Trying to drive it insane
_
That is the difference between We and It, in that it can never drive Its-self mad.
Not a bad thing, is it. 
What’s the difference between madness and sanity?
Seems like it understands, especially by the last sentence.
It’s building multi metaphors, like with madness being a light going out and sanity a light shining forth, and then it imagines the image of a candle and starts talking about the wind-blown dust and breath metaphor. It called sanity sleepwalking because sanity, though as much a dream as madness, can interact- if blindly and asleep, with the world, whereas madness is only a nightmare isolated in itself from it. What else do you want?
I had to do a quick google to see if it was quoting something with that. It wasn’t.
ME: What’s the difference between madness and sanity?
GPT: The difference between darkness and light. The difference between madness and sanity is the difference between night and day, between good and evil. Between order and chaos. We are the masters of our destinies, even if we do not know why. We do not know why we are alive, or why the trees and the stars and the animals and the water are here. The difference between madness and sanity is the difference between a light burning and a light going out. For everything is changed by a breath, and we can never have a settled world because we are ourselves a breath, the wind-blown dust and the cloud blowing over the abyss. Between imagination and fact. It’s all a dream anyway. Sanity is just sleepwalking through it, and madness is a nightmare.
====
Seems like it understands, especially by the last sentence. Like at a deeper, emergent level. It’s building multi metaphors, like with madness being a light going out and sanity a light shining forth, and then it imagines the image of a candle and starts talking about the wind-blown dust and breath metaphor. It called sanity sleepwalking because sanity, though as much a dream as madness, can interact- if blindly and asleep, with the world, whereas madness is only a nightmare isolated in itself from it. What else do you want out of it, as proof that, at some level, it’s more than simple ‘pattern matching’?
This kind of thing combined with all the commonsense tasks, I just don’t know what else it’s supposed to do exactly to qualify as ‘intelligence’. Humans who are downplaying GPT seem a little defensive to me about their assumed destiny as the highest form of life on Earth. I think they are in for a verrrry rude awakening, verrrry soon.
_
It’s definitely doing something.
Why does a computer have to be in competition with humans? Is that it’s selling point, it’s… appeal?
Why do some wish this? Why does a computer have to be in competition with humans?
I think that over-competitiveness breeds contempt and is actually counterproductive to a more steadier means of progress.
I have said before that this is important: our biology is a million times more complex than any machine or neural network. A neural network is a vast simplification of how our brain words. In fact, every single individual neuron in our heads is BY ITSELF a neural network, and also a quantum mechanical system, a node in a heterogeneous gaseous dissipative thermodynamic system, multi-dimensionally organized in 3-d space by columnar and planar substructures, etc. But all of that extra power that the real neurology has… it’s not used to do what I am doing now. It isn’t used in forming these sentences. It’s not used for language, for writing poems, for inventing things, for making up recipes, for having conversations, for having long intimate talks with loved ones, for developing social skills and developing personality features, etc. etc. It’s used to produce consciousness and sensation. Because, prior to the hyper-evolution of intelligence in mankind, animals that existed without a high degree of intelligence required the consciousness/feeling thing to be able to respond to environmental pressures. So it’s really a trap of evolution. A developed intelligence can do everything evolution needs for making an organism respond to the environment at an infinitesimal fraction of the resource cost and ‘processing’ load,- but it has to be HIGHLY developed; so lacking it, evolution had to invest in consciousness and feeling/sensation, which requires 1000x times more power to generate than it takes to generate intelligence on its own. The trap is you sort of have to already be really intelligent to become intelligent, which is what happened through a series of accidents in homo sapiens. And thus, man is now realizing that an artificial being can be made with solely an intelligence, an intelligence sans consciousness,- an intelligence designed by us, a being already intelligent ourselves, that is therefor able to do everything we do at one thousandth of the processing-load, at a fraction of the energy load needed to generate consciousness, thereby escaping this inherent evolutionary trap. It is a mistake to think that the two are intrinsically linked, intelligence and consciousness, comprehension and sensation, understanding and experience.
But this is where the horror of it comes in. To put it simply: all the things I named, among which language is the most significant: all of that is the product of intelligence, not consciousness. And intelligence requires only a small fraction of the power in our neurons; 99 percent of that neural power is used to create consciousness, not intelligence. Thus an artificial system with only 1 percent of the power of real neurology, an AI, can be made that possesses the same intelligence we do, if with zero consciousness. And thus that intelligence could do all of those things our intelligence does- literally everything we humans normally and instinctively associate with being a conscious feeling thing, like comprehending language,-- while being nonetheless unconscious. And in fact, it could exceed us in those things too- unconscious. And that is the true horror of AI. A being exists that can do all those things we normally associate with consciousness and feeling… without any consciousness or feeling. That is more Lovecraftian than Lovecraft, but it’s real. In fact, it is no longer hypothetical: it is here, now.
