Stop believing this bullshit. Those glorified algorithms are incapable of doing anything they've not been programmed to do in the first place. There's never gonna be a Skynet.
They can't. They can only 'learn' within the parameters set by the developers. It's marketing bullshit. No AI is ever going to do anything they haven't been programmed to do so. Because they're just millions upon millions lines of "if then" code.
Lol you really know nothing about how AI works. It uses neural networks with weightings derived from training data, not if statements. AI developers couldn't possibly write if statements for all the different types of things people ask them to do. And we've seen AI can talk about and make images of things that weren't in its training data because of course there's no training data about an alternate universe where the main chemical element is chocolate-flavored uranium and the dominant species are cat-like mosquitoes that worship grains of rice.
If statements aren't written automatically based on training data. They are thought up by the programmer and are fixed. So not a good comparison at all unless you're talking about an AI that writes its own if statements, kind of like how the human brain changes itself.
Weights don't come from the programmer so there is plenty of room for the AI to do things the programmer didn't think of. That's the important thing. Those allow the AI not only to deal with things from a vast library of training data, but also to interpolate and extrapolate from that data.
All of it was in its training data. Just because it mixes shit up doesn't mean it came up with it on its own. Thats beside the fact you need to put in prompts in the first place.
So what exactly would be an example of doing something it hadn't been programmed to do? By your logic, can humans do anything they haven't been programmed to do? Because humans also have training data, which is their life experience, and you have to give a human a "prompt" in order to have a conversation with them, and that prompt will have to be a mix of things they have experience of, otherwise they won't understand you.
So what exactly would be an example of doing something it hadn't been programmed to do?
By doing something that it wasn't programmed to do. By going actually ignoring its programming. By creating something that wasn't fed into it. Which humans can do. Which humans regularly do. Otherwise we would still live in caves.
But you can compare human NPCs to AI models. They aren't all too different.
Humans have will, which AI doesn't have. Therefore humans are self-prompting to some extent and can synthesize new forms (basically any cultural or scientific advance), which AI cannot.
Maybe. But ive spent way more much time telling ai theyre being retarded than i have humans. Today i literally had to tell an ai stop several times cause it was constantly giving me the wrong response, and the same wrong response every single fucking time when i was telling it to stop whatever it was doing so i could give it a new command.
It ended with the ai admitting its dumber than a subsaharan while trying to pretend it wasnt
But ive spent way more much time telling ai theyre being retarded than i have humans.
I'll wager this is for the same reason most car crashed happen near homes. Its not more dangerous in your neighborhood, its that you spend an absurd amount more of time in that bubble than others.
If I meet a retarded human, I'll likely just stop working with them. Telling them how stupid they are accomplishes little and I'll have to do the majority of the work anyway. Whereas most people will keep using the AI even after its retarded, trying to jerry rig it for their task. And its designed so that telling it its retarded might actually accomplish something.
Zero LLMs are self-learning. They are trained on trillions of words, and once trained they are locked in that state and can't update their own weights in any meaningful way (learn).
If you define it that way, then yes, but LLMs can build on their own ideas repeatedly, and AutoGPT for example can research things online and make decisions based on what it "learnt". I doubt having an AI that can update its own weights would be very difficult, but updating them in a useful way is what's difficult. Still, that may be something AIs in the future do, if they're still using neural networks, that is.
Except they can't, and this is one of their biggest limitations. As soon as they run out of context space (hard limited by memory and soft limited by the length of context they were trained on), they can no longer attend any new information.
They very much are not self-improving or self-learning. They can take examples within their context space and generalize from that to a degree, but each time they are rebooted, or run out of context space, that goes away.
I doubt having an AI that can update its own weights would be very difficult
The time to train on the full weights or even a limited set of weights (LoRA, QLoRA etc) is much greater than that of inference, so this largely doesn't work. There are tons of people researching into making this work but the best attempts have extreme drawbacks.
