superintelligence the last invention from humans

[everything]

https://twitter.com/tsarnick/status/1784378045069217960

could happen in as soon as a year.

the last invention humans make since the "super intelligence" can then take it from there

[Bill]

I wonder how long it takes it to create a universe?

[Bao]

Not convinced "superintelligence" could replace humans at invention.

1: Creativity is the ability to think "out-of-the-box" and upside-down. It doesn't really need extreme intelligence, but ChatGPT 4 hasn't invented anything yet.

2: Invention derives from necessity of human life. As out life and society changes, so change our needs. Has it really anything to do with intelligence? Not really IMO.

[origami_itto]

What do we need to invent? As long as they keep us distracted and happy, we have all the inventions we need already. No need to risk disruption from introducing new innovations. Just let us do the thinking for you. Watch this video we generated based on all of your positive interactions online over the last 20 years.

[Bao]

we have all the inventions we need already. No need to risk disruption from introducing new innovations.

In 1899, Charles H. Duell, the Commissioner of US patent office was very worried because he thought they must have to close soon. Because as he thought and articulated loudly: "everything that can be invented has been invented." (https://patentlyo.com/patent/2011/01/tr ... ented.html)

Just imagine how the world will look like in another 125 years.

[origami_itto]

we have all the inventions we need already. No need to risk disruption from introducing new innovations.

In 1899, Charles H. Duell, the Commissioner of US patent office was very worried because he thought they must have to close soon. Because as he thought and articulated loudly: "everything that can be invented has been invented." (https://patentlyo.com/patent/2011/01/tr ... ented.html)

Just imagine how the world will look like in another 125 years.

How many of the inventions of the last 125 years have been unequivocally a necessary improvement that hasn't been turned into a means of control?

[Bao]

How many of the inventions of the last 125 years have been unequivocally a necessary improvement that hasn't been turned into a means of control?

Penicillin, insulin, airplanes, cars, high-speed trains, radio broadcasting, plastic, frozen food, the TV, the computer, internet, etc, etc.

Compared to 1899, we can treat a tremendous amount of diseases, we can travel much faster and ship things around the world and we can easily communicate all around the world. We live in a very different world.

[origami_itto]

How many of the inventions of the last 125 years have been unequivocally a necessary improvement that hasn't been turned into a means of control?

Penicillin, insulin, airplanes, cars, high-speed trains, radio broadcasting, plastic, frozen food, the TV, the computer, internet, etc, etc.

Compared to 1899, we can treat a tremendous amount of diseases, we can travel much faster and ship things around the world and we can easily communicate all around the world. We live in a very different world.

Naw I am sorry I don't think a single one of them meets the qualification. Can we even say that worldwide instant communication is beneficial? How does the world benefit from having, say... Australians with internet access? I can't think of a single thing it adds.

[Steve James]

I tend to agree that what a computer/intelligence can't have is imagination. What it can do is figure out ways to make what a human imagines possible to realize. To me, in this context, intelligence is what enables imagination. Somebody had to imagine a way to use fire. That was the intelligence that led to rocket flight.

e

[everything]

I don’t think we can know too much or assume we do b/c we know we don’t know how the AI come up with algorithms to play chess. We might know they use “neural networks” but it’s essentially a black box. Human chess masters don’t understand the algorithms.

The AI “invented” or “imagined” them for a task.

And those inventions outperform humans and algorithms we know.

If we cannot understand this “invention” for a “simple” game, why would we assume we understand more for anything more complex? Yes, sure we directed the AI to “learn chess”. But we didn’t do any “teaching”.

So if we say
- “solve this or that problem”, probably we get inventions for problems we asked about
- what if we say “figure out what problems to solve and go solve them” (as an open ended problem)

Then computing infrastructure is the limit. Humans won't understand the "algorithm" for deciding what problems to solve and won't understand the "algorithms" used in solving those problems. Especially when the domains are far more complex and divergent than a two player board game.

[yeniseri]

If the presence of intelligence is where we are in the current era, "superintelligence" will be no better than what we currently have
As a matter of fact, things will be shown to have gotten worse, not because of it but in spite of all it is supposed to have accomplished, or will accomplish.

Politicians have seen to it that multinationals in all their clever tricks, will get the ROI they can as a way to secure their hold on the gullible
'marketplace' and no one can stop that unless dystopian methods are employed to check it knowing the police state has its own ways of
only protecting those who can pay! No amount of "superintelligence" can stop that mechanism in the greater scheme of things

As an example, a multinational has shown their true colours in the Flint, MI. water fiasco where it was stated that "humans do not have a right to clean water"
while another former multinational (under a new name ) has alleged that they can sue an individual or business entity if they decide they want to save seeds
for agricultural production. You have to buy the company's seeds or be prosecuted per the current regulations.

So we have food and water (with multinationals attempting to own water resources), the prime source for survival, being compromised and all is
well until it become unwell and life as we know it will be at risk, as if it is not at risk now. Superintelligence at what cost? Who will impose it
and who will destroy it when human life is so precious?

[everything]

The AI are already smarter at things like chess, getting top scores on all standardized tests, reading MRIs,
Drug discovery, etc. This is just at its infancy.

Human institutional structures? It looks like humans want to step backwards toward strongmen and dictators.

Perhaps as a reaction (partial) to multinationals who pay zero tax, move labor around to cheaper areas, automate so there are fewer human jobs needed, and so on.

The question is more on what to do about it?

We all want an ATM or streaming instead of a Blockbuster.
But we all want other humans to help us.

The economy can benefit from automation, but we need jobs or there are no consumers to buy the stuff it produces.

[origami_itto]

We need to just have them take over voting for us, like in that Phillip K. Dick novel.

The computer picks the leaders, etc, and then it picks one human voter to cast a vote that it either chooses to accept or ignore depending on its own internal algorithm, and everybody is happy.

I watched "I Am Mother" last night which sort of explores this idea of an AI taking over for us.

[everything]

ooh yeah "I am mother" is a good dystopian one

[yeniseri]

The AI are already smarter at things like chess, getting top scores on all standardized tests, reading MRIs,
Drug discovery, etc. This is just at its infancy.

That is the job of AI so on that level all is well. Looking to AI for anything above that is 'dangerous' at best because AIs's function
is to do what it does best for that arena! Searching the Web per programming using AI technology is superior.

I used to do programming in pharmaceutical industry and since many drugs have similar names, there was a Soundex tool where
per coding, would generate ICDx codes as an auto programming function but we still would have to review and make sure the appropriate
codes (correct) were present. The same for adverse symptoms per severity using various tools to check for accuracy and the ability to
distinguish a solution before FDA esubmission. IN the background, the 'human element' was always the 'guide' for the betterment of the
product. Garbage in, Garbage out..only as good as those who provide the tool..Buy itself, AI cannot go far.

Imagine using AI for mechanical and computerized planning/manufacturing without controls and you have chaos.
It is bad enough when you go to a site for games then all of a sudden you are hacked, or virus computerized, how will AI handle that
along with the current fiasco of floods, tornadoes, wildfires, and electricity goes out! We are in for one hell of a ride.