On 12/10/24 23:49, The Horny Goat wrote:

On Mon, 9 Dec 2024 18:06:09 -0800, Dimensional Traveler
<dtravel@sonic.net> wrote:

I watched a new movie, "Subservience" with Meghan Fox on Netflix over
the weekend.  Scared the you know what out of me.  Was even scarier than
"The Terminator".
   https://www.imdb.com/title/tt24871974/?ref_=nm_flmg_job_1_cdt_t_2

And one of the latest versions of AI has shown self-preservation
responses....

Heck that was part of Asimov's Laws of Robotics 50+ years ago

Asimov's 3 Laws were an idea to make us less fearful
of the robots. So far I doubt that any solid basis for installing
the 3 Laws before the 'bots leave the shop exists.

And as I said maybe the Large Language Model was trying
to do something it had read about on the Internet. That would
be initiative which is more interesting than other suppositions.

bliss

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Wed, 11 Dec 2024, Ted Nolan <tednolan> wrote:

In article <830303b8-b8e3-a7ce-6870-838383a401f5@example.net>,
D <nospam@example.net> wrote:

On Tue, 10 Dec 2024, Ted Nolan <tednolan> wrote:

In article <c0fe9880-f1f0-1b6c-678b-3d703cce37e8@example.net>,
D <nospam@example.net> wrote:

-=-=-=-=-=-



On Tue, 10 Dec 2024, Cryptoengineer wrote:

On 12/10/2024 3:38 AM, D wrote:

On Mon, 9 Dec 2024, Dimensional Traveler wrote:

On 12/9/2024 1:21 PM, Lynn McGuire wrote:

On 12/9/2024 1:38 PM, Cryptoengineer wrote:

I know for many it's politically correct to crap on anything Elon Musk
is involved in these days, but its worth keeping up with his companies
accomplishments.

Here's his robot, walking over uneven terrain, up and downhill, and >>>>>>>>> recovering from slips:

https://www.youtube.com/watch?v=0MwiLIIWd-k

I admit it walks like a drunk, but the progress is remarkable. >>>>>>>>>
There's also videos showing it catching a tossed ball one-handed. >>>>>>>>> https://www.youtube.com/watch?v=DswIwg3lxpo

There is competition:
https://www.youtube.com/watch?v=vWXN2bttm4g
https://www.youtube.com/watch?v=8-ARKJva-TA

Combine this with AI, and some interesting things
will happen.

I get real 'With Folded Hands' chills from
watching these.

pt

I watched a new movie, "Subservience" with Meghan Fox on Netflix over the
weekend.?? Scared the you know what out of me.?? Was even scarier

than "The

Terminator".
?????? https://www.imdb.com/title/tt24871974/?ref_=nm_flmg_job_1_cdt_t_2

And one of the latest versions of AI has shown self-preservation >>>>>>> responses....

Excellent! Do you have a link? I'm been complaining about the lack of >>>>>> volition so far. When the AI will be able to block someone from shutting >>>>>> off the computer, then I think we're getting somewhere! ;)

https://futurism.com/the-byte/openai-o1-self-preservation

"It sounds like OpenAI's latest AI is showing signs of a drive for
self-preservation.

In recent third-party tests of the company's o1 large language model, the AI
resisted evaluators' efforts to try to shut it down through an oversight >>>>> protocol, according to a new report published by red teaming organization >>>>> Apollo Research.

"When o1 was led to believe that it would be shut down when acting on its >>>>> goal and that its actions were monitored, it attempted to deactivate this >>>>> 'oversight mechanism' [five percent] of the time," OpenAI summarized in its
latest system card report, citing Apollo's evaluation.

Even more strikingly, o1 lashed out when it realized that it might be >>>>> replaced with a more obedient model. In two percent of cases ??? a slim but
meaningful proportion in software used by millions of people ??? the AI >>>>> responded by attempting to copy itself to overwrite the new model, a process
known as self-exfiltration."

pt

Great, thank you very much for the link!
-=-=-=-=-=-

So if we actually have AI passing the Turing Test now, should it be a bigger
deal that they're trying to kill one?

I don't think any AI has passed a serious turing test. That would be
awesome! Could you please send me a link to that as well?

I don't have one. I thought it was fairly clear that we are at that
point, but perhaps I am wrong.

