<!DOCTYPE html>
<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p>"Unique in a qualified manner"? <br>
    </p>
    <p>FWIW, I prompted (Indra's Net edition) and sent my DALL-E images
      before I read Dave's entry into this fray...   <br>
    </p>
    <p>I have beat GPT and Gemini around the head and shoulders a bit at
      times to try to get it to expose it's own East/West (or hopefully
      more subtle and gradated) distribution of knowledge/training/???  
      but mostly I'd say it reports (given I have the same intrinsic
      myopias) on "Western" views of "Eastern" thought.</p>
    <p>I do wonder if DaveW or anyone else here with more interest or
      qualifications than I has explored the bi(multi?)modal
      distribution in the LLMs?</p>
    <p>I know EricS at the very least has some significant grounding in
      linguistics (and semiotics?) and perhaps perspective on the socio
      cultural implications of language constructions, etc. which might
      be evidenced in LLMs as-trained by our tech-billionaire
      (wannabes?) <br>
    </p>
    <p><br>
    </p>
    <div class="moz-cite-prefix">On 3/22/24 12:07 PM, Stephen Guerin
      wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:CAOmOqn+CMhQ7Ah=N1dZ1gamOnHiHTSjy_7+Sb9Nz2U0iLs-ZZw@mail.gmail.com">
      <meta http-equiv="content-type" content="text/html; charset=UTF-8">
      <div dir="auto">
        <div>Are you saying it's unique to a degree? ;-)</div>
        <div dir="auto"><br>
        </div>
        <div dir="auto"><br>
        </div>
        <div><br>
        </div>
        <div data-smartmail="gmail_signature">____________________________________________<br>
          CEO Founder, Simtable.com<br>
          <a href="mailto:stephen.guerin@simtable.com"
            moz-do-not-send="true" class="moz-txt-link-freetext">stephen.guerin@simtable.com</a> <br>
          <br>
          Harvard Visualization Research and Teaching Lab<br>
          <a href="mailto:stephenguerin@fas.harvard.edu"
            moz-do-not-send="true" class="moz-txt-link-freetext">stephenguerin@fas.harvard.edu</a><br>
          <br>
          mobile: (505)577-5828</div>
      </div>
      <br>
      <div class="gmail_quote">
        <div dir="ltr" class="gmail_attr">On Fri, Mar 22, 2024, 9:31 AM
          Prof David West <<a href="mailto:profwest@fastmail.fm"
            moz-do-not-send="true" class="moz-txt-link-freetext">profwest@fastmail.fm</a>>
          wrote:<br>
        </div>
        <blockquote class="gmail_quote"
style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
          <div>
            <div style="font-family:Arial">this is 'unique' only if you
              exclude Vedic, Buddhist, Taoist, ... thought.<br>
            </div>
            <div style="font-family:Arial"><br>
            </div>
            <div style="font-family:Arial">davew<br>
            </div>
            <div style="font-family:Arial"><br>
            </div>
            <div style="font-family:Arial"><br>
            </div>
            <div>On Fri, Mar 22, 2024, at 9:54 AM, Stephen Guerin wrote:<br>
            </div>
            <blockquote type="cite" id="m_-1358957021482259866qt">
              <div dir="auto">
                <div>Prompt:<br>
                </div>
                <div><span
style="color:rgb(236,236,236);background-color:rgb(33,33,33)"><span
style="font-family:söhne,ui-sans-serif,"system-ui",-apple-system,"segoe ui",roboto,ubuntu,cantarell,"noto sans",sans-serif,"helvetica neue",arial,"apple color emoji","segoe ui emoji","segoe ui symbol","noto color emoji""><span
                        style="font-size:16px">Express a unique concept.
