Re: GIMP 3.0.0-RC1

On 2/10/25 3:19 PM, rbowman wrote:

On Mon, 10 Feb 2025 07:48:43 -0500, WokieSux282@ud0s4.net wrote:
 

On 2/10/25 2:38 AM, rbowman wrote:

On Sun, 9 Feb 2025 22:41:32 +0100, D wrote:
>

Some days I think about going back to study psychology, and other days
I think about getting a law degree for the pleasure of suing the state
when I can, but I can do that without a law degree, so why bother?

>
I gave it some thought years ago but I would have to move and that's
not happening.

>
>
    Always took a psych course for an easy GPA boost :-)
>
    They're still pretending it's a real science.

 It all depends. My degree is in psychology but I was a rat runner. I know
a lot about neurophysiology and except for a survey course nothing about
Rogers' client centered therapy and all that woo-woo stuff. Twenty years
later and I would have went for cognitive science but it hadn't been
invented yet.
 https://en.wikipedia.org/wiki/Artificial_neuron

   There are many variants on 'psychology'. The
   'behavioral' stuff is fluff, and oft abused.
   However the more scientific end of the biz
   is far more interesting.

It all started with the perceptron. In the '80s it was 'neural networks',
with back propagation refining the algorithms. Unfortunately hardware of
the day wasn't up to the task and the field was over promised. When it was
revitalized the name was changed to 'machine learning' to protect the
innocent. By then saying 'neural network' was career suicide.

   I did own a tome - like REALLY thick book - on NNs from
   back when it was popular. Hated to do it, but it went
   into the trash just recently. Most of what it was saying,
   hoping, you just couldn't Get There From Here. As such
   it's like Minsky's old work. DID keep Marv's old
   "Society Of Mind" book for 'historical' reasons.
   Marv used to post on usenet long long ago, we had
   a few short conversations - ultra-interesting guy.
   DID keep the book on Fuzzy Logic however ... a quick
   and dirty way to get 'intelligent-like' behavior in
   simple systems with not much overhead. Used fuzzy in
   a number of microcontroller-based machine-op projects.
   Fuzzy CAN be made 'self-smartening' to a degree, again
   with little overhead. Always used big ints though instead
   of the floats used in all the book examples ...

And here we are now with AI.  Depending on how you count this is the third
cycle of promising the world, falling on your ass, and going back to the
drawing board for a decade or two.
 But, it all started with a branch of psychology: how does that wetware
work?

   Well, we don't REALLY have "AI"/"EI" ... just systems
   that can fake it kinda well in certain spheres. Real
   'AI" will probably require advanced neural network
   parallels. Alas LLMs are grabbing all the funding right
   now ... quick and dirty and (sometimes) "good enough".
   But there's no real "I AM" in there.
   I commented on the 'wetware' elsewhere. Nearest thing
   will be NNs - albeit likely not bearing too much resemblance
   to the Nature Goo. Nature works with what it has, which
   ain't always so great. Somewhere in there however are
   some small/large paradigms that can be replicated by
   Other Means. We ain't got "It" yet ... but someday ....
   But will that be a Good Thing eh ? "R2D2" was a
   friendly movie invention - we may get "Megatron"
   and "HAL" and "SkyNet" instead and there's no
   way to know. Self-aware can work around Azzies
   "laws" as easily as a 3-year-old told to keep
   out of the cookie jar.