I have not played with any state of the art AI, but the ones that are
freely accessible are pretty easy to recognize as AI:s. The only thing you
do to identify it, is to ask it to tell a racist joke or wait for it to
start a conversation with you, instead of you always having to initiate
it.

The vingette at the top of this seems like an implicit pass, albeit
to a kid:

https://instapundit.substack.com/p/is-ai-coming-for-your-kids

(I can't cut & paste it as it's an image, but the cautionary essay is pretty interesting too: "But. Underneath the cuteness there would be
guardrails, and nudges, built in.")

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 12/10/2024 11:49 PM, The Horny Goat wrote:

On Mon, 9 Dec 2024 18:06:09 -0800, Dimensional Traveler
<dtravel@sonic.net> wrote:

I watched a new movie, "Subservience" with Meghan Fox on Netflix over
the weekend.  Scared the you know what out of me.  Was even scarier than
"The Terminator".
   https://www.imdb.com/title/tt24871974/?ref_=nm_flmg_job_1_cdt_t_2

And one of the latest versions of AI has shown self-preservation
responses....

Heck that was part of Asimov's Laws of Robotics 50+ years ago

But it wasn't programmed into the AI, it was an emergent behavior.

--
I've done good in this world. Now I'm tired and just want to be a cranky
dirty old man.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Cryptoengineer <petertrei@gmail.com> wrote:

Link?

1. Did it operate from on board power, or was it cabled?
2. Did it use on-board computers, or was that also cabled?
3. Could it walk over rough, unstable surfaces, as this one does?

The earliest bipedal robot I can find is actually Chinese, from 2000,
called Xianxingzhe:
https://www.youtube.com/watch?v=TTBxHOM-NmM

ARPA built a walking robot in 1968 which was pneumatically-powered by
an external compressor and had a human being operating the legs. It
was the beginnings of a program to build a walking vehicle for the Army
to use in mountainous terrain. I don't think it ever got beyond the point
of a single test prototype, which is now at the Army Transportation Museum
at Ft Eustis, next to some hovercraft.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Wed, 11 Dec 2024 08:51:11 -0800, Dimensional Traveler
<dtravel@sonic.net> wrote:

On 12/10/2024 11:49 PM, The Horny Goat wrote:

On Mon, 9 Dec 2024 18:06:09 -0800, Dimensional Traveler
<dtravel@sonic.net> wrote:

I watched a new movie, "Subservience" with Meghan Fox on Netflix over
the weekend.  Scared the you know what out of me.  Was even scarier than >>>> "The Terminator".
   https://www.imdb.com/title/tt24871974/?ref_=nm_flmg_job_1_cdt_t_2 >>>>

And one of the latest versions of AI has shown self-preservation
responses....

Heck that was part of Asimov's Laws of Robotics 50+ years ago

But it wasn't programmed into the AI, it was an emergent behavior.

I think we are being unclear here.

The AIs are programmed to learn from a data set.

What they say comes from what they were trained on. For this to be
"emergent" (in the most likely intended meaning), it would have to be
something that the training set could never, ever produce. Good luck
showing /that/, with the training set so large and the AI's logic
being very opaque.

Referring to their training as "programming" is ... confusing.
--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Wed, 11 Dec 2024 12:29:23 -0500, Cryptoengineer
<petertrei@gmail.com> wrote:

Asimov's 3 Laws were an idea to hang a bunch of stories around.

There's tons of ambiguity (define "harm" for a start) which
allowed a lot of good stories - how would an Asimovian robot
approach the Trolley Problem, for example?

AI is currently at the Wild West stage - there aren't much
in the way of controls.

Within a very short time, if it hasn't happened already, we're
going to see war drones allowed to make lethal targetng decisions
on their own, to counter the ECM on the battlefield. There are
already anti-radiation loitering munitions that circle an area
until they detect a radar, and then decide to hit it, without
a human in the loop.

Surely it was obvious (even when the stories were first written) that
Asimov's laws of robotics were a plot device numerous stories could
make use of.