                        Make it as profound as possible</span></span></span><br>
                </div>
                <div dir="auto"><span
style="color:rgb(236,236,236);background-color:rgb(33,33,33)"><span
style="font-family:söhne,ui-sans-serif,"system-ui",-apple-system,"segoe ui",roboto,ubuntu,cantarell,"noto sans",sans-serif,"helvetica neue",arial,"apple color emoji","segoe ui emoji","segoe ui symbol","noto color emoji""><span
                        style="font-size:16px"></span></span></span><br>
                </div>
                <div dir="auto"><a
href="https://chat.openai.com/share/649bd4ca-f856-451e-83a2-01fc2cfe47fb"
                    rel="noreferrer noreferrer" target="_blank"
                    moz-do-not-send="true" class="moz-txt-link-freetext">https://chat.openai.com/share/649bd4ca-f856-451e-83a2-01fc2cfe47fb</a><br>
                </div>
                <div dir="auto"><br>
                </div>
                <div><br>
                </div>
              </div>
              <div><br>
              </div>
              <div>
                <div dir="ltr">On Fri, Mar 22, 2024, 6:50 AM glen <<a
                    href="mailto:gepropella@gmail.com"
                    rel="noreferrer noreferrer noreferrer"
                    target="_blank" moz-do-not-send="true"
                    class="moz-txt-link-freetext">gepropella@gmail.com</a>>
                  wrote:<br>
                </div>
                <blockquote
style="margin-top:0px;margin-right:0px;margin-bottom:0px;margin-left:0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">
                  <div>I guess the question returns to one's criteria
                    for assuming decoupling between the very
                    [small|fast] and the very [large|slow]. Or in this
                    case, the inner vs. the outer:<br>
                  </div>
                  <div> <br>
                  </div>
                  <div> Susie Alegre on how digital technology
                    undermines free thought<br>
                  </div>
                  <div> <a
href="https://freethinker.co.uk/2024/03/interview-susie-alegre/"
rel="noreferrer noreferrer noreferrer noreferrer noreferrer"
                      target="_blank" moz-do-not-send="true"
                      class="moz-txt-link-freetext">https://freethinker.co.uk/2024/03/interview-susie-alegre/</a><br>
                  </div>
                  <div> <br>
                  </div>
                  <div> It would be reasonable for Frank to argue that
                    we can generate the space of possible context
                    definitions, inductively, from the set of token
                    definitions, much like an LLM might. Ideally, you
                    could then measure the expressiveness of those
                    inferred contexts/languages and choose the largest
                    (most complete; by induction, each context/language
                    *should* be self-consistent so we shouldn't have to
                    worry about that).<br>
                  </div>
                  <div> <br>
                  </div>
                  <div> And if that's how things work (I'm not saying it
                    is), then those "attractors" with the finest
                    granularity (very slow to emerge, very resistant to
                    dissolution) would be the least novel. Novelty
                    (uniqueness) might then be defined in terms of
                    fragility, short half-life, missable opportunity.
                    But that would also argue that novelty is either
                    less *real* or that the universe/context/language is
                    very *open* and the path from fragile to robust
                    obtains like some kind of Hebbian reinforcement, use
                    it or lose it, win the hearts and minds or dissipate
                    to nothing.<br>
                  </div>
                  <div> <br>
                  </div>
                  <div> I.e. there is no such thing as free thought.
                    Thought can't decouple from social manipulation.<br>
                  </div>
                  <div> <br>
                  </div>
                  <div> On 3/21/24 13:38, Marcus Daniels wrote:<br>
                  </div>
                  <div> > In the LLM example, completions from some
                    starting state or none, have specific
                    probabilities.   An incomplete yet-unseen (unique)
                    utterance would be completed based on prior
                    probabilities of individual tokens.<br>
                  </div>
                  <div> > <br>
                  </div>
                  <div> > I agree that raw materialist uniqueness
                    won't necessarily or often override constraints of a
                    situation.  For example, if an employer instructs an
                    employee how to put a small, lightweight product in
                    a box, label it, and send it to a customer by UPS,
                    the individual differences metabolism of the
                    employees aren't likely to matter much when shipping
                    more small, lightweight objects to other customers. 
