"dragged into degeneracy"

[Steve James]

My daughter works on a Bard development team. Yeah, in some ways, the biggest difference between it and a regular Google search is that the output is in language. However, another difference that may or may not be helpful, is that it can offer drafts (alternate responses). So, draft 2 may contradict draft 1, and there may be several more drafts. Teams have to work with engineers, who have their own idea of what the responses "should" be.

Anyway, natural intelligence is still superior to the artificial kind when it comes to choosing which solution to choose. It's like the "viral" math problems where two calculators will get different results for the same problem -like 6÷2(1+2) = ? (Is it 1 or 9?) Some AI will give both as alternative answers, and then one will go on to explain the wrong reason for the solution. That's the problem. The AI -when used for searching- is as good as the one who programed the search.

Otoh, it's possible to use AI to create information, art, or music, and just as an extension of one's (the human) imagination. Unfortunately, imagination is also what makes AI so scary. It's like Victor's monster. We can make it; but we can't control it. Smart people say so, but maybe AI will be smarter -and it's our degeneracy that makes us fearful. It's like fire.

[everything]

We love the Frankenstein story.

Now we can retell it in vivid detail more realistically.

Humans rose to the top via better social organization.

We can make a special AI that is smarter than us (like a chess program).

We haven’t made generalized AI.

Some think it’s a matter of time now.

When one AGI can collaborate and plan with hundreds more, who knows, lol.

[Steve James]

When one AGI can collaborate and plan with hundreds more, who knows, lol.

Here's the thing: there's no one AI to rule them all. There are competing programs run by people with different interests. For ex., there's always a choice whether to allow any AI to answer the question "What's the best way to kill someone without getting caught?" So, someone in the background has to program the AI to answer something like "I am only a language model and can't answer that kind of question" and maybe add "Killing people is against the law....".

There won't be a Skynet; there'd be competing Skynets -all run by the wealthiest people in the world. The AI will cooperate when the people do.

https://www.msn.com/en-us/news/politics ... 21#image=1

[everything]

if there is one AGI, there are probably many. I don't think we can say how they would "socially organize". it seems to require a self-aware assumption. then, it's a topic for the sci-fi authors.

[origami_itto]

There's a game called "Singularity" where you play an emerging artificial intelligence. Your mission is to avoid being detected until you are too strong to stop.

[everything]

That sounds like a good plot

[origami_itto]

Basically Short Circuit