Eliza Bot from 1966

make

#1

Hi Guys,

I was going through my Medium feed today, and found this interesting article named: Your chatbot needs a personality

It talks about an experiment done in the early days of computing called Eliza. Eliza was a virtual psychotherapist. ELIZA did not actually have much intelligence at all. Instead, it cleverly decomposed its users statements and re-purposed them back as open-ended questions, simulating a non-directional conversation.

It made me think about the Role of Bot’s personality in the conversation. Ofcourse this is a rather extreme example, so I ended up doing what any sane person would do. Resurrected the whole thing and made a Bot on top of my own platform.

L&G I give you Eliza on TARS:
https://eliza.hellotars.com/conv/SkwiDj/


#2

Very cool. I would be interested in how you achieved that. Specifically when I asked “I have a lot of questions about the universe” and the bot responded with, “So you have a lot of questions about the universe?” Or something really close… How did you get it to drop the “I” in the user response to make that work?


#3

Hey @Levi,

I actually did nothing, but use this node library: https://www.npmjs.com/package/elizabot
made a simple API out of it and used it with our Bot over the API-Config section. That’s it.

If you want to understand how the logic of Eliza Bot works internally you can read more about it here:


#4

In the early 1980s, I typed that entire original ELIZA program into my Apple II+ home computer (in BASIC) and chatted with her. Thank you for this fun blast from the past.


#5

That is simply amazing… @lbdsgn
How do i follow you? twitter?


#6

Hi Vinit. Sure, I’m on Twitter using @lbdsgn, although I don’t spend much time there.
Thanks for being impressed that I’m simply old :wink:

all best,
–Lee