Humanist Discussion Group, Vol. 37, No. 530.
        Department of Digital Humanities, University of Cologne
                      Hosted by DH-Cologne
                       www.dhhumanist.org
                Submit to: humanist@dhhumanist.org




        Date: 2024-04-05 17:27:24+00:00
        From: James Rovira <jamesrovira@gmail.com>
        Subject: Re: [Humanist] 37.527: talking to & from smart machines

I'm not sure I understand the question. If you define the answer in terms
of outputs, then you just need better prompts. There have been good answers
along these lines already. But a better output by an AI, one that resembles
a genuine conversation, still isn't a genuine conversation. The AI is
always calculating a probability. Genuine conversations with real people
only begin when the person consciously chooses to quit calculating the
probabilities of the outputs they desire from you, while the AI is blank,
completely, in terms of any expectations from you, just constantly
calculating outputs in response to prompts.

Jim R

On Fri, Apr 5, 2024 at 2:58 AM Humanist <humanist@dhhumanist.org> wrote:

>
> > I'm not out to establish the inferiority of the machine, however smart,
> > rather to question what would need to be done to give an LLM the ability
> > to engage in a "genuine conversation", as Gadamer says--one in which the
> > user was not so much in control of, and the LLM not so much eager
> > flatteringly to please (and so to help keep the research funding
> flowing).
> >
>
>


_______________________________________________
Unsubscribe at: http://dhhumanist.org/Restricted
List posts to: humanist@dhhumanist.org
List info and archives at at: http://dhhumanist.org
Listmember interface at: http://dhhumanist.org/Restricted/
Subscribe at: http://dhhumanist.org/membership_form.php