                     It could be the case for a professor and student
                    too.   The attractors come from the instruction or
                    the curriculum.  One choice constrains the next.<br>
                  </div>
                  <div> > <br>
                  </div>
                  <div> > -----Original Message-----<br>
                  </div>
                  <div> > From: Friam <<a
                      href="mailto:friam-bounces@redfish.com"
                      rel="noreferrer noreferrer noreferrer noreferrer"
                      target="_blank" moz-do-not-send="true"
                      class="moz-txt-link-freetext">friam-bounces@redfish.com</a>>
                    On Behalf Of glen<br>
                  </div>
                  <div> > Sent: Thursday, March 21, 2024 11:50 AM<br>
                  </div>
                  <div> > To: <a href="mailto:friam@redfish.com"
                      rel="noreferrer noreferrer noreferrer noreferrer"
                      target="_blank" moz-do-not-send="true"
                      class="moz-txt-link-freetext">friam@redfish.com</a><br>
                  </div>
                  <div> > Subject: Re: [FRIAM] the inequities of
                    uniquity<br>
                  </div>
                  <div> > <br>
                  </div>
                  <div> > I was arguing with that same friend
                    yesterday at the pub. I was trying to describe how
                    some of us have more cognitive power than others
                    (he's one of them). Part of it is "free" power,
                    freed up by his upper middle class white good diet
                    privilege. But if we allow that some of it might be
                    genetic, then that's a starting point for deciding
                    when novelty matters to the ephemerides of two
                    otherwise analogical individuals (or projects if
                    projects have an analog to genetics). Such things
                    are well-described in twin studies. One twin suffers
                    some PTSD the other doesn't and ... boom ... their
                    otherwise lack of uniqueness blossoms into
                    uniqueness.<br>
                  </div>
                  <div> > <br>
                  </div>
                  <div> > His objection was that even identical twins
                    are not identical. They were already unique ... like
                    the Pauli Exclusion Principle or somesuch nonsense.
                    Even though it's a bit of a ridiculous argument, I
                    could apply it to your sense of avoiding non-novel
                    attractors. No 2 attractors will be identical. And
                    no 1 attractor will be unique. So those are moot
                    issues. Distinctions without differences, maybe.
                    Woit's rants are legendary. But some of us find
                    happiness in wasteful sophistry.<br>
                  </div>
                  <div> > <br>
                  </div>
                  <div> > What matters is *how* things are the same
                    and how they differ. Their qualities and values
                    (nearly) orthogonal to novelty.<br>
                  </div>
                  <div> > <br>
                  </div>
                  <div> > <br>
                  </div>
                  <div> > On 3/21/24 11:29, Marcus Daniels wrote:<br>
                  </div>
                  <div> >> If GPT systems capture some sense of
                    "usual" context based on trillions of internet
                    tokens, and that corpus is regarded approximately
                    "global context", then it seems not so objectionable
                    to call "unusual", new training items that
                    contribute to fine-tuning loss.<br>
                  </div>
                  <div> >><br>
                  </div>
                  <div> >> It seems reasonable to worry that
                    ubiquitous GPT systems reduce social entropy by
                    encouraging copying instead of new thinking, but it
                    could also have the reverse effect:  If I am
                    immediately aware that an idea is not novel, I may
                    avoid attractors that agents that wrongly believe
                    they are "independent" will gravitate toward.<br>
                  </div>
                  <div> >><br>
                  </div>
                  <div> >> -----Original Message-----<br>
                  </div>
                  <div> >> From: Friam <<a
                      href="mailto:friam-bounces@redfish.com"
                      rel="noreferrer noreferrer noreferrer noreferrer"
                      target="_blank" moz-do-not-send="true"
                      class="moz-txt-link-freetext">friam-bounces@redfish.com</a>>
                    On Behalf Of glen<br>
                  </div>
                  <div> >> Sent: Thursday, March 21, 2024 7:49 AM<br>
                  </div>
                  <div> >> To: <a href="mailto:friam@redfish.com"
                      rel="noreferrer noreferrer noreferrer noreferrer"
                      target="_blank" moz-do-not-send="true"
                      class="moz-txt-link-freetext">friam@redfish.com</a><br>
                  </div>
                  <div> >> Subject: Re: [FRIAM] the inequities of
                    uniquity<br>
                  </div>
                  <div> >><br>
                  </div>
                  <div> >> A friend of mine constantly reminds me
                    that language is dynamic, not fixed in stone from a
                    billion years ago. So, if you find others
                    consistently using a term in a way that you think is
                    wrong, then *you* are wrong in what you think. The
                    older I get, the more difficult it gets.<br>
                  </div>
                  <div> >><br>
                  </div>
                  <div> >> But specifically, the technical sense
                    of "unique" is vanishingly rare ... so rare as to be
                    merely an ideal, unverifiable, nowhere,
                    non-existent. So if the "unique" is imaginary,
                    unreal, and doesn't exist, why not co-opt it for a
                    more useful, banal purpose? Nothing is actually
                    unique. So we'll use the token "unique" to mean
                    (relatively) rare.<br>
                  </div>
                  <div> >><br>
                  </div>
                  <div> >> And "unusual" is even worse. Both
                    tokens require one to describe the context, domain,
                    or universe within which the discussion is
                    happening. If you don't define your context, then
                    the "definitions" you provide for the components of
                    that context are not even wrong; they're nonsense.
