OK, I’ve had a play for myself and I think it’s pretty obviously a bot, because you can trick the text parsing - I made up some words and asked it the meaning of them:
When I asked it “What is Glar?”, it gave me responses such as:
“A facial expression used in anger.” (it interpreted ‘Glar’ ‘Glare’)
“Glau?” (it interpreted ‘Glar’ as ‘Glau’ - the surname of an actress)
“It stands for: Genetic Lifeform and Disk Operating System.” (it interpreted ‘Glar’ as ‘GlaDOS’ - a character from the game ‘portal’)
“Happiness. Joy. That warm fuzzy feeling.” (it interpreted ‘Glar’ as ‘Gladness’ - maybe?)
“Gladiator is a movie.” (it interpreted ‘Glar’ as ‘Gladiator’")
“It is a famous town in england. There is a big festival there in the summer.” (it interpreted ‘Glar’ as ‘Glastonbury’)
I asked it “What is Snid?”, it replied:
“Frozen rain” (it interpreted ‘Snid’ as ‘Snow’)
“Snow is frozen rain.” (ditto)
“Frozen water.” (ditto)
“Saturday Night Live” (it interpreted ‘Snid’ as ‘SNL’)
“Water in the sky that has cooled to the point of crystalization that falls as a form of precipitation.” (see above)
“Super Nintendo Entertainment System” (it interpreted ‘Snid’ as ‘SNES’)
“It’s my nose dribble.” (it interpreted ‘Snid’ as ‘Snot’)
And so on. It’s far too consistently easy to trick the text parsing. Turing test fail.