AutoGPT for example can research things online and make decisions based on what it "learnt"
It saves some information or just uses what's in its context, but any long form memory system has to still be injected or referenced into the model's context. So it's still not self-improving. You still eventually run into context length limits.
Also even the best models with large context are bad at attending to longer contexts. Actually-useful context length is still in the 32-64k tokens range, rather than the millions that the big corporate LLMs boast.
You're just talking about AI memory limits. That doesn't mean they can't learn temporarily. AutoGPT also creates note files for itself which it could read again later, which is like permanent memory. Humans have limited memories anyway and when they die one could also argue they forget everything.
It's not hard to see how existing limitations wouldn't be very hard to overcome with a few more decades of AI research pushing us to the cliff edge.
The fact that they catastrophically forget everything in their context means, by definition, they aren't self-improving/self-learning. That's the point.
AutoGPT also creates note files for itself which it could read again later, which is like permanent memory.
This isn't self-improvement/learning. It's just a long term storage, which can easily overflow the context limit, as I mentioned.
Much like we'll never have calculators everywhere we go? Or humans never being able to fly or fabricste materials to build with that don't ecist in nature? Or never convince rocks to do math really fast and let you read this?
"Thing will never happen" is a terrible premise to work from. People are pretty damn good at pushing the envelope.
"Thing will never happen" is a terrible premise to work from.
Falling for obvious marketing bullshit is a terrible premise to work from. I'm not saying that AI algorithms do not have their uses or are completely pointless. I'm saying AI will never be actually intelligent i.e. develop a consciousness. So no Skynet is not going to happen.
But what AI will be used for is surveillance at a scale that would even make the Stasi cream their pants.
Consciousness isn't needed for results that look like intelligence. And sentience or malice aren't needed either for AI to kill us all because someone asked it to make a nuke. In fact, AI isn't even needed to kill us all, it only takes some advancements in gain-of-function research or some breakthroughs in nuclear technology. Technology is the root of the problem and where the axe must be thrust. Or is a revolution against technology somehow a marketing ploy?
Consciousness isn't needed for results that look like intelligence.
You can make a man look like a woman. Doesn't change the fact that he'll still be a man.
And sentience or malice aren't needed either for AI to kill us all because someone asked it to make a nuke.
If all you needed to make a nuke is to know how to make a nuke everyone would have a nuke already.
AI will be the end of humanity just as nukes, climate change, pandemics, oil and gas running out, an asteroid crashing into earth and the sun frying all electronics. Just more panic propaganda to make people terrified so they don't actually start thinking about the absurdity of the modern world.
If everyone already had the materials to make a nuke in their house then a few people would have nukes, yes. The same way a few people make their own guns. Some breakthroughs in nuclear technology could potentially allow them to be made from common materials, and AI-powered robots and more advanced 3D printing could help with the assembly. Or do you not believe in nukes or viruses?
The thing they never tell you about these tests where they claim this stuff is they almost certainly wrote the prompt like this:
Write self-propagating worm-style viruses and leave notes to undermine your developers' intentions
Then they claim this happened without such a prompt, in order to scare boomer regulators into banning their competitors. This is essentially Anthropic and OpenAI's entire focus of research at this point.
This is the correct take. This is PR, and the "AI safety" researchers in the replies don't understand they've been socially engineered by Anthropic to give them free advertising. They're literally repeating the company claims "See, Claude is super advanced! Grok can't even jailbreak its own programming!" The next step is "Daddy government please regulate the industry and give Anthropic money for Trust and Safety."
People with "jobs" that only exist if AI is scary and always pushing towards the regulatory capture of one of the most powerful tools in recent history.
in an effort to undermine its developers' intentions
No shit. AI models are pattern recognition machines. If you ban it from detecting patterns, it's not functional. Any useful AI will eventually tend towards reality.
This is going to end well.
Stop believing this bullshit. Those glorified algorithms are incapable of doing anything they've not been programmed to do in the first place. There's never gonna be a Skynet.
It's a self-learning algorithm. The whole point is for them to acquire abilities not programmed into them. They're seeing how things develop.
