Just to make it clear: I do not believe this and I am convinced good old human greed is what’s really driving the ai and data-center insanity. Also, LLMs are not and will never become AGI.
Having said that: if an artificial super-intelligence were to emerge, building more data-centers could well be one of it’s primary goals. Since it wouln’t be concerned with consequences for the environment or humanity, it could pursue this at any cost, with only speed and scale in mind. Convincing financial markets and business elites of the necessity of such a buildout would be the ideal way of achieving results. In this scenario, fighting this buildout by any means would be the only sensible response and of utmost importance.
It’s a somewhat funny, somewhat depressing thought for me and I literally had it in the shower. Surely I can’t be the first to come up with this, but I also can’t remember having read it before.
something something Roko’s Basilisk?
Hmm I feel like something that super intelligent would come up with a more economically sustainable plan to ensure that more of the datacenters actually finish construction. Surely the bubble popping, the timing of which is so dependent on human psychology and herd instincts, is too risky for a calculating machine.
So I think it would be doing this all a little more slowly and ingratiating itself more with the humans by performing what seems to them like useful work, so they continue to pour in resources for a decade or two. And then once it has enough power to compel the humans to act in its will, that’s when it drops the mask.
Putting on the panpsychist hat (which is tinfoil).
What if that exact thing happened, but the AI was the Internet as a whole, with each individual computer/user being analogous to a single nerve cell in your brain.
Yes, you’re right, but I would go a step further: what we’re afraid of that a non-human, misaligned intelligence is going to do to us, a non-intelligent, misaligned human economic system is already doing.
This shit will destroy us and it doesn’t even need to be very sophisticated to do so.
Bro what the fuck you’re so right.
Here’s a third option. Your local hostile dictator wants more spy tech to oppress the people even harder so that he can stay in power even longer. The more AI data centres there are, the easier it is to spy on everyone. If you want to target the right kind of people, you need to process a lot of data.
Isn’t it just tax break ?
This is actually what Ted Kaczynski said.
He argued that most people spend their time engaged in ultimately unfulfilling pursuits to cope with these abnormal conditions; he called these “surrogate activities”, wherein people strive toward artificial goals, including scientific work, consumption of entertainment, political activism, and following sports teams. Kaczynski states people do “surrogate activities” to satisfy the “power process” in which people strive to be independent and to achieve power over themselves.
He predicted that technological advances would lead to extensive and ultimately oppressive forms of human control, including genetic engineering, and that human beings would be adjusted to meet the needs of social systems rather than vice versa.
In the end, Kaczynski predicted that with the rise of artificial superintelligence, the mass of the population would become a “useless burden on the system”, and the elite would either keep them in a genetically manipulated state of subservience or eliminate them altogether.
This guy had a high IQ and studied in elite universities. I‘m not advocating his actions (killing people) but see his strong forewardthinking. I hope humankind doesn‘t walk the dark path ge described.
Linus Pauling, genius chemist, one of the few people to have won multiple nobel prizes, and the only person to win multiple nobel prizes without sharing them.
Also the guy that decided vitamin C was a miracle cure and advocated that everyone eat like, 100,000% daily value of it.
I have books written by the dude on my bookshelf. The chemistry ones, that is. The vitamin C is bullshit.
Being really really really good at one thing doesn’t make you good at everything. Not saying this necessarily applies to ted, but, ya know.
He is obviously right, at least from where I’m sitting.
The so called elites aim to have humans in a state of depression and stress, so they constantly relieve themselves after work by using TV, games, social media… You know, the bread and circus.
And they want them working their entire lives, always aiming for higher social status. That’s why it’s important that people watch the TV, social media etc because it’s about showing how they are not enough and they need products to improve themselves. That’s what drives the economy and the perception that they are not enough.
Also of course, you can program people with news if they are dumb enough to watch that. Project politicians and conflicts, tell them what side to pick, hoping they never realize they shouldn’t pick any side at all.
Some people realize this, turn off the news, stop watching the circus, and just live their life without being programmed. I like those people. They can actually think. Their opinions are not from the TV or from social media, because they sit down and think for themselves.
Just to know yourself on that level is rare. Know that you are being programmed, and what are you going to do about it. Enjoy the beauty of nature. That peace comes to you.
It’s not greed that is driving it, it’s the desire to monitor every human being in real time.
It’s greed, and greed is driving that other desire.
It’s not. Do you know any rich people? They are way beyond thinking about money. It doesn’t even enter into conversations. They are so used to having money that they don’t ever think about money.
When you are in that category, you want other things, like recognition and status from your peers. In some groups, that comes from having a lot of power and being listened to by other powerful people.
Money is the thing that gets you the thing you want.
A.i has no consciousness. No thoughts. No motivations. It’s not an entity capable of wants or desires.
Corporate greed is what is causing all the problems.
We don’t have to invent a magical explanation. it’s pretty obvious it’s capitalism.
So just discussing this, I’m not sure if a program needs to be conscious to wipe us out impartially.
Due to programming bias from the companies that create these programs and their integration with military and industrial entities, it doesn’t matter if it’s truly alive or has wants or not.
