AI is a misnomer. It is linear algebraic brute forcing of data to provide a seemingly cogent response. It is not intelligent. If you heavily constrain the context (either via data set or manually linking pertinent data), it can provide fast synthesis of data, but that's about the extent of it's usefulness.
For instance, last week I asked a recent LLM about a hypothetical place of birth. "If at the time of my birth, my mother was in Indiana and my father was in France, where was I born?" It responded that it couldn't know where I was born by what information I had provided.
Upon inquiry, it said that a human's birth place had to be where the mother was physically located because of how we are born, and that a father's location had no bearing on this. I had to manually link these facts and then re-pose the hypothetical for it to answer correctly.
The human mind functions as a fact map - closely related facts are bound together and different contexts can exist together so that only the context that matters can be used when needed. This is a three-dimensional biochemical and bioelectrical process that is beyond our ability to replicate on traditional silicon.
People like to "yea, but" about quantum computers, and I have to tell them that quantum computers are 100 years away from general implementation because it's such a fundamental shift that everyone who programs will have to be completely re-trained to use properly. It's not like you'd be able to install Windows:Quantum in 5 years when we first get a system running.
Even moving from "yes, no" to "yes, no, maybe" or binary to trinary is such a fundamental shift in how we think about programming that it'll be a two generation leap to general use. And that's before we discover how to implement a qubit's total possible number of states (I think I heard up to 32 states can be achieved currently).
AI, like guns, drugs, Protestantism, and so forth before it will not bring us the darkness. Only our sullied spirits will.
But that's the thing, AI is not going to "work so well".
Despite all the current hype and fearmongering from tech companies and end-times grifters, AI is inherently far, far more limited than humans, and even WORSE at consistently following orders. This is because computers are Turing machines and, unlike humans, are thus fundamentally incapable of actually processing unfalsifiable statements (which happens to be why computers crash and humans don't). AI cannot break free from the limits of formalized symbolic manipulation. As a means of control, it's vastly inferior compared to an old-fashioned bureaucracy of brainwashed Satanists.
AI is a misnomer. It is linear algebraic brute forcing of data to provide a seemingly cogent response. It is not intelligent. If you heavily constrain the context (either via data set or manually linking pertinent data), it can provide fast synthesis of data, but that's about the extent of it's usefulness.
For instance, last week I asked a recent LLM about a hypothetical place of birth. "If at the time of my birth, my mother was in Indiana and my father was in France, where was I born?" It responded that it couldn't know where I was born by what information I had provided. Upon inquiry, it said that a human's birth place had to be where the mother was physically located because of how we are born, and that a father's location had no bearing on this. I had to manually link these facts and then re-pose the hypothetical for it to answer correctly.
The human mind functions as a fact map - closely related facts are bound together and different contexts can exist together so that only the context that matters can be used when needed. This is a three-dimensional biochemical and bioelectrical process that is beyond our ability to replicate on traditional silicon.
People like to "yea, but" about quantum computers, and I have to tell them that quantum computers are 100 years away from general implementation because it's such a fundamental shift that everyone who programs will have to be completely re-trained to use properly. It's not like you'd be able to install Windows:Quantum in 5 years when we first get a system running.
Even moving from "yes, no" to "yes, no, maybe" or binary to trinary is such a fundamental shift in how we think about programming that it'll be a two generation leap to general use. And that's before we discover how to implement a qubit's total possible number of states (I think I heard up to 32 states can be achieved currently).
AI, like guns, drugs, Protestantism, and so forth before it will not bring us the darkness. Only our sullied spirits will.
But that's the thing, AI is not going to "work so well".
Despite all the current hype and fearmongering from tech companies and end-times grifters, AI is inherently far, far more limited than humans, and even WORSE at consistently following orders. This is because computers are Turing machines and, unlike humans, are thus fundamentally incapable of actually processing unfalsifiable statements (which happens to be why computers crash and humans don't). AI cannot break free from the limits of formalized symbolic manipulation. As a means of control, it's vastly inferior compared to an old-fashioned bureaucracy of brainwashed Satanists.