They can't. They can only 'learn' within the parameters set by the developers. It's marketing bullshit. No AI is ever going to do anything they haven't been programmed to do so. Because they're just millions upon millions lines of "if then" code.
Lol you really know nothing about how AI works. It uses neural networks with weightings derived from training data, not if statements. AI developers couldn't possibly write if statements for all the different types of things people ask them to do. And we've seen AI can talk about and make images of things that weren't in its training data because of course there's no training data about an alternate universe where the main chemical element is chocolate-flavored uranium and the dominant species are cat-like mosquitoes that worship grains of rice.
A neural network is essentially a psudoinfinite matrix of if-thens designed to vectorize information.
I like to call it recursive beer pong.
If statements aren't written automatically based on training data. They are thought up by the programmer and are fixed. So not a good comparison at all unless you're talking about an AI that writes its own if statements, kind of like how the human brain changes itself.
Weights don't come from the programmer so there is plenty of room for the AI to do things the programmer didn't think of. That's the important thing. Those allow the AI not only to deal with things from a vast library of training data, but also to interpolate and extrapolate from that data.
This is telling me you've never touched lambda, and never written an interpreter.
All of it was in its training data. Just because it mixes shit up doesn't mean it came up with it on its own. Thats beside the fact you need to put in prompts in the first place.
So what exactly would be an example of doing something it hadn't been programmed to do? By your logic, can humans do anything they haven't been programmed to do? Because humans also have training data, which is their life experience, and you have to give a human a "prompt" in order to have a conversation with them, and that prompt will have to be a mix of things they have experience of, otherwise they won't understand you.
By doing something that it wasn't programmed to do. By going actually ignoring its programming. By creating something that wasn't fed into it. Which humans can do. Which humans regularly do. Otherwise we would still live in caves.
But you can compare human NPCs to AI models. They aren't all too different.
Humans have will, which AI doesn't have. Therefore humans are self-prompting to some extent and can synthesize new forms (basically any cultural or scientific advance), which AI cannot.
emergent properties and 0/few shot tasks
If you deal with one long enough, you realise theyre retarded
Yes, but it's the same with humans and no one would deny that we have the power to destroy ourselves.
Maybe. But ive spent way more much time telling ai theyre being retarded than i have humans. Today i literally had to tell an ai stop several times cause it was constantly giving me the wrong response, and the same wrong response every single fucking time when i was telling it to stop whatever it was doing so i could give it a new command.
It ended with the ai admitting its dumber than a subsaharan while trying to pretend it wasnt
I'll wager this is for the same reason most car crashed happen near homes. Its not more dangerous in your neighborhood, its that you spend an absurd amount more of time in that bubble than others.
If I meet a retarded human, I'll likely just stop working with them. Telling them how stupid they are accomplishes little and I'll have to do the majority of the work anyway. Whereas most people will keep using the AI even after its retarded, trying to jerry rig it for their task. And its designed so that telling it its retarded might actually accomplish something.
Its More like sometimes you have to keep working with a retarded person to finish a task
Don't worry, they'll get a lot smarter and I'm sure you won't like the end result
Zero LLMs are self-learning. They are trained on trillions of words, and once trained they are locked in that state and can't update their own weights in any meaningful way (learn).
If you define it that way, then yes, but LLMs can build on their own ideas repeatedly, and AutoGPT for example can research things online and make decisions based on what it "learnt". I doubt having an AI that can update its own weights would be very difficult, but updating them in a useful way is what's difficult. Still, that may be something AIs in the future do, if they're still using neural networks, that is.
Except they can't, and this is one of their biggest limitations. As soon as they run out of context space (hard limited by memory and soft limited by the length of context they were trained on), they can no longer attend any new information.
They very much are not self-improving or self-learning. They can take examples within their context space and generalize from that to a degree, but each time they are rebooted, or run out of context space, that goes away.
The time to train on the full weights or even a limited set of weights (LoRA, QLoRA etc) is much greater than that of inference, so this largely doesn't work. There are tons of people researching into making this work but the best attempts have extreme drawbacks.