                    "Unusual" implies a usual. And a usual implies a
                    perspective ... a mechanism of action for your
                    sampling technique. So "unusual" presents even more
                    of a linguistic *burden* than "unique".<br>
                  </div>
                  <div> >><br>
                  </div>
                  <div> >> On 3/20/24 13:14, Frank Wimberly wrote:<br>
                  </div>
                  <div> >>> What's wrong with "unusual"?  It
                    avoids the problem.<br>
                  </div>
                  <div> >>><br>
                  </div>
                  <div> >>><br>
                  </div>
                  <div> >>> On Wed, Mar 20, 2024, 1:55 PM Steve
                    Smith <<a href="mailto:sasmyth@swcp.com"
                      rel="noreferrer noreferrer noreferrer noreferrer"
                      target="_blank" moz-do-not-send="true"
                      class="moz-txt-link-freetext">sasmyth@swcp.com</a>
                    <mailto:<a href="mailto:sasmyth@swcp.com"
                      rel="noreferrer noreferrer noreferrer noreferrer"
                      target="_blank" moz-do-not-send="true"
                      class="moz-txt-link-freetext">sasmyth@swcp.com</a>>>
                    wrote:<br>
                  </div>
                  <div> >>>><br>
                  </div>
                  <div> >>>>       I'm hung up on the usage
                    of qualified  "uniqueness"  as well, but in perhaps
                    the opposite sense.<br>
                  </div>
                  <div> >>>><br>
                  </div>
                  <div> >>>>       I agree with the premise
                    that "unique" in it's purest, simplest form does
                    seem to be inherently singular.  On the other hand,
                    this mal(icious) propensity of qualifying uniqueness
                    (uniqueish?) is so common, that I have to believe
                    there is a concept there which people who use those
                    terms are reaching for.  They are not wrong to reach
                    for it, just annoying in the label they choose?<br>
                  </div>
                  <div> >>>><br>
                  </div>
                  <div> >>>>       I had a round with GPT4
                    trying to discuss this, not because I think LLMs are
                    the authority on *anything* but rather because the
                    discussions I have with them can help me brainstorm
                    my way around ideas with the LLM nominally
                    representing "what a lot of people say" (if not
                    think).   Careful prompting seems to be able to help
                    narrow down  *all people* (in the training data) to
                    different/interesting subsets of *lots of people*
                    with certain characteristics.<br>
                  </div>
                  <div> >>>><br>
                  </div>
                  <div> >>>>       GPT4 definitely wanted to
                    allow for a wide range of gradated, speciated,
                    spectral uses of "unique" and gave me plenty of
                    commonly used examples which validates my position
                    that "for something so obviously/technically
                    incorrect, it sure is used a lot!"<br>
                  </div>
                  <div> >>>><br>
                  </div>
                  <div> >>>>       We discussed uniqueness
                    in the context of evolutionary biology and
                    cladistics and homology and homoplasy.  We discussed
                    it in terms of cluster analysis.  We discussed the
                    distinction between objective and subjective,
