Playing Catch-Up (also a rant on data centres)
You know, it feels like I should have my own Members blog just to write this stuff...but I don't wanna. If I ever did, it'll probably exist for Marble related projects.
That aside, I've been doing pretty alright, though I am a bit worried about my current job. But Im not one to talk about my personal life indepth because I still clamour onto my privacy, also because who cares? "Fuuuck you might be at a risk of a lay-off, you're a failure to your family or whatever. I LIVE IN EASTERN EUROPE DUDE!"
I am working on something Marble related though, it's nothing complicated though, just a graphic thing. Infact, you might've seen one of them already. I've been chipping at it whenever I got the time, it's safe to say that I'm pretty close to finishing the entire line-up. Stay tuned for that...whenever I get around to releasing it.
Just like the OpenMBU server, I gatekeep hard. I don't wanna show my hand on planned threads or projects:
There's an easy solution to this right? Just ask the robot to do the research for you. If that was your first thought, then you should leave this thread and kill yourself.
I wanna work for my info, that's what makes research fun. Yeah it's frustrating to find constant dead-ends, but who said that shit was easy? All this skeptism about the world made people forget researching and writing citations is worthy enough to be a job. If anyone could do this shit, we would've figured out every abstract theory that ever existed, hell maybe we'll figure out if infinity is possible!
Despite my arrogance, Im not entirely against robots. They work best as tools rather than a magic wand, more to leviate reputation and tedium rather than do the whole job for you. I wanna own the best of computing, not rent one part of it for a monthly payment. I don't count "finding inspiration" as a tool though, if you want a prompt for writing, drawing, or just creating in general, then maybe try engaging with your interests more? There are billions of art that exist in many forms throughout our whole existence for you to consume, you won't go through all of it in one life, so just start somewhere!
If you don't know where to start, just think over what your favourite form of art is, it could be music, books, movies. Look over the arts made during your childhood, your parent's childhood, or just a period of time that interests you. Look into what arts your own country makes, or search for something unique outside. Skim through your favourite genres, or discover defunct ones like westerns. It's fine if you wanna catch up on the popular shit, everyone's gotta watch The Sopranos at some point in their life, but you also gotta try and look for the obscure and infamous too.
How I approach art is by whatever catches my eyes, I know the kinda genres I like, Im pretty open to whatever region of the world it was made in, seeing how other cultures create art interests me. That's if I know what kinda art Im looking for. What I often do is snatch whatever interests me by the box old school style and go into it blind, it might evoke an impression on me, or I come out of it feeling like I wasted my time, eitherway it's fine! Sometimes you gotta take the good with the bad, truthfully, seeing wasted potential could inspire you to seek a new take by another author, or just create that potential yourself.
I'd be more trustful of LLMs if they did what they say on the can, yknow shit like "figure out the cure to cancer" or "read through every research paper" or "prove impossible mathematical theories," but that's not a good sell. Why limit yourself to a research team when you can just make it all-purpose? Why read from a small library when you can just constantly scrape one fraction of the internet instead?
Now look where that got us, here we are summerising documents because reading two paragraphs is hard for us now. If we're not paranoid enough about imaginary scenarios, we can just generate one just to get mad at it. Our dependence on this shit is killing critical thinking skills. It gotten so bad that it's starting to suck up its own reguritation, and they're doing all that at a loss anyways.
All the while dumbasses like you who are still reading and haven't killed yourselves seven paragraphs ago will parrot shit like "you're standing in the way of progress, just like mathematicians did with calculators." while in the same breath bitch about surveilence and computers parts. Not to claim that the factory workers who destroyed computers back then were in the right, their action did stand in the way of progress, but those two periods shouldn't be compared.
I trust that you're smart enough to know the difference between the breakthroughs of computers and LLMs as a product. But just to get one point across, computers has gotten smaller since their invention, they don't take up a room anymore because clever motherfuckers figured out a way to compute shit with smaller shits.
Meanwhile, LLMs are using the tride and true method of building the same shit a buncha times instead of optimising what they got to work better. It makes sense why, our current computing power can't scrape a massive library, let alone the entire internet, without it taking decades. So instead of putting resources into research (we're operating at a loss afterall), lets just ruin your neighbourhood instead, we'll also pass all the expenses on you by the way, unless you decide to fight back against it.
Can you imagine the world we'd be in if computers led by the LLM practice? I'd be writing all this in my neighbourhood's designated computer room that I rented for the hour, sweating my ass off from all the processing machines heating the room up.
Data centres rolls together the worst practices in the industry, a lacking care for computing optimisation, invasion of privacy, generalised demographics, and the subscription model. These four are common enough practices that warrant their own discussion, but they all points to a greater issue in our economy. But that's when we start bordering on politics and Im not gonna be the one to break our only means of "escapism", so let's stop here before you get cagey on behalf of your favourite billionaire.
Could there be a better way to create these tools? Definitely! Though, the only way we could create a good LLM is by starting from scratch, pivot to research instead of constantly making shit up, maybe have competent people running these projects instead of doomsday cultists. Will it be perfect? Not really, a software can only be as smart as its creator(s), but it'll be better than whatever dystopian shit we got going on right now.
That aside, I've been doing pretty alright, though I am a bit worried about my current job. But Im not one to talk about my personal life indepth because I still clamour onto my privacy, also because who cares? "Fuuuck you might be at a risk of a lay-off, you're a failure to your family or whatever. I LIVE IN EASTERN EUROPE DUDE!"
