I’ve put together a report about my experience at the Loebner Prize this year. The contest has to change next year, as there is no more sponsorship money since Hugh’s death and I encourage you all to comment on Andrew’s thread with any comments or suggestions about the future of the contest.

https://medium.com/pandorabots-blog/mitsuku-wins-loebner-prize-2018-3e8d98c5f2a7

That’s an enjoyable read as always. Your account of the contest event was compelling. From the rest of it, there seems to be no real future for Turing tests. Nobody understands the human equation. When there are so many facets of intelligence that you can’t enumerate them, it is really just too much of a hassle to try and imitate that somehow. Is that what you’re saying? People make cartoons, why compare the cartoon’s behavior to a human?

Think about dogs. They have adapted to living with people, and people recognize emotions and activity in a way that is familiar, which is easily to anthropomorphize the dog as having human traits and characteristics. We do that with other things that we communicate with. Did we adopt the dog’s emotive content as part of our own? Or did dogs figure out how to talk to us in ways that we could understand. Perhaps the natural growth of all our kinds is such that we tend to develop with features that we may share with other beings as humans or dogs because of this simulation we’re in. It is the will of Landru.

Somewhere, Hugh Loebner is ready to collect the Marvin Minsky Prize from somebody.

Best — 5 points
2nd — 3 points
3rd — 2 points
4th — 1 point

Just thinking out loud, one could only be considered 100% human if half the judges had mistaken a bot for a human, which wasn’t the case, so “best” isn’t the maximum score possible on that scale. It would come closer if “best” equalled 50% human (x14/20 = 35%).