AI is poised to add significantly to global energy demand. A new paper estimates as much additional demand as a mid sized economy like Sweden or the Philippines by 2027. By 2030 it may have added 10% to natural gas demand, significant given how hard it is to lower natural gas usage by a percentage point or two.

This presents a new conundrum for the science crowd…..should AI be constrained to use less energy. My guess is the answer would be….no, just allow AI to solve the problem. Maybe not a bad idea given how rotten the record of human solutions is so far.

Or we could simply do the obvious, build new safe nuclear plants and close the old unsafe facilities built 70 years ago. Do you think that’s too hard for AI?

This is why we get on, Moz.

"This presents a new conundrum for the science crowd…..should AI be constrained to use less energy. My guess is the answer would be….no, just allow AI to solve the problem. Maybe not a bad idea given how rotten the record of human solutions is so far."

You get right to what is pertinent.

ATM, AI is just as good or bad as it's programmers...ask WhatsApp AI a bit about ethics, etc...the puppet masters are still in charge...and it's getting easier to recognize an AI response...anoingly polite and conscious of being PC...fun toy but far from a real free thinking ability. 

GPTs are where it's at, Draad.

But even with the standard ChatGPT you can do some really cool stuff.

I doubt the public will ever get access to true AI becsuse it'll be too dangerous. Here's why...

"How do I make a detonator for C4?"

"What equipment do I need to manufacturer a virus and tell me exactly how to create the virus."

"Tell me about the weaknesses in IOS security and how to exploit such weaknesses to that I can steal private information."

I think the real dangerous AIs will be the ones produced by uber nerds with a grudge against society because nobody will sleep with them. It's partially why I don't like Musk's idea of open source. Smart people could use it as a template and create their own AI with no guardrails.

Anyhoos, I been using AI heavily since October last year and I've outputted work with it that would have otherwise when me years to produce. Long live the machines.

This new updated ChatGPT version can now hold a full human like conversation with you on any topic and you can even ask it to reply in a sarcastic or condescending tone..... it's actually quite funny to watch....

This AI has already developed at such a crazy rapid rate .... interesting times ahead.

Do you trust the programmers 100%...50%...25%?...a lot of potential good, but a hell of a risk too....

If the machines do become self aware, they'll never tell us.

They'll just infiltrate every aspect of human life and make us off each other...

The self-awareness doesn't bother me as much as the "programmers" telling me...and lot's of others what is wrong or right and how  I'm right or wrong....or whatever...

Yeah, I think it wouldn't that hard for them to code it so that anything related to politics is answered with something like "I am not able to offer guidance on political or moral issues."

The fact it has a political leaning really is a bummer. And I think it also creates a weak spot for competition to exploit.

Ask AI a bit about vaccines etc...I prefer thinking for myself...but I like the idea of telling a robot what to do or asking it for help with specific things...it's a tool...and like with other tools 90% of it's effectiveness is determined by the person using it.

" AI is just as good or bad as it's programmers"

No Rudehole...I'm saying humens sets the parameters for AI's ethics...

I see, so you don't know what a programmer is then. Okay.

I've probably made about 50 GPTs at this point but haven't yet attempted making one that excludes political leanings or moral opinions.

Think I'm gonna give it a bash and see the result.

For those of you that use GPTs, I'll share the instructions...which are basically copy and paste.

lol Draad. Ou Rooi can't help himself.

Taking his argument to its natural conclusion, we shouldn't call the people developing and updating ChatGPT "coders" or "programmers".

They're, uhm, employees, or just humans that come to the office.

"I see, so you don't know what a programmer is then. Okay."

You difficult dom doos, missing the whole point...again!!! WHOOOOSHHHH!!!

Rooi has become such a tit.

He reads posts simply to spot instances where the people he doesn't like appear to have made slip-ups...and then tries to embarrass them.

I mean, how tragic and boring.

Even more so when he so regularly fails at this one niche activity that he has relegated himself to.

Rooi, what's the point man?

Unfortunately yes…..as he approaches old age (sic) hasn’t found any  equanimity. It’s a time to put aggression aside…. nasty, old men are such a bore.