In case you guys didn’t get this I made a chatbot which tries to impersonate me. It was a pretty fun casual project.

It’s interesting how AI becomes total uncanny valley when you know how much it is filling in the gaps of things. Obviously this thing isn’t me… but it’s trying hard and it has a tinge. Hard to explain.

I trained it only on Twitter data, I’m sure it would act different if I added my notes as context. Give it a spin and see if it’s accurate or not.