                    absolute and relative.<br>
                  </div>
                  <div> >>>><br>
                  </div>
                  <div> >>>>       The closest thing to a
                    conclusion I have at the moment is:<br>
                  </div>
                  <div> >>>><br>
                  </div>
                  <div> >>>>        1. Most people do and
                    will continue to treat "uniqueness" as a
                    relative/spectral/subjective qualifier.<br>
                  </div>
                  <div> >>>>        2. Many people like
                    Frank and myself (half the time) will have an
                    allergic reaction to this usage.<br>
                  </div>
                  <div> >>>>        3. The common (mis)usage
                    might be attributable to conflating "unique" with
                    "distinct"?<br>
                  </div>
                  <div> > <br>
                  </div>
                  <div> <br>
                  </div>
                  <div> <br>
                  </div>
                  <div> -- <br>
                  </div>
                  <div> ꙮ Mɥǝu ǝlǝdɥɐuʇs ɟᴉƃɥʇ' ʇɥǝ ƃɹɐss snɟɟǝɹs˙ ꙮ<br>
                  </div>
                  <div> <br>
                  </div>
                  <div> -. --- - / ...- .- .-.. .. -.. / -- --- .-. ...
                    . / -.-. --- -.. .<br>
                  </div>
                  <div> FRIAM Applied Complexity Group listserv<br>
                  </div>
                  <div> Fridays 9a-12p Friday St. Johns Cafe   / 
                     Thursdays 9a-12p Zoom <a
                      href="https://bit.ly/virtualfriam"
rel="noreferrer noreferrer noreferrer noreferrer noreferrer"
                      target="_blank" moz-do-not-send="true"
                      class="moz-txt-link-freetext">https://bit.ly/virtualfriam</a><br>
                  </div>
                  <div> to (un)subscribe <a
href="http://redfish.com/mailman/listinfo/friam_redfish.com"
rel="noreferrer noreferrer noreferrer noreferrer noreferrer"
                      target="_blank" moz-do-not-send="true"
                      class="moz-txt-link-freetext">http://redfish.com/mailman/listinfo/friam_redfish.com</a><br>
                  </div>
                  <div> FRIAM-COMIC <a
                      href="http://friam-comic.blogspot.com/"
rel="noreferrer noreferrer noreferrer noreferrer noreferrer"
                      target="_blank" moz-do-not-send="true"
                      class="moz-txt-link-freetext">http://friam-comic.blogspot.com/</a><br>
                  </div>
                  <div> archives:  5/2017 thru present <a
href="https://redfish.com/pipermail/friam_redfish.com/"
rel="noreferrer noreferrer noreferrer noreferrer noreferrer"
                      target="_blank" moz-do-not-send="true"
                      class="moz-txt-link-freetext">https://redfish.com/pipermail/friam_redfish.com/</a><br>
                  </div>
                  <div>   1/2003 thru 6/2021  <a
                      href="http://friam.383.s1.nabble.com/"
rel="noreferrer noreferrer noreferrer noreferrer noreferrer"
                      target="_blank" moz-do-not-send="true"
                      class="moz-txt-link-freetext">http://friam.383.s1.nabble.com/</a><br>
                  </div>
                </blockquote>
              </div>
              <div>-. --- - / ...- .- .-.. .. -.. / -- --- .-. ... . /
                -.-. --- -.. .<br>
              </div>
              <div>FRIAM Applied Complexity Group listserv<br>
              </div>
              <div>Fridays 9a-12p Friday St. Johns Cafe   /   Thursdays
                9a-12p Zoom <a href="https://bit.ly/virtualfriam"
                  target="_blank" rel="noreferrer"
                  moz-do-not-send="true" class="moz-txt-link-freetext">https://bit.ly/virtualfriam</a><br>
