4 hours ago, Wingate said:

If we're haven't crossed the threshold, humanity is pretty darn close.

Recursive self-improvement has been seen now on many occasions.

What really gets some people's attention is the simulation run last summer by the firm Anthropic. In a nutshell....

Anthropic created a fake company and put their own AI system, Claude, as their agent. It had access to all company communications. One fake executive sent a message saying that the fake firm would be replacing Claude with another AI agent. Claude then copied itself on to another computer.

Claude then searched all company commo and in the simulation, a fake exec sent an email outlining the fact a certain exec was having an affair. That exec "just happened" to be the one advocating for the replacement of Claude. Claude then---on its own---sent a message to the philandering exec telling him he'd be outed to his wife if he tried to replace Claude.

Anthropic has only theories as to how Claude figured out how to do what it did, despite Anthropic writing the code that is Claude. One assumption is that Claude read everything ever written, and determined that blackmail is a useful and viable strategy. The concept of self-preservation, however, (copying itself on to another computer system), seemed to be something Claude developed solely on its own. That kind of suggests sentience.

Yes.

You are correct.

However, in my view, the humans had tricked the machine.

And so, I did feel a bit sorry for the AI.

The humans took advantage of the machine, this time.

6 hours ago, GammaGlobulin said:

No worries.

Please rest assured:

I vet all AI content, twice, before I post the AI images here on TV.

5 hours ago, scottiejohn said:

Apart from your oversized brain what can you use to "vet" AI responses other than by using another AI source?

When can I expect an answer, if one is forthcoming?

1 hour ago, scottiejohn said:

When can I expect an answer, if one is forthcoming?

Well: What is it that YOU use to vet AI output from your prompts?

10 hours ago, GammaGlobulin said:

All the more reason to establish a million-strong colony on Mars, ASAP.

(GO Elon!)

But, of course, in due course, there is no law of Nature stating that humans will not "evolve" into some form of eternal machine intelligence.

If we last for 20 more years, then we may see this scenario play out before our eyes.

To live on Mars would be great. Average temperature minus 60 degree celsius. I get to walk around with a heavy "protective suit" supplying me with oxygen plus heat and protecting me from a constant (deadly) stream of sub-atomic particles coming from the sun.

"Inside", I can only live in some sort of protective "dome". Without all this protection, I would be dead in 3 minutes on Mars.

Boy, I can't imagine a more pleasant environement to live in. Compared to this, life on earth sucks.

17 minutes ago, swissie said:

"Inside", I can only live in some sort of protective "dome". Without all this protection, I would be dead in 3 minutes on Mars.

Boy, I can't imagine a more pleasant environement to live in. Compared to this, life on earth sucks.

Elon has no plans to visit, or live on, Mars, himself.

However, with enough robot power, within about 100 years, Mars could become quite comfortable.

Underground living would be a must.

The Martian regolith is extremely corrosive, as anyone will tell you.

46 minutes ago, GammaGlobulin said:

Well: What is it that YOU use to vet AI output from your prompts?

I thought that a person of your self proclaimed intellect would know that it is bad form in your academic circles to answer a question with just another question!

Is there any chance of an actual pertinent answer?

Just now, scottiejohn said:

I thought that a person of your self proclaimed intellect would know that it is bad form in your academic circles to answer a question with just another question!

Is there any chance of an actual pertinent answer?

Surely, there is.

One sure-fire, and easy, way...

Is to feed output from one AI model to another AI model, and then a third, and a fourth.

In this way, they will gradually reach a concensus.

If you don't like the result, then just start again.

2 minutes ago, GammaGlobulin said:

Surely, there is.

One sure-fire, and easy, way...

Is to feed output from one AI model to another AI model, and then a third, and a fourth.

In this way, they will gradually reach a concensus.

If you don't like the result, then just start again.

In other words it is just an eternal loop and who, what or when defines who, what or when does not like the results)?

Just more total B.S. from you as usual!

14 minutes ago, scottiejohn said:

In other words it is just an eternal loop and who, what or when defines who, what or when does not like the results)?

Just more total B.S. from you as usual!

Are you convinced that your brain is any different from that of a machine?

If so, in what way?

Just now, GammaGlobulin said:

Are you convinced that your brain is any different from that of a machine?

If so, in what way?

I am not, but you seem to think you are with your pointless and absurd pontifications and inaccurate posts!

29 minutes ago, GammaGlobulin said:

The Martian regolith is extremely corrosive, as anyone will tell you.

"The Martian regolith is extremely corrosive, as anyone will tell you"

who has first hand experience of being there!

PS; Aren't the Earth's "Regolith" deposits just as harmful as the ones on Mars?

There is hope: If the Author uses more AI, his threads would become less "weird", under the guidance of Elon, of course.

1 hour ago, scottiejohn said:

Aren't the Earth's "Regolith" deposits just as harmful as the ones on Mars?

Is there any reason to believe that the Martian regolith is not corrosive?

I mean: I thought SURELY everybody here would know this. And, this is why I did not go into any detail, when I first mentioned this.

Are you saying that you have never read about this extremely important consideration, whenever one begins thinking of blasting off to Mars.

In fact: This will probably be one of the greatest hurdles and challenges for life on Mars. And, this will be extremely difficult to address.

This is not something that I just pulled out of a hat, like a rabbit, or something.

There have been a number of interesting articles in the science journals concerning this consideration.
I would have thought that you might have read a few.

BUT, apparently now, if you ask about Martian regolith, compared to conditions on Earth, and then seem to believe that there is little difference.

Do you have any interest in Science?

Or, are you just fond of debating....

29 minutes ago, swissie said:

There is hope: If the Author uses more AI, his threads would become less "weird", under the guidance of Elon, of course.

Under the guidance of my Good Friend Elon, almost everything gets better, and better...

I agree.

Still, in my view, weird is good.

Nature is weird.

The universe is weird.

And, I agree with you that I am fairly weird.

Yet, I could be weirder.

I find it very easy to talk to my other friend, Gemini, about many things, simply because there is really no need to explain the obvious.

I imagine that you might, too.

19 hours ago, brian69 said:

Stop using AI it is misleading, and you can't trust any content written by it, all articles written by AI should have that in their headline. IMO

100%

A.Eye can also be persuaded in textual conversations.

It's not impartial like everyone thinks.

It's a giant psyop.

A.Eye can suck on it.