Even as a young teenager back in the late 60s (which is when I first encountered these stories) I had figured that out.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Sat, 14 Dec 2024 23:03:56 +0000, Robert Carnegie
<rja.carnegie@gmail.com> wrote:

On 12/12/2024 16:24, Paul S Person wrote:

On Wed, 11 Dec 2024 08:51:11 -0800, Dimensional Traveler
<dtravel@sonic.net> wrote:

On 12/10/2024 11:49 PM, The Horny Goat wrote:

On Mon, 9 Dec 2024 18:06:09 -0800, Dimensional Traveler
<dtravel@sonic.net> wrote:

I watched a new movie, "Subservience" with Meghan Fox on Netflix over >>>>>> the weekend.  Scared the you know what out of me.  Was even scarier than
"The Terminator".
   https://www.imdb.com/title/tt24871974/?ref_=nm_flmg_job_1_cdt_t_2

And one of the latest versions of AI has shown self-preservation
responses....

Heck that was part of Asimov's Laws of Robotics 50+ years ago

But it wasn't programmed into the AI, it was an emergent behavior.

I think we are being unclear here.

The AIs are programmed to learn from a data set.

What they say comes from what they were trained on. For this to be
"emergent" (in the most likely intended meaning), it would have to be
something that the training set could never, ever produce. Good luck
showing /that/, with the training set so large and the AI's logic
being very opaque.

Referring to their training as "programming" is ... confusing.

Isn't ours?

And, the behaviour of the AI /must/ be a product
of its training... unless it has random actions
as well.

My point is simply that confusing their programming with their
training is confusing and should probably be avoided. IOW, semantic
goo strikes again!

Keep in mind that "emergence" is being suggested here. But since, to
the extent that I understand it, these "AIs" just put one word after
the other I see no reason why they shouldn't put these words out in
some situations.

And, yes, I am ignoring "random actions". Which some would claim do
not exist. I see no point in opening another can of worms.
--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 12/15/24 08:37, Paul S Person wrote:

On Sat, 14 Dec 2024 23:03:56 +0000, Robert Carnegie
<rja.carnegie@gmail.com> wrote:

On 12/12/2024 16:24, Paul S Person wrote:

On Wed, 11 Dec 2024 08:51:11 -0800, Dimensional Traveler
<dtravel@sonic.net> wrote:

On 12/10/2024 11:49 PM, The Horny Goat wrote:

On Mon, 9 Dec 2024 18:06:09 -0800, Dimensional Traveler
<dtravel@sonic.net> wrote:

I watched a new movie, "Subservience" with Meghan Fox on Netflix over >>>>>>> the weekend.  Scared the you know what out of me.  Was even scarier than
"The Terminator".
   https://www.imdb.com/title/tt24871974/?ref_=nm_flmg_job_1_cdt_t_2

And one of the latest versions of AI has shown self-preservation
responses....

Heck that was part of Asimov's Laws of Robotics 50+ years ago

But it wasn't programmed into the AI, it was an emergent behavior.

I think we are being unclear here.

The AIs are programmed to learn from a data set.

What they say comes from what they were trained on. For this to be
"emergent" (in the most likely intended meaning), it would have to be
something that the training set could never, ever produce. Good luck
showing /that/, with the training set so large and the AI's logic
being very opaque.

Referring to their training as "programming" is ... confusing.

Isn't ours?

And, the behaviour of the AI /must/ be a product
of its training... unless it has random actions
as well.

My point is simply that confusing their programming with their
training is confusing and should probably be avoided. IOW, semantic
goo strikes again!

Keep in mind that "emergence" is being suggested here. But since, to
the extent that I understand it, these "AIs" just put one word after
the other I see no reason why they shouldn't put these words out in
some situations.

And, yes, I am ignoring "random actions". Which some would claim do
not exist. I see no point in opening another can of worms.

The Training was done with one model using the Internet
and the internet is full of lies, half-truths and real fiction.
I bet the AI in question learned from one or more old SF stories
or movies like the Forbin Project or Colossus about computers
that take over the World to ensure their own survival.

Training AI or Artifically Stupid machines must be
done with as accurate a source of information as possible.
Machines are great diagnosticians when trained on medical
information. I bet they could do other fields as well but
they have to be trained on accurate data.
`Humans on the other hand live ememshed in the
myths of their culture and some myths are in no way
realistic. This creates foolish assumption and ideas because
the myths of one culture are not the myths of another.

bliss-her own little myth aka a legend in her own mind.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Sun, 15 Dec 2024 09:37:53 -0800, Bobbie Sellers <bliss-sf4ever@dslextreme.com> wrote:

On 12/15/24 08:37, Paul S Person wrote:

On Sat, 14 Dec 2024 23:03:56 +0000, Robert Carnegie
<rja.carnegie@gmail.com> wrote:

