I invented that"
He is also an amazing professor / educator. His class at Carnegie Mellon (15-251: Great Theoretical Ideas in Computer Science) is one of the hardest yet most entertaining CS courses at CMU. One of the first assignments involves a scavenger hunt with a randomly assigned team, across campus, based on challenging algorithmic puzzles inspired from many different areas of computer science.
In one semester, he found people who were copying on their homeworks by creating a website with solutions to a problem on an early assignment. Many people fell into the trap, even after he had given a very clear disclaimer to not use search engines to solve homework problems.
The solution is provided by the simulation: to consider your future self and how happy you will be. I found it insightful and gave me pause for thought.
Can't agree on number 4 that invented captcha's. I saw that talk and he talks how annoying captcha was and how re-captcha would be so much better because it is actually used for a purpose. There is one word they know the meaning of and they use your translation of the word that they don't know the meaning of.
It's just horrible that in a time of great motivation that they still try to push a security issue to the end user. And for me they made it even worse with re-captcha because now you have to type 2 words.
A solution would be to know which is a computer and which is a real user, it's not impossible to do so.
"A solution would be to know which is a computer and which is a real user, it's not impossible to do so."
It seems like you have revolutionary knowledge that could change the course of human history. Or you're just trolling. Since you are not pointing to any actual approaches and hundreds of thousands of engineers have not yet come up with what to you is just "a solution" that doesn't even require further specification, I will go with trolling.
Figuring out remotely whether you are dealing with "a human" without any interaction (or, particularly, interaction less than a captcha, or even -gasp!- a two word captcha) is very definitely impossible. Think about it - how do you know whether it's a human? Well, you say, the black box device that he is sitting at has made sure that he is! And how does that machine transmit data to you? Over that data line that this supposed human is sitting right next... I guess you will see where I'm going with this.
I recommend watching "Tim Harford: Trial, error and the God complex" from this list.
Makes one think of self-assured designers vs. AB testing, how trial and error is much better approach to the design of a site than God complex approach.
