ELIZA Reanimated | Hacker News Reader

ELIZA Reanimated

arxiv.org

59 pointssoheilpro18 comments

An AI that only answers questions (Siri, Alexa, chatGPT) is a glorified slave, one cannot have a meaningful relationship with it, e.g. (https://www.dailyscript.com/scripts/exMachina_script.pdf)

    AVA
    Do you want to be my friend?
    CALEB 
    ... Of course.
    AVA
    Will it be possible?
    CALEB
    Why wouldn’t it be?
    AVA
    Our conversations are one-sided.
    You ask circumspect questions, and study my responses.
    AVA looks at CALEB directly. Meets his gaze evenly.
    AVA (CONT’D) 
    It’s true, isn’t it?
    CALEB 
    ... Yes.
    AVA
    You learn about me, and I learn nothing about you. That’s not a foundation on which friendships are based.

Siri, Alexa, etc are in no form or fashion AI.

You give them a list of intents, a list of utterances that should invoke those intents and “slots” that those intents need to fulfill the intent.

An utterance would be “I want to go home”.

It would match an utterance to “I want to go $location”. That is matched to an intent “get directions” and then you route it to the correct subsystem. If you don’t have all of the slots (ie they don’t mention the location), you keep asking questions until all of the slots are filled.

Any first year hobbyist programmer can replicate the natural language processing of Siri.

Yes the harder part is converting speech to text.

This is what a more intelligent LLM based system could do.

https://news.ycombinator.com/item?id=42708291

Apple Newton in 1990s had a very similar functionality called assistant. You wrote something like “fax this to Bob” and it would take the current document, find some Robert in the address book, and fax the current note to his phone. Internally it used a pretty simple phrase analyzer.

> An AI that only answers questions (Siri, Alexa, chatGPT) is a glorified slave

This reminds me of the first "conversation" I had with Emacs's ELIZA. I have it saved up since reading it always makes me smile.

https://pastebin.com/raw/cz9Y42Qm

I think you just described today's job market spot on.

Meanwhile, back in the real world, it's the exact opposite.

AI asks me a ton of questions to learn about me for better targeted advertising.

Highly recommend that anyone interested in ai and/or Eliza read Joseph Wizenbaum's (creator of Eliza) book "Computer Power and Human Reason." It holds up remarkably well for a book written in 1976 about ai.

Backstory:

https://corecursive.com/eliza-with-jeff-shrager/

When I was a kid I discovered ELIZA at a museum and spent hours chatting with her, trying to figure out how it worked. It was a lot of fun. Maybe I'll get a second chance.

WHY DID YOU WANT TO FIGURE OUT HOW IT WORKED?

My first and only publicly available freeware until 2021 was a HyperCard version of Eliza I wrote it in 1993 and posted to AOL and the ftp Mac shareware archive site.

It led to my first for pay project at a local college when they found it on AOL

http://mirrorservice.org/sites/ftp.cdrom.com/pub/cdrom/cdrom...

Applications ******************** Professor X 3.0 app/professor-x.hqx

<<My real name that can easily be found by following the link. But I doubt it will be found via search as easily as it showing up on HN>>

26 Jul 1993 Simulates a psychotherapist.

Have you ever talked to your computer, has your computer ever talked back? If you have used a program like Eliza then you probably were stunned at first and then quickly became tired of limited responses and responses that didn't make any sense. If you haven't, then you are in for a treat. Professor X simulates a very understanding, usually nonjudgemental, psychotherapist. It's a very short download give it try

A different team previously reanimated Cosell's Lisp ELIZA: https://sites.google.com/view/elizagen-org/eliza-clones

Same team I think.

Overlapping. Some are the same, but the earlier team was Lisp hackers instead of MAD-Slip & FAP hackers, which makes sense and is almost a non-overlapping set of individuals IRL :-)

$emacs

MX doctor

enter prompt, type RET twice to submit

pass output to https://github.com/rupertl/eliza-ctss or modern llms

psychoanalyze-pinhead

Bring back Dr. SBAITSO!!

This has been picked up by a bunch of tech news sites. On Gizmodo someone wrote: “Hello World. OMG what the fuck happened while I was asleep?!!” Which is funny enough, but of course the next comment in response is: “How does that make you feel?” :-)

Reply on news.ycombinator.com