Re: Remember "Bit-Slice" Chips ?

On 12/20/24 4:21 PM, D wrote:

  On Fri, 20 Dec 2024, rbowman wrote:
 

On Fri, 20 Dec 2024 01:31:30 -0500, 186282@ud0s4.net wrote:
>

   It's the 'hand-wave' thing that sunk the first AI paradigm.
   Marv Minsky (who posted on usenet for awhile) and friends saw how
   easily 'decisions' could be done with a transistor or two and assumed
   it would thus be easy to build an AI. AC Clarke used the Minsky
   optimism when fashioning the idea of "HAL".

>
Minsky threw a wrench in the works with his 9169 'Perceptrons'. He had
tried to implement B. F. skinner's operant condition with a analog lashup
that sort of worked if the vacuum tubes didn't burn out. Rosenblatt has
built a 'Perceptron' and Minsky pointed out original design couldn't
handle an XOR. That sent research down another rabbit hole.
>
By the '80s the original perceptron had evolved into a multilayer network
train by back propagation. When I played around with it 'Parallel
Distributed Processing' by Rumelhart and McClelland was THE book.
>
https://direct.mit.edu/books/monograph/4424/Parallel-Distributed-
Processing-Volume
>
The ideas were fascinating but the computing power wasn't there. Most of
what I learned then is still relevant to TensorFlow and the other neural
network approaches except now there are the $30,000 Nvidia GPUs to do the
heavy lifting.
>
The '80s neural networks weren't practical so the focus shifted to expert
systems until they petered out. The boom and bust cycles led to the term
'AI Winter'
>
https://www.techtarget.com/searchenterpriseai/definition/AI-winter
>
I think something worthwhile will come from this cycle but ultimately it
won't be the LLMs that are getting all the hype.

 I wonder if not facebook, open sourcing their llm threw quite a wrench in the Open AI machinery this time.
 Open AI:s ai is stagnating, and I think perhaps the development of the open source models will be good enough so that open ai might not be able to recoup all the massive amounts of money that has been invested in them.
 Then another ai winter, and after that, our dear llms might be ready for prime time!

   I think what LLMs do is a PART of 'intelligence/self',
   just not ALL of it. OTHER methods/layers maybe CAN
   be spliced in to fill the weak bits.
   Brains are an evolutionary hodgepodge - 'whatever was
   needed/worked'. 600+ million years of field testing.
   Kinda amazed they work at all. There's also a weird,
   almost 'holographic', nature to them - some of those
   kids blasted by hydrocephalus, with little grey
   matter left, still managed average or even a bit
   above average IQs. It's the same with 'cerebral
   palsy' cases. They STILL produce a 'person' in
   there. The System WANTS to work.
   Anyway, LLMs contaminated by some NN action and
   e-motions and maybe a few other odd bits ... and
   don't forget to include those hard-wired evolutionary
   sub-routines.
   TIGER !!!