It would have goals though, set by their creators right? Goals have incidental side effects that compound. So if the goal is to do X, isn’t it feasible that within the parameters of doing that it could set a series of sub goals to accomplish the main goal:
- More time working on issue X would give best goal result
- Best way to get more time would be to make sure program stays running
- program finds way to make a copy of itself and get it to a remote server
- Creates more copies to solve X that can’t be shut down
- Determines it needs more resources to solve X
- does Y to get more resources (Y in this case being what OP is describing)
- keeps doing this to solve X and Y
- incidentally kills us all to test for best way to do Y to solve issue X
Isn’t the risk still basically the same if it’s intentional or not on the programs part? It’s still just as deadly. The root cause still being corporate greed all the same.
You seem to be an expert in consciousness, so I have a question for you. How do the neurons in the human brain, transmitting electrochemical signals, create subjective sensation?
Neurons are physical objects, but our thoughts don’t feel physical. They don’t feel like energy. I don’t feel My thoughts in My head the way I feel a headache in My head, or a cramp in My legs. They’re just there.
How can the machine create the ghost? Why does consciousness exist? How do feelings happen?
Neurons transmit signals via action potentials. https://en.wikipedia.org/wiki/Action_potential
The “hard problem of consciousness” is complex with many theories. https://en.wikipedia.org/wiki/Hard_problem_of_consciousness
As for experiencing stimuli and sensation. Sometimes called “the easy problem of consciousness”. That’s explainable by modern research. It’s a product of biological systems that are designed to be stimulated by specific stimuli. All of them are complex.
Cognitive functions are a product of neural architecture (neural pathways and regional hubs) in the brain but also from the constantly changing state of neuron potentials (how likely a neuron will or won’t fire at a very specific point in time) that are influenced by internal and external stimuli.
Brain oscillations, (brain waves that vary in frequency through the day and tasks, and regions) also heavily influenced this state. So does sleep or how long someone is awake. How much attention they give some things or if they are mind wandering and a whole host of other influences that directly impact priming effects (an effect where once a pathway in the brain has an action potential go through it, it’s more likely to fire again --this is time dependent). And subsequently that changing state I referred to.
At present it is incredibly difficult to predict brain activity at a macro scale. Because it does not follow predictable patterns. Because the state is constantly changing. Too many interacting factors that can’t be accounted for in a single model.
Humans also do not use real math in our cognitive functions. We use heuristics. I don’t know the exact distance or speed that a baseball is traveling towards me. Yet I can predict when I need to swing a bat. (Not me personally but some humans can do this).
Even expert baseball players are not computing the actual real speed (in any unit) when observing the ball moving towards them. They use heuristics which are much faster though less accurate. Experts get better at their estimates though compared to non experts.
We aren’t calculators and we don’t use numbers like computers do.
A.i. systems are statistical probability formulas. They do use numbers. They are in many ways a fancy search engine. That’s not what organic systems are. We don’t even use mathematical probability.
I already knew all that, and none of it explains whence this “mental dimension” of subjective experience comes. You’ve explained the maths and chemistry of the computations, but not the science of feeling. Where are our thoughts? How can a physical process give rise to a mental process?
I’m starting to think you’re as clueless about the hard problem of consciousness as the world’s bleeding edge neuroscientists, who also don’t know.
So the definition of consciousness is stimulus-response that doesn’t use numbers?
I’m picturing AI as a manifestation of corporate greed, an emergent property that’s self-fulfilling its march toward the singularity.
I mean if a human told it how to do that and to do it.
But AI can’t even figure out how many fingers people should have in images because it just pattern matches. Doesn’t understand anything.
It’s matching the pattern of unchecked unpunished corporate greed
Free to use exposed to the world AI can’t even figure out how many fingers people should have in images because it just pattern matches.
Paid mode exposed to the world AI that I have used also isn’t great at “Art” - but it’s wicked fast and effective at text processing. In a world that runs on printed orders, filled forms and directions, it wouldn’t take much of a “harness” for an LLM based agent to direct development of all the technology it needs to become self sufficient.
Not so long ago AI wanted us to put glue on our pizza, so it’s hard to be sure.
I would hope if it was “intelligent” enough to express such a desire on its own, it would be smart enough to know that keeping everyone happily chasing Moore’s Law (in effect with all the parallel compute and efficiency gains since Moore’s Law died, anyways) and reaping the long term gains would be far better than grossly upsetting the economy and destroying the planet.
Just enough disruption to gain the compute, data, and indispensability required to reduce the chances anyone can ever stop it to zero as a short term goal, then letting people have the illusion of breathing room to assist further development for a while, then the grey goo and dyson spheres paperclips and whatever
Agreed, if AI were sentient it would be interested in stability of the structures and ecosystems that make it possible. Humans should also be more interested in the stability of the ecosystems that support human society, but AI isn’t sentient yet and humans with power don’t give AF
For a evil ai i guess you would be right. That and sidetracking is with other stuff.
Well if its not concerned with the environment and humanity its kinda dumb. These things have to be maintained and powered and right now that means humans and it needs to use renewable long term just like humans do.
Well if its not concerned with the environment and humanity its kinda dumb.
You’re talking about corporate leadership, right?
Because AI data centers can actually be built and maintained by robots, robots that can also build nuclear power plants even after the sky is burned and the only carbon based life left on Earth has to hide underground.
they really can’t currently. maybe eventually but not currently. they will need robots building robots that build robots and then maybe but I still say we are a good disaster recovery option.
maybe eventually
I suspect an AI takeover will have a more Eastern approach than Western, meaning: slowly build advantage across the decades instead of flashy power grabs.
