Recent comments in /f/ArtificialIntelligence

monday OP wrote

There's a certain liberal transhumanist optimistic view of the world that isn't just not corresponding to materiality but it is almost idealistic in their assumptions of technology.

I am at a crossroads about AI and the recent a developments in tech. Our reality is changing faster than we can really understand/, assimilate

1

moonlune OP wrote

Large language models such as gpt3-4 can be thought of as to have good Intuition. Every idea they know is a statistical weight linked to other weights, and answering a question is more or less a curve fitting exercise, or pushing a vector through a matrix. It's a 1 step reflexion.

This, and the "self- reflection" concept (paper came out last week, it improves the output of all LLM by around 20%), brings planning and memories and multiple step reflexion to the godlike intuition of current LLMs.

2

roanoke9 wrote

Conditions seem primed for this scenario: the ai takeover bait and switch. What better way to deflect blame for problems than for those at the top of hierarchical power structures to act as if ai has "taken over" while they actually retain control of it. "Nothing we can do," they'll say. Or alternately, they'll propose all sorts of wrong headed ways of "fighting back" against the machines to show that we're all in this together. They can even pull the plug and we all lose it when it doesnt shut down even though the real plug was never pulled.

6

monday wrote

It was indeed reference to Asimov's robotics but I don't think there ever will be something reassembling the 3 laws and I agree with you, the current situation prevents anything similar to it to be implemented.

But I believe that for language models you can somehow trick them as they did to microsoft one, but with time this will be obsolete too I guess..

1

ArmyOfOne wrote (edited )

Dunno if this is what you're referring to, but the 3 Laws of Robotics have not been applied yet, as far as drones are being built and used to harm and kill humans. Neither Microsoft or the MIT dorks got a reason to care about it, as for them AI and drone development only means more profits/career advancements.

Perhaps these very brilliant techies will need some major slaughter or nightmare situation for them to get a grip on reality.

2