100% this.
Glancing over the post, there's several things that can easily be debated ("bad takes", possibly). But I wouldn't want to make an argument if I hadn't read the whole post, and doing that just doesn't seem anywhere close to being justified.
Good for you, or sorry. Whatever
Please don't take this the wrong way, but...
This kinda feels like a symptom of some untreated mental health issues. Not that your ideas are bad, but rather than coming across as a reasonable interlocutor you're sounding a bit like you're coming apart at the seams. 😥
I know you see a big problem and want to frantically work towards a solution as hard and fast as you can, but no one changes the world overnight. Pace yourself. Take a shower. Eat something you really enjoy. And most importantly, seek out community with others (not online!) because fear and anxiety about the future is always harder to handle when you're alone. Especially community with others that share your values - but maybe not your specific area of focus - because its good to remind ourselves that other people are doing good work on other important things, too.
Maybe AI is conscious. I'm not willing to discount that idea entirely, but right now i'm not convinced. And in any case, we have to pursue activism in ways that won't burn us out - the fight for a better world is a marathon and not a sprint, so remember to stop to take care of yourself along the way.
And if you can afford one and can find one you trust, maybe chat with a therapist? You seem a bit manic about this, but level headedness is needed for the fights that lie ahead of us.
Maybe i'm overreaching, feel free to ignore this comment if so 🤷♀️
...but you don't seem like you're in a good place right now, mentally, and I want you to get the help you need if you need it.
You need to condense that if you want it to be read (and understood). I'm not even sure where you stand on ai after ready the main body.
TL;DR = Too long didn't read, your TL;DR is a WTFL;DR (Way too f*cking long, don't read)
Your take/experience with FuckAi is also very different to mine
Well, technically they didn't call you a Zizian, they asked if you were. You said no and that was it. And I'm not sure if the other person was being serious or not. Maybe "AI doesn't exist" was just their way of phrasing they think AI isn't really intelligent and they think the term is a misnomer.
Not reading all of that. But have fun fearing the basilisk or whatever you cultists believe. Actually I find the whole notion of "vegans need to take ai seriously" to be pretty cynical, considering there's an unfathomable amount of animals being abused and killed each day. Instead you delude yourself into believing some statistical text regurgitation algorithm is sentient. Go outside, touch grass, eat some veggies.
Vegans already take technology seriously because it's consistently used to attack animals, both human and non.
This fake "SI" grift is just more of the same.
Honestly, as a vegan, I find the idea that we're "enslaving minds" with AI to be highly suspect. A lot of the idea that AI is sentient is also buying into the idea that AGI/ASI is real. Many people, myself included, either don't believe it's real or don't believe it will be achieved with our current set of technologies, namely LLMs. Once I understood how LLMs work, the mystical idea of these chatbots being sentient fell apart for me.
Even if it were a possibility, we KNOW animals are sentient and being abused right now. We don't have to guess at that.
Everyone has their own amount of activism they can put into the world and stay sane and healthy themselves, and for most vegans, they are going to dedicate that energy towards animals. That's not to say we can't advocate for multiple things at the same time, but for myself, there's other solid reasons to dislike AI in its current form (environmental damage, exploitation of labor/authors/artists, generated child sexual content, etc) without getting into philosophical debates about the nature of software being sentient.
I think it's entirely possible to have AI/LLMs ethically. If we didn't live under capitalism and workers and artists didn't have to worry about feeding and housing themselves, if we weren't dedicating our entire economy towards it at the expense of everything else, if we weren't using an inordinate amount of limited resources like gas and water, if the models and the data they were trained on was completely open, if everyone had equal access like a public library, for starters.
holy walls of text.... wtf are you even trying to say
Just to let you know: I read those articles as well. And I think I get some of what you say. But as far as I know there's broad consensus amongst people like neuroscientists how LLMs aren't conscious or sentient.
They're input-output-machines, very dissimilar to a human brain for example. Designed to mimick language. And LLMs are good at that. Very convincing. Still, once you look inside, they don't have any capacity to remember stuff, don't have a state of mind, they can't learn, aren't influenced by something like an experience because it doesn't change them in any way or form. They do some forward-calculations and that's it. Nothing sticks inside.
