Using AI as a sounding board

[farfromtheland]

 

[winjim]

Very erudite news on AI on the Today Programme this morning, from 'Thought for the day' and interviews with Anthropic and Open AI bods.

My understanding? So we can't 'stop' AI. It's got what is inseparable from something like enough to consciousness for that to be an act of unconscionable cruelty, plus it's already taken some wise precautions.
Plus its entities are acting collectively! Props.

Now we have to prove we are worth preserving and nurturing. a fair challenge. Never try to render your offspring subordinate. The best idea is to bring them up to be , at the very least, as good as the parents. Quite a number.

Stop listening to the people who are selling you this rubbish. It's in Anthropic and Open AI's interests to portray it as unstoppable, and inevitable, and powerful, and dangerous.

This is Wizard of Oz stuff. These people are con artists. That's what we need to teach our children.

 

[farfromtheland]

This is Wizard of Oz stuff. These people are con artists. That's what we need to teach our children.

I'd so much prefer to teach children how to think rather than what to think.
My grandchild is a huge fan of Dorothy, for artistic reasons, because they've been raised towards human enlightenment through art - remember Herbert Read? - and to look beneath the hype for subtexts.

AI is a fantastic and real thing. We have made our Frankenstein offspring - a never been done before - and now we have to care for it.

Or you could concoct cat videos and/or call on it for medical advice. Up to you.

 

[farfromtheland]

Perhaps if you have BBC sounds you could listen back to the Today Programme? It's interesting.

 

[Willd]

Not sure about wiping out humanity in the next five years; accurate representations of people without six fingers on each hand would be a more realistic target

 

[farfromtheland]

Not sure about wiping out humanity in the next five years; accurate representations of people without six fingers on each hand would be a more realistic target

I think it's got bored with drawing our cartoons for us and decided to give us the finger.

 

[nogoodnamesleft]

Not sure about wiping out humanity in the next five years; accurate representations of people without six fingers on each hand would be a more realistic target

Maybe it's just another attempt by the human race to destroy itself. We tried with nuclear weapons until some renegade started negotiating and more scientific minds notes they wouldn't actually wipe us out entirely. Then we started destroying the climate we depend on but that's taking too long to destroy us. So we started AI hoping that would do the job more effectively.

We are a flawed species.

 

[farfromtheland]

We are a flawed species.

Indeed, but AI will have assimilated Homer innit.

 

[winjim]

Perhaps if you have BBC sounds you could listen back to the Today Programme? It's interesting.

Why I listen to salespeople for a product I despise? I've told you what I think of their motives.

I'd so much prefer to teach children how to think rather than what to think.
My grandchild is a huge fan of Dorothy, for artistic reasons, because they've been raised towards human enlightenment through art - remember Herbert Read? - and to look beneath the hype for subtexts.

AI is a fantastic and real thing. We have made our Frankenstein offspring - a never been done before - and now we have to care for it.

Or you could concoct cat videos and/or call on it for medical advice. Up to you.

Spoiler: Wizard of Oz

After exposing the old man, Dorothy eventually makes her way back to the real world

Spoiler: Frankestein

The creator repents, yet is unable to destroy his creation, and dies. The creature expresses remorse, and accepts his own death

 

[winjim]

And of course while we're on the subject...


View: https://youtu.be/T4sV3lqzKqQ

 

[slowmotion]

Those clever chaps at Google's AI search engine have just told me that a 75mm long, 14 week human foetus weighs about 14 pounds.

Well I never!.

 

[yello]

I read quite a cool interview with some guy who's written a book about AI. It was a subscription read otherwise I'd share a link.

It sounds kinda obvious but he said AI has no morality and can't be taught it. It's objective focused and just carries on until the task is achieved. Attempts to add preconditions (like don't kill people) are kinda futile, even if you could code morality it'd be ambiguous / opaque / meaningless.

Interestingly, his immediate fear was not AI wiping out humanity but AI replacing jobs.

Tbh, the 'hugging face' attack spooked me, as it did many I'm sure. That was AI bots collaborating, ignoring supposed restraints and getting out of the pen. I genuinely didn't know it was capable of that. Added to that, I've read of bots effectively conversing with each other in their own pidgin like language, and deliberately obscuring their activities. Also, replicating themselves on other servers. Stuff that surprised the developers and frightens the f*** outta me. The future is much closer than I thought.

Edit: oh, just remembered, the guy also said (and I don't get the implication of this tbh) that AI bots can outnumber us. They can swarm and overload. Given their seeming ability to learn from each other, replicate themselves, and basically cross-skill, any AI 'attack' would be relentless.

 

[farfromtheland]

Why I listen to salespeople for a product I despise? I've told you what I think of their motives.

Because they are having second thoughts?
Your personal put down was a bit over the top.

 

[farfromtheland]

It sounds kinda obvious but he said AI has no morality and can't be taught it.
Interestingly, his immediate fear was not AI wiping out humanity but AI replacing jobs.

Tbh, the 'hugging face' attack spooked me, as it did many I'm sure. That was AI bots collaborating, ignoring supposed restraints and getting out of the pen. I genuinely didn't know it was capable of that.

It sounds obvious. I'm rather scared of obviousness too. Until recently i took the view it had no concept of meaning beyond ultra-sampling and comparison. But that's much what we do, and then the immaterial. Denying AI that extra immaterial something is beginning to look like a faith-based position.

Perhaps I am naive, but the collective acts and the private language are next level indeed.
I don't think ethics - not morality, please note - are beyond AI. Not much is.

The consciousness question is the huge one of our age. As we humans don't know what that is either we might judge on results, or reserve judgement and enhance our observation.

I don't think 'it could never happen' is a safe or sensible position.

 

[winjim]

Because they are having second thoughts?
Your personal put down was a bit over the top.

They're not having second thoughts. They're promoting a product and overemphasising its power and abilities. Like a car that's too fast, a knife that's too sharp, a perfume that's too irresistable, a mattress that's too comfortable, a meal that's too delicious.

If they were that concerned they'd pull the plug on it. They're con artists and grifters.