XPost: talk.politics.misc, alt.survival, alt.politics
XPost: soc.culture
https://www.foxnews.com/media/elon-musk-develop-truthgpt-warns-civilizational-destruction-ai
. . .
[Musk]
"I'm going to start something which I call TruthGPT, or a
maximum truth-seeking AI that tries to understand the nature
of the universe," Musk told Carlson. "And I think this might
be the best path to safety in the sense that an AI that
cares about understanding the universe is unlikely to annihilate
humans because we are an interesting part of the universe."
CAREFUL Elon ! The Truth may be that we are NOT all that
interesting to whatever kind of intelligence you're creating,
no more than any other bug to be squashed underfoot.
Carlson noted during "Fox & Friends" that Musk helped fund
the development of ChatGPT in order to "control it," but
unfortunately, it didn't pan out that way.
"Effectively, that was taken away from him, and now you have
it as a commercial product owned by Microsoft and Google,"
Carlson said Monday. "But his position has never changed,
which is if we don't put the brakes on this, it could end
up destroying everything that we have. This is as dangerous
as nuclear weapons."
The software is programmed by left-wing experts, which
train the chatbots to lie, Musk explained during the interview.
Carlson said the partisan influence on the app's development
can even be characterized as an "attack on democracy" through
the "control" of American minds - all through the avenue of
access to information.
. . .
That "Chat" and friends are being taught to slant Truth,
even lie to protect the 'truths' of their masters, IS
very concerning. It would be the most insidious and
expansive type of disinfo/propaganda ever invented.
We've already seen these things make up 'references'
to support their claims - probably copying what people
tend to do these days. The next half-gen of Chat will
be able to WRITE those faux-references and insert them
into the stream in a way difficult to dislodge/expose.
My biggest concern about Musk's vision of AI as an
angel of death however is that it's almost surely
not going to be THAT way. The destruction will not
be by force of arms. We are already beginning to
see the true destructive force in the plans of
McDonalds to replace all humans and turn their
joints into nothing but big AI-operated vending
machines. WalMart plans to de-humanize it's whole
warehouse/shipping operation.
It is JOBS that are most at risk - and not just
burger-flipping jobs but a lot of professional/semi-pro
types of people jobs.
Vast quantities of jobs do not really require
hands-on interaction but are instead info-centered.
In any big corp likely 75% of jobs are clerical or
managerial, perhaps 'research' - they can ALL go.
The AI won't form a union, won't call in sick, won't
steal office supplies, won't sue, does not need
insurance plans or OSHA protection. Biz Logic sez
"Dump All Humans".
This IS a fatal trap - unemployed people cannot BUY
what the corps are selling. Brief profits turn into
a meteoric crash of the global economy. Of course the
top management/boards will all pop their golden
parachutes and land at their survival mansions in NZ
or where-ever. They won't be as safe as they think
however because soon their money will be toilet paper.
Hopefully the Kiwi's will eat them ...
ANYway - normally I avoid Tucker Carlson, gone kinda
'round the bend IMHO. However I will watch THIS interview
because Musk is hardly alone in his concerns - large
numbers of top tech/sci people back him. The current
trajectory of 'AI' might be even more dangerous than
China/NK/Russia - and attack from unexpected directions.
--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)