Companies want to profit from it without spending much
Startups, major companies, and venture capital investment groups implement new tech services quickly to capitalize on new trend
Stonks go up
Users coalesce around some major options that have actual ideas outside of “let’s start a [new technology] company,” most others fail to innovate or offer actual value that meets the wild expectations of speculative investors
Stonks go down, people lose their jobs, tech divisions crumble
Who could have foreseen this!? (See also: blockchain, NFTs, etc.)
Who would be Enron in this scenario? I’m thinking of openai or Nvidia but can’t imagine they’re at risk of much. The game seems far more rigged in favor of corpos than 20 yrs ago. Can’t imagine how they could ever fall off at this point. Or, say a company like Netflix which I’m is using tons of algorithms…
Wait a min. Isn’t AI now used as a catch all for any and all tech laymen don’t understand? Do you specifically mean llamas are being overhyped (had to leave the autocorrect version of LLMs)?
Sorry, if I’m slow to the anti-AI party. It’s been hard figuring out what’s sensationalism, what’s contrarianism, and what’s legit nowadays (or maybe I’m just getting dumber :D)
The companies aren’t really at stake here, it’s the human cost and the devaluing of labour that’s my personal issue.
I think it also makes it worse that people are being laid off to pay for “AI” (quotes because LLMs and ML in general are not actually real AI as understood by computer science, so yes I’m referring to LLMs here) technologies that literally require human intervention to function in the way they’re advertised to.
A lot of these models output is tweaked by people in India (for example) that get paid shit wages to make sure that it didn’t spit out absolute useless gibberish or straight up wrong information. The low wages don’t really incentivize the worker to give a fuck, so at every level you’re basically going for the minimum viable product (MVP).
Not a great foundation for the future of technology, in my opinion. These things will have to be actually useful in a practical way to not simply be a fad. Also, art theft and plagiarism are pretty real issues that are inherent to the implementation of the technology unless you only train a model on your own content, which a lot of people/companies straight up refuse to do.
To refer back to another example of tech that was overhyped, blockchain technology has like, 1 real-life use case, and that’s really cool, but it’s not really an earth-shattering development despite every tech bro proselytizing about it like the second coming. Laying off productive workers to focus on it was a bad move in hindsight, an obvious one. I think that will apply here too.
I was actually just hearing about this Karla Ortiz copyright case. Sounds like it’s going to shine a big ol spotlight on how these systems work and how exactly they copy other people’s work.
I wonder if it could make things so prohibitively expensive it’s no longer worth investing in for most companies. At the very least, it hopefully gives artists/content creators some kind of legal protection
I’m hoping the trend’ll mostly burn itself out while people figure out how to scale due to the business model’s inherent lack of sustainability and rising energy costs, but if it doesn’t the least we can do is help artists defend what little they do have.
I really fear for 2D artists and indie music producers because 2D art and instrumental music seem like the easiest media to rip off without credit or compensation. Even 3D artists might not be safe depending on how good these models can get at creating 3D meshes that don’t look like Eldritch homunculi.
https://en.m.wikipedia.org/wiki/Dot-com_bubble
Oh, look! Relevant history! (Hint: companies don’t care about actual value if they think there’s easy money to be made.)
Uhh… what?
Could you pretend I don’t already know what you mean and explain what you’re trying to communicate a bit?
Who could have foreseen this!? (See also: blockchain, NFTs, etc.)
Thanks. That makes sense
Who would be Enron in this scenario? I’m thinking of openai or Nvidia but can’t imagine they’re at risk of much. The game seems far more rigged in favor of corpos than 20 yrs ago. Can’t imagine how they could ever fall off at this point. Or, say a company like Netflix which I’m is using tons of algorithms…
Wait a min. Isn’t AI now used as a catch all for any and all tech laymen don’t understand? Do you specifically mean llamas are being overhyped (had to leave the autocorrect version of LLMs)?
Sorry, if I’m slow to the anti-AI party. It’s been hard figuring out what’s sensationalism, what’s contrarianism, and what’s legit nowadays (or maybe I’m just getting dumber :D)
The companies aren’t really at stake here, it’s the human cost and the devaluing of labour that’s my personal issue.
I think it also makes it worse that people are being laid off to pay for “AI” (quotes because LLMs and ML in general are not actually real AI as understood by computer science, so yes I’m referring to LLMs here) technologies that literally require human intervention to function in the way they’re advertised to.
A lot of these models output is tweaked by people in India (for example) that get paid shit wages to make sure that it didn’t spit out absolute useless gibberish or straight up wrong information. The low wages don’t really incentivize the worker to give a fuck, so at every level you’re basically going for the minimum viable product (MVP).
Not a great foundation for the future of technology, in my opinion. These things will have to be actually useful in a practical way to not simply be a fad. Also, art theft and plagiarism are pretty real issues that are inherent to the implementation of the technology unless you only train a model on your own content, which a lot of people/companies straight up refuse to do.
To refer back to another example of tech that was overhyped, blockchain technology has like, 1 real-life use case, and that’s really cool, but it’s not really an earth-shattering development despite every tech bro proselytizing about it like the second coming. Laying off productive workers to focus on it was a bad move in hindsight, an obvious one. I think that will apply here too.
You’re makin great points!
I was actually just hearing about this Karla Ortiz copyright case. Sounds like it’s going to shine a big ol spotlight on how these systems work and how exactly they copy other people’s work.
I wonder if it could make things so prohibitively expensive it’s no longer worth investing in for most companies. At the very least, it hopefully gives artists/content creators some kind of legal protection
I’m hoping the trend’ll mostly burn itself out while people figure out how to scale due to the business model’s inherent lack of sustainability and rising energy costs, but if it doesn’t the least we can do is help artists defend what little they do have.
I really fear for 2D artists and indie music producers because 2D art and instrumental music seem like the easiest media to rip off without credit or compensation. Even 3D artists might not be safe depending on how good these models can get at creating 3D meshes that don’t look like Eldritch homunculi.