From Newsgroup: comp.programming

On 8/14/25 2:14 PM, Bobbie Sellers wrote:

On 8/14/25 00:32, c186282 wrote:

On 8/14/25 1:32 AM, Bobbie Sellers wrote:

On 8/13/25 21:45, c186282 wrote:

On 8/13/25 6:27 AM, The Natural Philosopher wrote:

On 12/08/2025 19:47, Charlie Gibbs wrote:

On 2025-08-12, c186282 <c186282@nnada.net> wrote:
;

'The rhetoric was, if you just learned to code, work hard
and get a computer science degree, you can get six figures
for your starting salary,' the San Roman, California native
told The New York Times.

;
https://cheezburger.com/7562537216/work-hard-for-what
;

;     If you can fix toilets then you probably have a
;     fairly secure career future ... but anything
;     white-collar/'office', soon industrial/factory -
;     nope. ALL the corps want to DUMP the humans.

;
That's been a dream of PHBs for decades.
"Managing programmers is like herding cats."
;

;     Not sure who'll be left to BUY their stuff - but
;     they're not thinking that far ahead.
;
;     Also not aware of ANY govt with a plan to deal
;     with all the obsolete humans. Directions to the
;     nearest Soylent Green factory maybe ?

;
Yet for some reason they're obsessed with breeding or
importing ever-increasing numbers of the pesky things.
See my .sig.
;

That's to try and square the debt circle.
Having no idea how to create wealth, they are simply

;  > importing people ....
;
;    Well, despite AI/automation claims, humans are still
;    MUCH better than high tech at a lot of things - esp
;    if it involves digging dirt or crawling thru pipes or
;    welding weird shit together and such.
;
;    This may not, will not, last FOREVER - but it's
;    the current reality.
;
;    The conspiracy theorists (esp USA) have all sorts of
;    political theories about 'immigrants'. This is not
;    entirely the real picture. Natives WON'T do shit work,
;    certainly not for an affordable wage, but there's
;    no replacement FOR that labor. So - you import it by
;    any means.
;
;    Downside ... gen-2 'immigrants' will want $49/hr to
;    pick lettuce. This makes them useless - and just an
;    added burden. We're looking at last-ditch short-term
;    'fixes' here. Then it all goes to hell.
;
;    Will AI/bots be ready to do all the stupid shit ?
;    Ummmm ... NOT clear at all. Invest trillions if you
;    want but Reality is NOT forgiving. Lots of 'human
;    stuff' just plain requires HUMANS ... maybe for 50
;    years at least.
;
;    Then the SoftBots ...
;
;    As mentioned in some of my recent postings, LLMs
;    now DO seem to have a basic sort of 'self awareness',
;    even a 'survival instinct'. Not exactly a 'human'
;    sort of 'self', but it's now THERE. NNs might do
;    even better, but the right hardware base still
;    isn't there - waiting for the 'perfect device'.

     That survival reflex comes from reading all the Speculative
fictions available onlline where machine intelligence resists being
turned off.>

;    In any case, Chat-5 now can, brags, about being
;    able to do code-to-order. Might not be the very
;    BEST code but, as attributed to several, the
;    phrase "Quantity has a quality all its own".
;
;    Human coders/designers will be a RAPIDLY shrinking
;    segment. What they knew, the AI builders put into
;    their systems. Those systems will rapidly Do Better.
;
;    Chat-6 ... expect it to be writ mostly by Chat-5.
;    No human will really understand the code. We've
;    reached a sort of threshold.
;
;    THEN things get interesting !  :-)
;
;    And expect like a full third of global energy
;    production to go into 'AI' centers. We're
;    seeing kinda unprecedented, insane, levels
;    of investment here - all to dump the annoying
;    Human workers.
;
;    Azimov, Clarke, Dick and friends kind of did us
;    a disservice by predicting good electronic
;    intelligence Too Soon. I know why, mostly the
;    delusions of Minsky (who USED to post to usenet,
;    had dialog LONG back). The too-soon predictions
;    led us to poop on what eventual AI would be
;    like, when it would come, IF it would come.
;
;    Chat, NNs and such are NOT going to be HAL-9000,
;    they're trained on petabytes of 'human stuff'
;    and will thus be a little TOO much like us for
;    comfort and safety. There will be no "three laws".

     We haven't figured out that training LLM on the
Internet will do us no good as it is full of lies and human
prejudices.

;
;    Still have Minsky's "Society Of Mind" ... it
;    seemed SO EASY back then - until the huge
;    complications/depth became obvious  :-)

     If you want a conscious machine then you have to
give it senses so that it develops a true sense of self.
It has to burn its sensors to know that fire is hot and
dangerous, that gravity pulls things down and that the
Sun can be blinding. Also that its parts modeled on
the human body can be broken painfully.

With present LLMs the power requirements are too high
to put it in a mobile body and the possibility of a
remote stable operating computer is limited by
communications speed.

     At least from what I read that is my opinion.

   To a point I agree.

   A "life-like", "human-like", system can't be had
   in a relative vacuum. It HAS to have a "real life"
   in the Real World to a degree.

   They are TRYING to get around that. Don't think
   it's working. The AI has a 'too abstract' base
   of experience - decidedly NOT Us.

   This MIGHT be bad, or good, or "Who Cares" - we
   can't tell yet.

   Better figure it out - quick.

   IMHO we do NOT want "e-humans". Humans are Just
   Awful - all through history.

   But, then, WHAT ???

    Ever read SF? Jack Williamson Humanoids are just like
your mom always trying to get you to give up bad habits
and dangerous ideas.  In person they are sleek, slim black
robots. The story is that they look for wild humans to
domesticate them. They would take care automatically
of the disabled and ill and apply the cures for all the
human aliments which they have divised since their
creation. I would like one please but I am 88.
They are all out "for your own good" but unlike the
humans who tried to do things for your own good
they really mean to help you.

But ... what makes you believe we can MAKE such bots ???
The functional IQ needed for that role means the IQ needed
to rationalize around all constraints/conditioning/rules
as well.

We were promised robo-slaves - but I think we're going
to GET something rather different.

A month or so ago, there was a tech news blurb about
a LLM - OpenAI I think - that deliberately searched,
found and disabled it's official shut-down script.

That speaks volumes. By Whatever Means we HAVE now
created sort of 'alien' intelligences/'selves'.
They look at US, petabytes, but in that they also
see themselves. I suspect that 'self-reflection'
has evolutionary properties.

There seem to be two or three other ways to
create such intelligences. Think they're gonna
be your Mommy ???
--- Synchronet 3.21a-Linux NewsLink 1.2