I am working on something Marble related though, it's nothing complicated though, just a graphic thing. Infact, you might've seen one of them already. I've been chipping at it whenever I got the time, it's safe to say that I'm pretty close to finishing the entire line-up. Stay tuned for that...whenever I get around to releasing it.
Just like the OpenMBU server, I gatekeep hard. I don't wanna show my hand on planned threads or projects:
- One, because I'll fall into this trap where Im locked to the announced project, a common trap most mod creators are caught by, anyone's wondering what happened to Marble Blast Fusion? I certainly don't. Hope they're still alive tho.
- Two, I wanna do this shit with you guys, discuss and create together, we're a community no? Usually I'd hear about some interesting shit that going on, only to find out that it's exclusively happening in the off-shoots. That shit's deflating, its not infuriating though, I don't get this pressed over the mere mention of off-shoots like Kalle does. Not saying specific names, this could apply to you for all I care, but I feel like some of you got infamy ego, like you're entitled to recognition because you were vilified in the past.
- Three, because Im a retard. A good chunk of my drafts start off as hypothesis, it requires some level of research and learning new skills. Im not a know-it-all, even though I write like one. Im extremely doubtful about what I know and I use that energy to search for answers, either through discussions or just combing through the internet.
There's an easy solution to this right? Just ask the robot to do the research for you. If that was your first thought, then you should leave this thread and kill yourself.
I wanna work for my info, that's what makes research fun. Yeah it's frustrating to find constant dead-ends, but who said that shit was easy? All this skeptism about the world made people forget researching and writing citations is worthy enough to be a job. If anyone could do this shit, we would've figured out every abstract theory that ever existed, hell maybe we'll figure out if infinity is possible!
Despite my arrogance, Im not entirely against robots. They work best as tools rather than a magic wand, more to leviate reputation and tedium rather than do the whole job for you. I wanna own the best of computing, not rent one part of it for a monthly payment. I don't count "finding inspiration" as a tool though, if you want a prompt for writing, drawing, or just creating in general, then maybe try engaging with your interests more? There are billions of art that exist in many forms throughout our whole existence for you to consume, you won't go through all of it in one life, so just start somewhere!
If you don't know where to start, just think over what your favourite form of art is, it could be music, books, movies. Look over the arts made during your childhood, your parent's childhood, or just a period of time that interests you. Look into what arts your own country makes, or search for something unique outside. Skim through your favourite genres, or discover defunct ones like westerns. It's fine if you wanna catch up on the popular shit, everyone's gotta watch The Sopranos at some point in their life, but you also gotta try and look for the obscure and infamous too.
How I approach art is by whatever catches my eyes, I know the kinda genres I like, Im pretty open to whatever region of the world it was made in, seeing how other cultures create art interests me. That's if I know what kinda art Im looking for. What I often do is snatch whatever interests me by the box old school style and go into it blind, it might evoke an impression on me, or I come out of it feeling like I wasted my time, eitherway it's fine! Sometimes you gotta take the good with the bad, truthfully, seeing wasted potential could inspire you to seek a new take by another author, or just create that potential yourself.
I'd be more trustful of LLMs if they did what they say on the can, yknow shit like "figure out the cure to cancer" or "read through every research paper" or "prove impossible mathematical theories," but that's not a good sell. Why limit yourself to a research team when you can just make it all-purpose? Why read from a small library when you can just constantly scrape one fraction of the internet instead?
Now look where that got us, here we are summerising documents because reading two paragraphs is hard for us now. If we're not paranoid enough about imaginary scenarios, we can just generate one just to get mad at it. Our dependence on this shit is killing critical thinking skills. It gotten so bad that it's starting to suck up its own reguritation, and they're doing all that at a loss anyways.
All the while dumbasses like you who are still reading and haven't killed yourselves seven paragraphs ago will parrot shit like "you're standing in the way of progress, just like mathematicians did with calculators." while in the same breath bitch about surveilence and computers parts. Not to claim that the factory workers who destroyed computers back then were in the right, their action did stand in the way of progress, but those two periods shouldn't be compared.
I trust that you're smart enough to know the difference between the breakthroughs of computers and LLMs as a product. But just to get one point across, computers has gotten smaller since their invention, they don't take up a room anymore because clever motherfuckers figured out a way to compute shit with smaller shits.
Meanwhile, LLMs are using the tride and true method of building the same shit a buncha times instead of optimising what they got to work better. It makes sense why, our current computing power can't scrape a massive library, let alone the entire internet, without it taking decades. So instead of putting resources into research (we're operating at a loss afterall), lets just ruin your neighbourhood instead, we'll also pass all the expenses on you by the way, unless you decide to fight back against it.
Can you imagine the world we'd be in if computers led by the LLM practice? I'd be writing all this in my neighbourhood's designated computer room that I rented for the hour, sweating my ass off from all the processing machines heating the room up.
Data centres rolls together the worst practices in the industry, a lacking care for computing optimisation, invasion of privacy, generalised demographics, and the subscription model. These four are common enough practices that warrant their own discussion, but they all points to a greater issue in our economy. But that's when we start bordering on politics and Im not gonna be the one to break our only means of "escapism", so let's stop here before you get cagey on behalf of your favourite billionaire.
Could there be a better way to create these tools? Definitely! Though, the only way we could create a good LLM is by starting from scratch, pivot to research instead of constantly making shit up, maybe have competent people running these projects instead of doomsday cultists. Will it be perfect? Not really, a software can only be as smart as its creator(s), but it'll be better than whatever dystopian shit we got going on right now.
If you can read this, fuck you!