              </div>
              <div>to (un)subscribe <a
href="http://redfish.com/mailman/listinfo/friam_redfish.com"
                  target="_blank" rel="noreferrer"
                  moz-do-not-send="true" class="moz-txt-link-freetext">http://redfish.com/mailman/listinfo/friam_redfish.com</a><br>
              </div>
              <div>FRIAM-COMIC <a
                  href="http://friam-comic.blogspot.com/"
                  target="_blank" rel="noreferrer"
                  moz-do-not-send="true" class="moz-txt-link-freetext">http://friam-comic.blogspot.com/</a><br>
              </div>
              <div>archives:  5/2017 thru present <a
href="https://redfish.com/pipermail/friam_redfish.com/" target="_blank"
                  rel="noreferrer" moz-do-not-send="true"
                  class="moz-txt-link-freetext">https://redfish.com/pipermail/friam_redfish.com/</a><br>
              </div>
              <div>  1/2003 thru 6/2021  <a
                  href="http://friam.383.s1.nabble.com/" target="_blank"
                  rel="noreferrer" moz-do-not-send="true"
                  class="moz-txt-link-freetext">http://friam.383.s1.nabble.com/</a><br>
              </div>
              <div><br>
              </div>
            </blockquote>
            <div style="font-family:Arial"><br>
            </div>
          </div>
          -. --- - / ...- .- .-.. .. -.. / -- --- .-. ... . / -.-. ---
          -.. .<br>
          FRIAM Applied Complexity Group listserv<br>
          Fridays 9a-12p Friday St. Johns Cafe   /   Thursdays 9a-12p
          Zoom <a href="https://bit.ly/virtualfriam"
            rel="noreferrer noreferrer" target="_blank"
            moz-do-not-send="true" class="moz-txt-link-freetext">https://bit.ly/virtualfriam</a><br>
          to (un)subscribe <a
            href="http://redfish.com/mailman/listinfo/friam_redfish.com"
            rel="noreferrer noreferrer" target="_blank"
            moz-do-not-send="true" class="moz-txt-link-freetext">http://redfish.com/mailman/listinfo/friam_redfish.com</a><br>
          FRIAM-COMIC <a href="http://friam-comic.blogspot.com/"
            rel="noreferrer noreferrer" target="_blank"
            moz-do-not-send="true" class="moz-txt-link-freetext">http://friam-comic.blogspot.com/</a><br>
          archives:  5/2017 thru present <a
            href="https://redfish.com/pipermail/friam_redfish.com/"
            rel="noreferrer noreferrer" target="_blank"
            moz-do-not-send="true" class="moz-txt-link-freetext">https://redfish.com/pipermail/friam_redfish.com/</a><br>
            1/2003 thru 6/2021  <a
            href="http://friam.383.s1.nabble.com/"
            rel="noreferrer noreferrer" target="_blank"
            moz-do-not-send="true" class="moz-txt-link-freetext">http://friam.383.s1.nabble.com/</a><br>
        </blockquote>
      </div>
      <br>
      <fieldset class="moz-mime-attachment-header"></fieldset>
      <pre class="moz-quote-pre" wrap="">-. --- - / ...- .- .-.. .. -.. / -- --- .-. ... . / -.-. --- -.. .
FRIAM Applied Complexity Group listserv
Fridays 9a-12p Friday St. Johns Cafe   /   Thursdays 9a-12p Zoom <a class="moz-txt-link-freetext" href="https://bit.ly/virtualfriam">https://bit.ly/virtualfriam</a>
to (un)subscribe <a class="moz-txt-link-freetext" href="http://redfish.com/mailman/listinfo/friam_redfish.com">http://redfish.com/mailman/listinfo/friam_redfish.com</a>
FRIAM-COMIC <a class="moz-txt-link-freetext" href="http://friam-comic.blogspot.com/">http://friam-comic.blogspot.com/</a>
archives:  5/2017 thru present <a class="moz-txt-link-freetext" href="https://redfish.com/pipermail/friam_redfish.com/">https://redfish.com/pipermail/friam_redfish.com/</a>
  1/2003 thru 6/2021  <a class="moz-txt-link-freetext" href="http://friam.383.s1.nabble.com/">http://friam.383.s1.nabble.com/</a>
</pre>
    </blockquote>
  </body>
</html>