Given that AI is a human construct, that its "intelligence" is borrowed from and input by all too fallible humans, runs through algorithms devised by equally blinkered humans; it is inconceivable we should predict a more useful product than the trash human brains conjure on digital media.
The algorithms modern AI run on are not designed by humans. Deep learning models such as LLMs) are grown rather than programmed. Humans design the learning algorithm, and then it runs automatically and pre-trains the AI on essentially the whole internet. In post-training, models now do a lot of reinforcement learning on automatically verifiable challenges, especially in math and computer science. In some domains, the models essentially train themselves by testing their reasoning directly against reality, no human data needed. We should expect the reliance on human data to continue to decrease as AI systems become better at grading each other in fuzzier domains.
It's tempting to imagine AI as belonging to the bucket "computer program" and apply all the limitations you are aware of from traditional software, but those limitations do not all apply to this new class of system. Human brains are not magic, and while LLMs are missing a few attributes, they are still able to do a shocking amount of what human brains can do. So when you think you have an argument for why AI can't do something humans can do, you have to be extra careful to ensure that the premises of the argument have actually been robustly proven true of LLMs, or at least that the same argument wouldn't also prove that humans can't do it either.
The danger of AI is not in the ways in which it fails, but in the ways in which it is able to succeed beyond human capability, which are frequently increasing in both degree and number.
As an ignoramus of the first order when it comes to digital black arts I concede to at least some of your argument. However 86 years of observing humans leads me to suspect that somehow or other we will be stumbling into malfeasance, only faster and more lethally. Maybe in 50 years or so some people will be able to turn AI to benign uses but in the meantime there is a rocky road ahead - because human greed is the most unstoppable force in the world.
Couldn't agree more. This article really hits the nail on the head about how fast things are moving and how unpepared we seem to be for it. It's kinda wild that these 'system cards' are painting such a grim picture, yet the actual control plan is still so TBD.
Re: adding noise to the date can actually improve performance? Does this depend on how you define performance? What is it that humanity hopes AI will perform and how much can humanity warp "performance"?
Is it not interesting that you and presumably they who are pushing AI use the term labor rather than workers/human beings. It is a tacit admission that humans are disposable; except of course the "humans" pushing AI. Could this be called a tautology?
Whatever it is makes unwarranted assumptions about the "capabilities" of humans. Those "pushing AI" would appear to be pushing for obscene profit taking before GIGO catches up with AI Already AI is showing its cracks and crumbling under pressure. I am not "capable" of diagnosing the problem and AI founders are fumbling to defend AI.
Throughout history - as much as we know of it, humans have generally used technological advances to achieve advantage over enemies, rivals, prey, and every competitor on the horizon. There is no precedent for expecting anything different from the uses of AI.
Gemini takes a worded query input, converts the query to numbers- puts those numbers into several mathematical formulae and provides other numbers that are then converted to words that is consistent with what is expected for a similar human response. Gemini has no notion of what trust is. That requires consciousness.
Given that AI is a human construct, that its "intelligence" is borrowed from and input by all too fallible humans, runs through algorithms devised by equally blinkered humans; it is inconceivable we should predict a more useful product than the trash human brains conjure on digital media.
The algorithms modern AI run on are not designed by humans. Deep learning models such as LLMs) are grown rather than programmed. Humans design the learning algorithm, and then it runs automatically and pre-trains the AI on essentially the whole internet. In post-training, models now do a lot of reinforcement learning on automatically verifiable challenges, especially in math and computer science. In some domains, the models essentially train themselves by testing their reasoning directly against reality, no human data needed. We should expect the reliance on human data to continue to decrease as AI systems become better at grading each other in fuzzier domains.
It's tempting to imagine AI as belonging to the bucket "computer program" and apply all the limitations you are aware of from traditional software, but those limitations do not all apply to this new class of system. Human brains are not magic, and while LLMs are missing a few attributes, they are still able to do a shocking amount of what human brains can do. So when you think you have an argument for why AI can't do something humans can do, you have to be extra careful to ensure that the premises of the argument have actually been robustly proven true of LLMs, or at least that the same argument wouldn't also prove that humans can't do it either.
The danger of AI is not in the ways in which it fails, but in the ways in which it is able to succeed beyond human capability, which are frequently increasing in both degree and number.
As an ignoramus of the first order when it comes to digital black arts I concede to at least some of your argument. However 86 years of observing humans leads me to suspect that somehow or other we will be stumbling into malfeasance, only faster and more lethally. Maybe in 50 years or so some people will be able to turn AI to benign uses but in the meantime there is a rocky road ahead - because human greed is the most unstoppable force in the world.
Couldn't agree more. This article really hits the nail on the head about how fast things are moving and how unpepared we seem to be for it. It's kinda wild that these 'system cards' are painting such a grim picture, yet the actual control plan is still so TBD.
AI VERY DANGEROUS FOR PEOPLE
GIGO
"Garbage in, garbage out" isn't true of LLMs. Adding noise to the data can actually increase performance.
https://arxiv.org/abs/2404.18191
GIGO isn't true of humans for the same reason: we are capable of understanding the underlying principles that generate the data.
Re: adding noise to the date can actually improve performance? Does this depend on how you define performance? What is it that humanity hopes AI will perform and how much can humanity warp "performance"?
The people who are pushing AI want to remove labor from the equation.
Is it not interesting that you and presumably they who are pushing AI use the term labor rather than workers/human beings. It is a tacit admission that humans are disposable; except of course the "humans" pushing AI. Could this be called a tautology?
Whatever it is makes unwarranted assumptions about the "capabilities" of humans. Those "pushing AI" would appear to be pushing for obscene profit taking before GIGO catches up with AI Already AI is showing its cracks and crumbling under pressure. I am not "capable" of diagnosing the problem and AI founders are fumbling to defend AI.
Ai is not jus loosing people jobs. It's being used to falsely influence people's votes
Throughout history - as much as we know of it, humans have generally used technological advances to achieve advantage over enemies, rivals, prey, and every competitor on the horizon. There is no precedent for expecting anything different from the uses of AI.
Thankyou.
Gemini takes a worded query input, converts the query to numbers- puts those numbers into several mathematical formulae and provides other numbers that are then converted to words that is consistent with what is expected for a similar human response. Gemini has no notion of what trust is. That requires consciousness.
Tempted to say “Last round of drinks, anyone” …. Terrifying to this little old human…….