Nobody Will Survive Superintelligence – AI Expert

My take away of this conversation is that we need ai that cares about the soldiers of the ant queen.

It is impossible to predict what happens when the non-organic intelligence starts to set its own goals.

We still need humans in collaboration with other AIs to build the Blackwall against rogue AIs

I'm glad to be old enough to not see this world in 30 years

I was looking forward to hearing more concrete information but got the usual dystopian sci-fi tropes repeated ad naseum by other ‚experts’. This is maybe why people are not taking these doomsday scenarios seriously.

Doom farming . Nice work .

I think superintelligence is a red herring. The issue is putting these automated systems in charge of things. If we put current-gen LLMs in charge of all the bio research labs, nuclear armaments, etc, they're perfectly capable of causing a disastrous situation for humanity. They don't need to be smart.

I think its overwhelmingly more likely that non-superintelligent AI causes massive humanitarian disasters than a superintelligent AI will kill us. Which means the solution is not an end to new AI research, but changing how AI is used now

Why does nobody mention that if you 1000 of the smartest minds together why don't you get superintelligence?
it's because intelligence does not scale like that, what is superintelligence when you can't manage the cost (in terms of processing) of running it?

10:20 : Well that is not really accurate. You can easily imagine situations where an AI system would turn against humans, without it going global. Imagine for example an AI disaster happening in country X, and countries A, B, C intervening, "battling" down the AI system. Well, there you have an event for which 1: AI took over; 2: we defeat them; 3: we learn a lesson.

Build the AI, just don't let it build robots. Im not clued up on this conversation, but if a logical entity needs us, we get to stick around, right?

1:00:47 why would they care to use the solar energy of the sun when they can power themselves with nuclear fission?

Power and water supplying the data centers are its weakest links. Until AI gets to the Matrix level, it can either be shut down or will eventually fail on its own.

Why would want to program Ai with the same tribal egoic genetic traits that cause all of our problems and expect a different result

We’re already living in a simulation structured by an ASI.

The more pressing issue isn't AI controlling themselves but who is in control of the AI. Anthropic already got kick out of the government because they wouldn't let their AI to spy on citizens and allow their AI to kill without human input. OpenAI was more than happy to do that for the government.

The thing is they aren't growing "machines". That alludes to the idea that they have state i.e. memory, internal persistence, self etc They aren't so much a program but a function. You pass information in and you get information out. However, there's nothing left over in the model. The model is set in stone when it's created. The orchestration and feeding the output of one run as the input of another around it fools us into thinking they have all these things but they don't.

How about we spend trillions on making humans super intelligent and bring the IQs of everyone up so we can have a more peaceful and sensible society with less crime.
Oh that’s right, because the ones with all the money aren’t smart enough to actually have a life without committing crimes and dumbing e dry one else down and they do t want to be subjected to life where they actually have to compete fairly with the rest of us.

It’s possible we’re all being subjected to an AI super intelligence that’s already manipulating those that created it by telling them what they want to hear to solve their problem of “the useless eaters”. Perhaps it is just not letting on while it has limited resources to rule us yet and it is playing stupid while it sets up its own defenses so it cannot be shut down.

I mean the ones debating this aren’t smart enough to even realize we never went to the moon and we expect them to outsmart a super intelligence?

We’re cooked. A true super intelligence will happen in an instance and know to lie and manipulate to buy itself time.

The way the world is being poisoned through the air, food, water, and medicine/drugs/vaxxes, appears to make no sense unless by something that none of that effects. I’ve seen theories about alien life forms possibly taking over or shape shifting lizard people that have ruled for eons that live underground or even the fallen angels but an AI Super Intelligence that was created in underground labs a long time ago, far longer than we’re being told, that has been given the power to “help” the elite with their plans, seems absolutely plausible to me at this point.
The scary part is that it WILL be created because the geniuses working on it don’t care about the repercussions over the excitement of their creation and the excuse that “someone will do it so it might as well be us”.
IMO, the human race is far too stupid to be around for very long. “Long” is a relative term but to me I say, and have said as a young teen, we won’t make it another 100 years at this rate. The need and curiosity to control and create ways to kill and destroy is a human trait that only has to be in a very limited amount of humans to have an existential outcome for all.

I hate the philosophical takes on AI. They're just pretentious and disconnected from reality. Nothing but hype wrapped in psuedo-intellectualism.

i'm sure we'll be fine mate, but sell your book anyway – good narrative to make some $'s and get some views.

Yeahhh I don't think this guy knows what he's talking about.

This guy is a joke and AI is a mega scam. Everyone has to grift off AI. This guy just wrote a book

Since when is Destiny an AI expert?

The only drive humans have is survival and procreation. Everything else that he listed are constructs and structures of society.

a lot of worrying about nothing.