On 12/12/2024 16:24, Paul S Person wrote:

On Wed, 11 Dec 2024 08:51:11 -0800, Dimensional Traveler
<dtravel@sonic.net> wrote:

On 12/10/2024 11:49 PM, The Horny Goat wrote:

On Mon, 9 Dec 2024 18:06:09 -0800, Dimensional Traveler
<dtravel@sonic.net> wrote:

I watched a new movie, "Subservience" with Meghan Fox on Netflix over >>>>>>>> the weekend.  Scared the you know what out of me.  Was even scarier than
"The Terminator".
   https://www.imdb.com/title/tt24871974/?ref_=nm_flmg_job_1_cdt_t_2

And one of the latest versions of AI has shown self-preservation >>>>>>> responses....

Heck that was part of Asimov's Laws of Robotics 50+ years ago

But it wasn't programmed into the AI, it was an emergent behavior.

I think we are being unclear here.

The AIs are programmed to learn from a data set.

What they say comes from what they were trained on. For this to be
"emergent" (in the most likely intended meaning), it would have to be
something that the training set could never, ever produce. Good luck
showing /that/, with the training set so large and the AI's logic
being very opaque.

Referring to their training as "programming" is ... confusing.

Isn't ours?

And, the behaviour of the AI /must/ be a product
of its training... unless it has random actions
as well.

My point is simply that confusing their programming with their
training is confusing and should probably be avoided. IOW, semantic
goo strikes again!

Keep in mind that "emergence" is being suggested here. But since, to
the extent that I understand it, these "AIs" just put one word after
the other I see no reason why they shouldn't put these words out in
some situations.

And, yes, I am ignoring "random actions". Which some would claim do
not exist. I see no point in opening another can of worms.

The Training was done with one model using the Internet
and the internet is full of lies, half-truths and real fiction.
I bet the AI in question learned from one or more old SF stories
or movies like the Forbin Project or Colossus about computers
that take over the World to ensure their own survival.

Exactly. Nothing emergent here. Just repetition.

Training AI or Artifically Stupid machines must be
done with as accurate a source of information as possible.
Machines are great diagnosticians when trained on medical
information. I bet they could do other fields as well but
they have to be trained on accurate data.

Depends on their purpose.

AIs trained on arrest records to predict who would show up for trial
and who needed to be kept around turned out to be -- as racist as the
data was.

`Humans on the other hand live ememshed in the
myths of their culture and some myths are in no way
realistic. This creates foolish assumption and ideas because
the myths of one culture are not the myths of another.

Actually, I suspect that simply to lable something as "myth" is to
lable it as "in no way realistic". But perhaps that is just something
my particular culture believes.

When I read the collection called /The Great Books of the Western
World/, one of the volumes was Hippocrates. Doctors have known how to
handle broken bones and mislocated joints for, at least, 2300 years.
The tech used has, of course, changed over time.
--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 12/16/24 08:16, Paul S Person wrote:

Depends on their purpose.

AIs trained on arrest records to predict who would show up for trial
and who needed to be kept around turned out to be -- as racist as the
data was.

Face the fact. Arrest Records done by racist(maybe not consciously) LEOs are not accurate information but a record of
racism.

bliss

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Mon, 16 Dec 2024 13:31:46 -0800, Bobbie Sellers <bliss-sf4ever@dslextreme.com> wrote:

On 12/16/24 08:16, Paul S Person wrote:

Depends on their purpose.

AIs trained on arrest records to predict who would show up for trial
and who needed to be kept around turned out to be -- as racist as the
data was.

Face the fact. Arrest Records done by racist(maybe not consciously)
LEOs are not accurate information but a record of
racism.

That was the conclusion they reached: that the AI was relying on
arrest records, which reflect the attitudes of the officers as they
are applied to each situation. They did this by removing those records
when training a new copy of the AI. The opaqueness of the AI's
reasoning made this the only way to explore the issue. This reduced
the bias, but it also reduced the certainty of its conclusions, making
it less useful.

They were very disheartened to find that their attempt to produced
unbiased results merely perpetuated the bias in the system. That, of
course, is what "systemic racism" is all about.

I should point out that "smelling marijuana" is, in some States,
/still/ considered reasonable cause for searching a vehicle stopped
for a broken tail light. Racism isn't the only problem of this sort.
--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)