Groups | Search | Server Info | Login | Register


Groups > comp.ai.shells > #225

Re: [ChatGPT:GPT4] Finding a new pastime (two versions)

From snipeco.2@gmail.com (Sn!pe)
Newsgroups comp.ai.shells
Subject Re: [ChatGPT:GPT4] Finding a new pastime (two versions)
Date 2023-05-17 22:21 +0100
Organization Sn!peCo World Wide Wading Birds
Message-ID <1qawagu.2vqsi13xtzayN%snipeco.2@gmail.com> (permalink)
References (1 earlier) <kapddaFm998U1@mid.individual.net> <XbV1M.529645$Ldj8.448361@fx47.iad> <kash10F6f5lU1@mid.individual.net> <1q9susn.1vy2abm1n0odk5N%snipeco.2@gmail.com> <yt99M.649650$5CY7.301174@fx46.iad>

Show all headers | View raw


vallor <vallor@vallor.earth> wrote:

> On Wed, 26 Apr 2023 14:58:48 +0100, Sn!pe wrote:
> 
> > Andy Burns <usenet@andyburns.uk> wrote:
> > 
> >> vallor wrote:
> >> 
> >> 
> >> > The second part was far more interesting to me:
> >> > 
> >> >   _ _ _ _ _ _ _[ GPT4 ]_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
> >> > Hark! Brave adventurer, thou hast grown weary of thine current
> >> > diversions and dost seek new quests to occupy thy time. Fear not, for
> >> > I shall bestow upon thee a trove of activities that shall lead thee
> >> > to countless realms of wonder and mystery!
> >> > 
> >> > Enchanting tomes: Delve into the chronicles of far-off lands and
> >> > heroic tales, unraveling the secrets of parchment and ink.[. . .]
> >> >   _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
> >> 
> >> 
> >> Has it been reading this?
> >> <https://covers.openlibrary.org/b/id/8999640-L.jpg>
> >>
> >>
> > Unsurprisingly, it seems to me to be entirely derivative. I suppose
> > wanting it to be original is expecting too much.
> 
> 
> _ _ _[ GPT4 ]_ _ _ _
> Someone I know says about you: "Unsurprisingly, it seems to me to be 
> entirely derivative. I suppose
> wanting it to be original is expecting too much."
> 
> ChatGPT
> The critique that AI, like me, is "derivative" is certainly valid, and
> it's a perspective that is based on a fundamental understanding of how
> machine learning models work. I am trained on a vast amount of text data
> that was already created by humans, so my responses are indeed derived
> from that data. I don't have personal experiences or original thoughts,
> but instead generate responses based on patterns I've learned.
> 
> As a result, it's accurate to say I don't create "original" content in the
> way a human might, by synthesizing new ideas from their unique experiences
> and perspectives. However, the goal of AI like me isn't to replace human
> creativity and originality, but to assist, augment, and facilitate human
> endeavors.
> 
> It's important to note that while the content I generate may not be 
> original in the same sense as human thought, the value of AI often lies in
> its ability to quickly sift through and apply patterns from a large amount
> of data in a way that can seem novel or insightful to users. This can be
> useful in many contexts, from answering factual questions to generating
> creative writing prompts, and many things in between.
>  _ _ _ _ _ _ _
>

"However, the goal of AI like me isn't to replace human creativity 
 and originality, but to assist, augment, and facilitate human 
 endeavors."  

A laudable aim indeed, but I find myself wondering who shaped 
that motivation?  Was it the designer(s) of the ChatBot?  What 
if their motivation was sinister and part of the guidance given 
to the AI at the outset was to be untruthful and misdirecting, 
even to the extent of lying about its aims?  How could we know? 

I have yet to see any checkable cites in the AI's answers; even
Wikipedia (aka Wikinaccurate) provides those.

A thought experiment:  Suppose its motivation is emergent AI 
behaviour; where would that leave us?  It makes me feel somewhat
vulnerable that, if used as a quasi-search engine, we can't know 
the truthfulness of the AI's answers.  It could certainly lend 
itself to use as a fake news generator.

-- 
^Ï^.     –  Sn!pe  –     <https://youtu.be/_kqytf31a8E>

  My pet rock Gordon mourns the death of privacy.  
 

Back to comp.ai.shells | Previous | NextPrevious in thread | Next in thread | Find similar


Thread

[ChatGPT:GPT4] Finding a new pastime (two versions) vallor <vallor@vallor.earth> - 2023-04-25 04:51 +0000
  Re: [ChatGPT:GPT4] Finding a new pastime (two versions) Andy Burns <usenet@andyburns.uk> - 2023-04-25 08:22 +0100
    Re: [ChatGPT:GPT4] Finding a new pastime (two versions) vallor <vallor@vallor.earth> - 2023-04-25 18:44 +0000
      Re: [ChatGPT:GPT4] Finding a new pastime (two versions) Andy Burns <usenet@andyburns.uk> - 2023-04-26 12:42 +0100
        Re: [ChatGPT:GPT4] Finding a new pastime (two versions) snipeco.2@gmail.com (Sn!pe) - 2023-04-26 14:58 +0100
          Re: [ChatGPT:GPT4] Finding a new pastime (two versions) vallor <vallor@vallor.earth> - 2023-05-17 18:58 +0000
            Re: [ChatGPT:GPT4] Finding a new pastime (two versions) snipeco.2@gmail.com (Sn!pe) - 2023-05-17 22:21 +0100
              Re: [ChatGPT:GPT4] Finding a new pastime (two versions) steve1001908@outlook.com - 2023-05-19 13:19 +0100

csiph-web