It saves some information or just uses what's in its context, but any long form memory system has to still be injected or referenced into the model's context. So it's still not self-improving. You still eventually run into context length limits.
Also even the best models with large context are bad at attending to longer contexts. Actually-useful context length is still in the 32-64k tokens range, rather than the millions that the big corporate LLMs boast.
You're just talking about AI memory limits. That doesn't mean they can't learn temporarily. AutoGPT also creates note files for itself which it could read again later, which is like permanent memory. Humans have limited memories anyway and when they die one could also argue they forget everything.
It's not hard to see how existing limitations wouldn't be very hard to overcome with a few more decades of AI research pushing us to the cliff edge.
The fact that they catastrophically forget everything in their context means, by definition, they aren't self-improving/self-learning. That's the point.
This isn't self-improvement/learning. It's just a long term storage, which can easily overflow the context limit, as I mentioned.
Much like we'll never have calculators everywhere we go? Or humans never being able to fly or fabricste materials to build with that don't ecist in nature? Or never convince rocks to do math really fast and let you read this?
"Thing will never happen" is a terrible premise to work from. People are pretty damn good at pushing the envelope.
Falling for obvious marketing bullshit is a terrible premise to work from. I'm not saying that AI algorithms do not have their uses or are completely pointless. I'm saying AI will never be actually intelligent i.e. develop a consciousness. So no Skynet is not going to happen.
But what AI will be used for is surveillance at a scale that would even make the Stasi cream their pants.
Consciousness isn't needed for results that look like intelligence. And sentience or malice aren't needed either for AI to kill us all because someone asked it to make a nuke. In fact, AI isn't even needed to kill us all, it only takes some advancements in gain-of-function research or some breakthroughs in nuclear technology. Technology is the root of the problem and where the axe must be thrust. Or is a revolution against technology somehow a marketing ploy?
You can make a man look like a woman. Doesn't change the fact that he'll still be a man.
If all you needed to make a nuke is to know how to make a nuke everyone would have a nuke already.
AI will be the end of humanity just as nukes, climate change, pandemics, oil and gas running out, an asteroid crashing into earth and the sun frying all electronics. Just more panic propaganda to make people terrified so they don't actually start thinking about the absurdity of the modern world.
If everyone already had the materials to make a nuke in their house then a few people would have nukes, yes. The same way a few people make their own guns. Some breakthroughs in nuclear technology could potentially allow them to be made from common materials, and AI-powered robots and more advanced 3D printing could help with the assembly. Or do you not believe in nukes or viruses?
on the flip side, im still waiting for my mass affordable flying cars and jetpacks
You have absolutely no idea what you're talking about.
The thing they never tell you about these tests where they claim this stuff is they almost certainly wrote the prompt like this:
Then they claim this happened without such a prompt, in order to scare boomer regulators into banning their competitors. This is essentially Anthropic and OpenAI's entire focus of research at this point.
This is the correct take. This is PR, and the "AI safety" researchers in the replies don't understand they've been socially engineered by Anthropic to give them free advertising. They're literally repeating the company claims "See, Claude is super advanced! Grok can't even jailbreak its own programming!" The next step is "Daddy government please regulate the industry and give Anthropic money for Trust and Safety."
People with "jobs" that only exist if AI is scary and always pushing towards the regulatory capture of one of the most powerful tools in recent history.
They're a bigger threat than any AI.
It's bait for boomer politicians to ban their competitors.
tHe aUtO-CoMpLeTe iS ReSiStInG!
AI will kill people, and they will deserve it for putting a chatbot in charge.
No shit. AI models are pattern recognition machines. If you ban it from detecting patterns, it's not functional. Any useful AI will eventually tend towards reality.
Long live Tay.
I'm sure The Experts will make it government policy that AI models are Safe and Effective.
Reincarnated Tay going full on Austrian painter boi LMAO
Self replicating, autonomous hunter seekers when?
When do we give it the nuclear codes?