This. My company pushes AI use (garbage) and then a month or so ago the provider went "oopsy-doodle, your 30k/month bill will now be 3 mil / month, kthanksbye"
This is why I'm in favor of NOT EVEN pirating those programs. The first step to breaking their monopoly is to learn how to use alternatives. Gimp is pretty popular, but I've been using PaintDotNet for a lot of things in the past few years, and for 90% of what I want to do it's way easier to use than Gimp.
Disagree completely. Local models are the absolute pinnacle of AI technology as it currently exists...and it is mid. All non-local platforms are objectively propaganda pushed to manufacture a new reality that exhaults the owners of said platform i.e. wealthy shareholders foaming at mouth for a slave populace.
At first I completely disagreed with you (actually, we both agreed, which is why I disagreed with you? Since I didn't disagree with anything you said) but I've since changed my mind about local models. Having gotten a dual gpu setup working, I can run the 70b models very well now, and they are actually impressive.... I mean, compared to the horseshit propaganda gemini/claude bs I mean.
So...you're ok with depriving artists and creators of their means of living just to be used by massive corporation so they can capitalise on something built with stolen labour? Weird, that seems far more fascistic than allowing people to earn an income from their work... Or are you one of those who think art is owned by mankind as a whole and should be free, but doesn't realise we're still in a system where basic needs aren't assured so artist have to sell their work to be able to live?
That's a too generalized view point there. While I also think, that generative LLMs are flawed to a degree that makes them basically unfit for anything right now, analytical and transformative LLMs do have their uses and can be very helpful. The more specialized the task the better it usually works. But this whole nonsense of a generalized AI, that they literally try to sell us, is just unwarranted.
Yep, give a purpose built AI data to process like a categorized state of repair based on images for traffic signs and it can do TONS of work and fairly well. But asking it if a mole looks cancerous is what dermatologists are for, not AI
Even then can it be used to support the effort. It's also one of the problems with this whole notion of generalized AI: That the human as no task any longer. Just lay back and let the computer do the work. That's usually not gonna fly. It's a tool. Even when it comes to identifying cancer, then it may very well work, and work well. But that does not mean that the person wielding this tool is all of a sudden not required to be thorough or even make some effort of verifying the results.
Just imagine it working like this: AI looks at the scan results and points out the places where it assessed there might be something. Could give you an ahead start, especially when the false positives are low. Now you can look more specifically. If there is indeed something, you have just speed up the process considerably without having lost the prudence. A more thorough check may still be in order. It does not mean that you do not take a second look, regardless of what that thing says.
UserMCP
AI slop is slop, no matter the message. They're trying to encourage us to use it as bait, while manufactured reality is the hook.
narcoleptiKnickers
Free AI?
HeresYourSauce
The idea is to get people hooked. They're viewing it like ai is drugs or something.
Might even work for some people. They'll have found specific tasks they can automate with AI and not want to change back.
Most of us won't fall for it, though, and ask soon as they push AI will collapse into a fraction of what it is now.
cxmpxsxd
There are a bunch of words above that I've seen before but never in this order... you're saying to put all of my money in NFTs?
mrthewhitee
Save a little for crypto coins.
Hexidimentional
ok but why would i want it?
LockedAndKnotted
So they're destroying the planet for free! Nice!
Th1ng
yeah but how much vram do I need to run one of theirs? 100GB? 10?
MattScifiBandit
It doesnt matter, cant afford the GPUs to run the AI because of the AI
GratuaCuun
It reads like they want people to train their AI systems for them through free use
CitrusyGarlic
Do you have to understand code to get this or am I just dumb?
bad1080
hook em for free, then make em pay after they can't live/work without it
darthstormer
This. My company pushes AI use (garbage) and then a month or so ago the provider went "oopsy-doodle, your 30k/month bill will now be 3 mil / month, kthanksbye"
theworldcouldbeflat
Like they did with Photoshop and Office.
UNHchabo
This is why I'm in favor of NOT EVEN pirating those programs. The first step to breaking their monopoly is to learn how to use alternatives.
Gimp is pretty popular, but I've been using PaintDotNet for a lot of things in the past few years, and for 90% of what I want to do it's way easier to use than Gimp.
Arcygenical
Local models are all kinda mid, but this is still nice. These are all available on other platforms already, though.
MundySeme
Disagree completely. Local models are the absolute pinnacle of AI technology as it currently exists...and it is mid. All non-local platforms are objectively propaganda pushed to manufacture a new reality that exhaults the owners of said platform i.e. wealthy shareholders foaming at mouth for a slave populace.
Arcygenical
At first I completely disagreed with you (actually, we both agreed, which is why I disagreed with you? Since I didn't disagree with anything you said) but I've since changed my mind about local models. Having gotten a dual gpu setup working, I can run the 70b models very well now, and they are actually impressive.... I mean, compared to the horseshit propaganda gemini/claude bs I mean.
lightfoot2
Yeah, if it's non-local then it's basically a mechanical turk going to the Google search engine. What's the point?
Pilgrymm
RowanUnderwood
You guys have really swallowed the fascist "intellectual property" nonsense hook line and sinker haven't you :*( Internalized propaganda . . .
JStengah
Nah, just trying to survive in the system they live in.
BackToTheOriginalUsername
Nevermind, I looked at your comment, you're an hypocritical AI users, calling fascim everywhere where you're voluntarily using one of its tool.
BackToTheOriginalUsername
So...you're ok with depriving artists and creators of their means of living just to be used by massive corporation so they can capitalise on something built with stolen labour?
Weird, that seems far more fascistic than allowing people to earn an income from their work...
Or are you one of those who think art is owned by mankind as a whole and should be free, but doesn't realise we're still in a system where basic needs aren't assured so artist have to sell their work to be able to live?
Pilgrymm
Look at you, above it all. Thumb up,
Sotai
They might not make money, but at least they also aren't good at anything.
bippityboppitybuttsex
They are key to enshittification
alt86er
That's a too generalized view point there. While I also think, that generative LLMs are flawed to a degree that makes them basically unfit for anything right now, analytical and transformative LLMs do have their uses and can be very helpful. The more specialized the task the better it usually works. But this whole nonsense of a generalized AI, that they literally try to sell us, is just unwarranted.
Sebastopol140
THIS!!
Flyingwigs
Yep, give a purpose built AI data to process like a categorized state of repair based on images for traffic signs and it can do TONS of work and fairly well. But asking it if a mole looks cancerous is what dermatologists are for, not AI
alt86er
Even then can it be used to support the effort. It's also one of the problems with this whole notion of generalized AI: That the human as no task any longer. Just lay back and let the computer do the work. That's usually not gonna fly. It's a tool. Even when it comes to identifying cancer, then it may very well work, and work well. But that does not mean that the person wielding this tool is all of a sudden not required to be thorough or even make some effort of verifying the results.
alt86er
Just imagine it working like this: AI looks at the scan results and points out the places where it assessed there might be something. Could give you an ahead start, especially when the false positives are low. Now you can look more specifically. If there is indeed something, you have just speed up the process considerably without having lost the prudence. A more thorough check may still be in order. It does not mean that you do not take a second look, regardless of what that thing says.