Why even use Eugene? It’s a crappy chatbot to say the least.
“10 out of 30 judges were fooled”
My hobby mail grabber chatbot with ~20 premade answers fooled 400 people to give me their email address while 1 gave a bogus one because he figured it out, so what, did I achieve 99.75% success rate? Is my 120-LoC mirc script better than Cleverbot and Eugene combined?
I don’t think it is like to compare the two. They serve two different purposes in the domain of AI/ computing personalities.
Why even use Eugene? It’s a crappy chatbot to say the least.
“10 out of 30 judges were fooled”
My hobby mail grabber chatbot with ~20 premade answers fooled 400 people to give me their email address while 1 gave a bogus one because he figured it out, so what, did I achieve 99.75% success rate? Is my 120-LoC mirc script better than Cleverbot and Eugene combined?
The difference is that in the Turing test, the judges all know that there is a possibility they are talking to a machine.