Yes I’m referring to LLMs, and image classification models. And image generation models. And even the code that controls the Creepers in Minecraft. AGI isn’t a thing, but we’ve had AI for a looong time. It’s just not as flashy as it often looks in Sci-Fi movies.
Okay, great. AI as you describe exist, but are still things. Not sentient beings. Never will be. My point is the only people that think that they could be, are people that humanize pencils. Or gods. Or other things.
What makes humans different? If someone perfectly simulated my entire brain, would that digital brain be sentient? what even is sentience? I think it’s strange to say that AI will never be sentient.
Complexity for one. A cramped foot has an influence on the brain, as does apparently the gut bacteria. Focusing on the brain is a starting point and we don’t even understand that that well.
If someone perfectly simulated your entire brain, would that digital brain be sentient?
I don’t know. It could be. For now I don’t think so. Are you comparing that to an LLM? That would be like comparing the paths of snail slime to a comic. One could compare story lines and art styles to something that just isn’t there. And never will be.
What is sentience?
Sentience is the ability to experience feelings and sensations (wiki). A word not based on a clear understanding, but rather an attempt to categorize. Nonetheless, an LLM doesn’t experience anything. It uses pattern recognition and human provided categorization to try and create different stuff. All in the confines of the recognitions.
I think it’s strange to say that AI will never be sentient.
It’s why it’s important to distinguish between “AI” and “LLM”. AI, as an AGI, is something we might be able to build one day. LLMs might be a step on the way to this. But not the way they are now.
You have a point with most of the things you said, it’s mostly a matter of perspective and how you define stuff. the only thing I really fundamentally disagree with is equating AI to AGI.
AI refers to lots of things, including image recognition or generation models. AGI only refers to artificial general intelligence, aka the kind of AI you would see in science fiction movies. we have ai, we don’t have AGI
Yeah, I see how this looks. I was trying to comment about how for some people an AI (as in LLM) seems like a real person (or something different, but sentient), so I was reducing the category “AI” to LLMs.
AI is also, as you said, used for ie pathfinding algorithms in games. I never liked the word “AI” for that. But I came to terms with it as the AI got more sophisticated and rounded, making the figurines in games appear more natural in their behaviour. Also I don’t have a better word for it.
I used AGI because that is the only subpath of AI that I can consider having a chance of being/becoming sentient. That’s why I went into that direction, to oppose LLMs, despite LLMs being perceived by some as being sentient.
So yeah, the categorization was a bit off to drive home a point. I didn’t realize you wanted to discuss semantics (I know this sounds sarcastic, but I also tend to correct people on semantics if I can, therefore I don’t intend to be sarcastic.)
Yes I’m referring to LLMs, and image classification models. And image generation models. And even the code that controls the Creepers in Minecraft. AGI isn’t a thing, but we’ve had AI for a looong time. It’s just not as flashy as it often looks in Sci-Fi movies.
Okay, great. AI as you describe exist, but are still things. Not sentient beings. Never will be. My point is the only people that think that they could be, are people that humanize pencils. Or gods. Or other things.
So yes, AI exist. But not as sentient beings.
What makes humans different? If someone perfectly simulated my entire brain, would that digital brain be sentient? what even is sentience? I think it’s strange to say that AI will never be sentient.
Complexity for one. A cramped foot has an influence on the brain, as does apparently the gut bacteria. Focusing on the brain is a starting point and we don’t even understand that that well.
I don’t know. It could be. For now I don’t think so. Are you comparing that to an LLM? That would be like comparing the paths of snail slime to a comic. One could compare story lines and art styles to something that just isn’t there. And never will be.
Sentience is the ability to experience feelings and sensations (wiki). A word not based on a clear understanding, but rather an attempt to categorize. Nonetheless, an LLM doesn’t experience anything. It uses pattern recognition and human provided categorization to try and create different stuff. All in the confines of the recognitions.
It’s why it’s important to distinguish between “AI” and “LLM”. AI, as an AGI, is something we might be able to build one day. LLMs might be a step on the way to this. But not the way they are now.
You have a point with most of the things you said, it’s mostly a matter of perspective and how you define stuff. the only thing I really fundamentally disagree with is equating AI to AGI.
Why do you disagree with that? No, that’s a stupid question. How do you disagree with that? Can you elaborate your point?
AI refers to lots of things, including image recognition or generation models. AGI only refers to artificial general intelligence, aka the kind of AI you would see in science fiction movies. we have ai, we don’t have AGI
Yeah, I see how this looks. I was trying to comment about how for some people an AI (as in LLM) seems like a real person (or something different, but sentient), so I was reducing the category “AI” to LLMs.
AI is also, as you said, used for ie pathfinding algorithms in games. I never liked the word “AI” for that. But I came to terms with it as the AI got more sophisticated and rounded, making the figurines in games appear more natural in their behaviour. Also I don’t have a better word for it.
I used AGI because that is the only subpath of AI that I can consider having a chance of being/becoming sentient. That’s why I went into that direction, to oppose LLMs, despite LLMs being perceived by some as being sentient.
So yeah, the categorization was a bit off to drive home a point. I didn’t realize you wanted to discuss semantics (I know this sounds sarcastic, but I also tend to correct people on semantics if I can, therefore I don’t intend to be sarcastic.)
Yeah I love this Train of thought because it’s an interesting thing to consider.