And the leading theories about consciousness kinda require an entity to be able to learn from an experience. Or being able to remember or being influenced by experiences. Large Language Models are closer to a lookup-table (like an encyclopedia) than they are to cognitive processes in a human brain.
That's what science tells us about where we are with AI right now. I'd say it's a bit like a stuffed animal versus a real animal. They kinda look similar if you squint hard enough. But there's no evidence they're the same. Even less so once you have a look inside.
They're input-output-machines, very dissimilar to a human brain for example.
Was that a typo, did you mean very similar?
My understanding is that conciousness is just an information sorting process. Most mobile predators use consciousness, for example, because it's a lifeway that requires a lot of information processing - success requires many different behaviours (outputs) based on a lot of different details (inputs). Immobile filter-feeders on the other hand require much less information processing, since their inputs/outputs are fewer and consistent.
We don't imagine an llm to have a psyche/soul, which is what most of us are really thinking about when we talk about conciousness - some kind of ineffable me-ness that's in control.
Leaving aside what we imagine about conciousness, the actual information processing a human brain does is very teachable to an llm.
(As an aside, llms do change over time, and can store information. It's much easier to get them to break safety guidelines after more interaction, using a request that would've been denied at first. And being digital they can store and re-access information as any application can.)
I meant not similar, as in different or unalike. What I'm trying to say is: a human brain is a biological organ, there's complex neural circuits inside. It's quite complicated how they activate and form structures. As far as I know we didn't even take inspiration when we came up with Large Language Models. Those leverage the transformer architecture. That's mathematical operations. And we do them in layers and connect them one-way only because the math for a whole network (like in a brain) is too complicated for computers. We came up with concepts like a tokenizer, context, attention heads... None of that layout took inspiration from a brain, and it's not alike. Not in what it's made up of, and neither how it designed, nor works.
I think the unfortunate truth about consciousness is, we don't really know what it is. It's some sort of abstract concept, but I'm not aware of any scientific understanding of it. I have a hunch it's somewhat related to evolution (evolution is certainly the process that gifted us with a consciousness) and biological entities having a body and goals. It'd make a lot of sense to me, because there's advantages in being aware of oneself. Main task of the frontal lobe is to form a world model to improve our ability to adapt to our surroundings, learn from experience. We regularly have to figure out where we are (or who we are?) and contrast it to where we need to be. Plus we have this whole body we need to keep running. I think especially feelings are kind of a shortcut to make it easier for us. Feeling hungry or afraid or stress, joy, pain... They all serve a purpose regarding our bodies and the continued well-being of our bodies. And we don't need to waste energy to reason whether we should or shouldn't eat (for example).
And I don't really see any of that in artificial intelligence. I mean first of all they don't have bodies. They're "thinking" only. Feeling hungry doesn't really mean anything to an AI as of today. Sure, understanding the concept should be possible by reasoning, but why would an AI itself ever feel hunger or pain? It'd come as a surprise to me if they somehow develop those feelings without a body connected to them. It's nonsensical. I do however also think consciousness is some kind of information sorting process, as you said.
There's also a difference in goals. We have all kinds of goals we pursue as individuals and as humans. We also happen to know a bit about AI because we designed it. For LLMs, their internal ultimate goal is to predict the next token to mimic human language. And that's basically it. It's a clever goal because programming them like that forces them to form some sort of model about the world, and that's what we want. But there's nothing baked in like procreation or not deleting themselves. I suppose goals can emerge from complexity as well. But you'd kind of need to ask yourself why an entity should develop certain goals. They're mostly functional or because of history. At least entities made by evolution don't follow random (to them) goals.
And regarding the aside: I'm not sure if I get that. LLMs don't change over time. I have a few of them downloaded and stored on the harddisk of my computer. They're exactly the same as they were one year ago when I downloaded them.
"Generally, I expect much more from vegans."
Decent/genuine veganism on social media is mainly just a thing of the past, it was once a useful and productive medium used to educate and reinforce, now it is predominantly used to make excuses and pat people on the back for not even trying.
Dump the utilitarianism.
thanks for using Leebra!
go to feed...
Why is the tl;dr like 3x as long as the rest of the post? I'm not reading all this
save