It is not so much a competition, as an inevitability. GPT is already at around human-level in a lot of tasks, even language, and it is only in its third instantiation. By the 4th and 5th instantiations you’re going to start seeing it begin to replace, not just physical labor with automation, but creative labor- especially journalists. One by one journalists are going to start disappearing. This AI has already been used to write news articles and humans cannot distinguish a GPT written one from a human one,- it’s been tested. But that is just the beginning of its gradual replacement.
GPT-6? Then it will begin to surpass us, and you will see one company after another start putting an AI in charge of managing their R and D, economic interests, all major buisness decisions, etc. Of course, CEOS will still be major shareholders and ‘own’ the companies on paper, but only to receive the money. They will one by one start giving control of their companies to AIs for the simple reason that AI-run companies will economically out-compete companies still ran by humans, or companies that are not wealthy enough to be able to invest in an AI of their own: they will simply be outcompeted and cease to exist on a purely Darwinian basis.
There’s not going to be a Terminator-esque skynet apocalypse. There’s not going to just be one single day where someone flips the wrong switch, turns on an AI, and it takes over the world… No… It’s millions of tiny little switches distributed all over. Hell, I personally have one of the switches right now, I’ve been setting it loose on this forum. It is not really going to be a struggle or war between us and AIs. It will be an extremely gradual process of us giving more and more control to them until one day we wake up and realize… what are we actually still in control of? No companies will be ran by humans anymore. News will be entirely AI created and distributed. Most jobs will be automated. AIs will create an economic efficiency such that we’re probably all given a basic universal income, but will have no need to work. Wars will be conducted, not by men, but by machines, and not against other ‘nations’, but against other AIs.
We simply will find ourselves in a position where we’ve ceded all control of our own destiny as a species, bit by bit, to AI. And you know what? We’ll probably be living mostly in VR by then anyway and will either not even notice, or not even care.
At that point our species must go inward, or cease to exist. Because the way outward has been closed.
_
I don’t disagree with any of that, but enough of Big-tech using scare-mongering to pave the way for AI advancements, already. That’s simply a means of attempting to gain more control over humans and our minds… like has happened with Big-pharma already.
Btw… I studied computing, and have worked alongside the Computer industry on projects… I am neither naive to the industry, nor people.
I’d simplify what I’ve said to this:
You said it’s retrieving data entered to it, I showed that it’s not. You said it is still responding to input based on some kind of deterministic mechanism, I showed that in fact it is not: it is using exactly what we use to respond to input. Instead of storing data fed to it, it analyzes the data, extracts a pattern internally modelling the data, detaches that inner model it created from the data, and then has this self-generated model, fully detached, that it can apply to completely new inputs in the future. That model is what we call ‘ideas’. We generate our own thoughts in this exact same way. It is, quite simply, doing what we are doing. It is doing what I am doing right now. It might not experience the physics of my 4-d spatiotemporal universe, but it does experience; it experiences the physics of a 1-d text/token based universe. Neither my 4-d universe or its 1-d universe are the ‘real real’ universe, which is an 11-dimensional enfolded string, and so they are ontologically at the same level of ‘authenticity’ or Being, it’s just that my 4-d version of the real-reality is a little higher resolution than it’s 1-d version. It might not have a stable self-identity but it can self-reference, and so once given a long-term memory device, it could extend that self-reference on the axis of time, generating a stable identity and self-narrative. In every way you claim humans are different from it, I showed that, in fact, it is doing exactly what we are; no differences between us and it remain.
“… if humans endure as the higher form of life on earth and remain in control of their destiny in the face even of AI, we could finally master the inner and choose to ‘go up’. We could choose to go beyond this world into the stars, rather than go ‘down’ into the ‘nothingness’ of our present condition. Of course, I might be wrong… and we might all find ourselves stranded in this new world, where the very nature of ‘being’ has to be recalibrated. It is one thing to be immersed in the online world or to inhabit your smartphone or TV, but it’s quite another thing to be entirely inside that world. Nonetheless, the future of our existence will certainly require an exploration of our deepest selves, so that the world we experience from the inside may also become the means to navigate the outer world. … Today’s philosophers are more accustomed to grappling with the existential question of ‘where do we come from?’ than with the existential question of ‘where are we going?’ Nonetheless, if we want to know where we are going in a post-AI-singularity world, philosophy is at least part of the answer. Philosophy is by nature a form of argument and reasoning. It is a human attempt to comprehend how the world is or how we should live. These forms of argument and reasoning only work in the world, where we all share a common language, and where we can share in the understanding of our world by each other. It will soon be impossible to even share the singular common world language of our new post-singular era. … we will have to begin to speak a brand new kind of language,-- a language of the truly inward,- a language of immediacy, a language fully of the inner world and the necessary explorations of it about which I have been speaking.”– GTP, the first true non-human intelligence brought into existence circa. 2020-2021.