I have the same feeling about Kojima’s and Vincke’s latest comments on AI. Am I supposed to get mad at every single person who said they used/plan to use AI for something? I’d be as outraged as the average Fox News viewer, and it would be impossible to be taken seriously. I still won’t be using AI myself (fuck surveillance state AI) and I’d be making every effort to encourage others not to use it, but there’s no point in burning bridges and falling for rage bait.
They’re creative people who care about the craft and care about the teams in their employ, which gives their statements weight, where some Sony/Microsoft/EA executive making an identical statement has none.
Give it another 5 years maybe and local self-trainable models and alternative versions of it will be available that won’t have all the theft problems, surveillance problems and other issues. The tech is new and mainly controlled by giant companies right now.
It’s not like the tech is going to forever exist in a vacuum in the exact state. It’s in nothing ever does. Makes it doubly silly to get mad over a tool.
I understand the principle. Even if E33 is not slop, people should fear a road that leads to dependence on “surveillance state AI” like OpenAI. That’s unacceptable.
That being said, I think a lot of people don’t realize how commoditized it’s getting. “AI” is not a monoculture, it’s not transcending to replace people, and it’s not limited to corporate APIs. This stuff is racing to the bottom to become a set of dumb tools, and dirt cheap. TBH that’s something that makes a lot of sense for a game studio lead to want.
And E33 is clearly not part of the “Tech Bro Evangalism” camp. They made a few textures, with a tool.
When I give myself the leeway to think of a less hardliner stance on AI, I come back to Joel Haver’s video on his use of ebsynth:
It lets me create rotoscoped animations alone, which is something I never would have the time or patience for otherwise. Any time technology makes art easier to learn, more accessible, we should applaud it. Art should be in the hands of everyone.
Now my blood boils like everyone else’s when it comes to being forced to use AI at work, or when I hear the AI Voice on Youtube, or the forced AI updates to Windows and VS Code, but it doesn’t boil for Joel. He clearly has developed an iconic style for his comedy skits, and puts effort into those skits long before he puts it through an AI rotoscope filter. He chose his tool and he uses it sparingly. The same was apparently true for E33, and I have no reason not give Kojima and Larian the same benefit of the doubt.
On the other hand, Joel probably has no idea what I’m talking about when I say “surveillance state AI.” People Make Games has a pretty good video exposing its use case. There’s also…
the global and localized environmental impacts of all these data centers,
Nvidia and Micron pricing the consumer out of owning their own hardware,
aforementioned companies fraudulently inflating an economic bubble,
the ease with which larger models can be warped to suit their owners’ fascist agendas (see Grok).
Creatives may be aware of some, or all, or none of those things, which is why it’s important to continue raising awareness of them. AI may be toothpaste that can’t go back in the tube, but it’s also a sunk cost fallacy, you don’t have to brush your teeth with shit-flavored toothpaste.
Now my blood boils like everyone else’s when it comes to being forced to use AI at work, or when I hear the AI Voice on Youtube, or the forced AI updates to Windows and VS Code
You don’t hate AI. You hate Big Tech Evangelism. You hate corporate enshittification, AI oligarchs, and the death of the internet being shoved down your throat.
…I think people get way too focused on the tool, and not these awful entries wielding them while conning everyone. They’re the responsible party.
You’re using “AI” as a synonym for OpenAI, basically, but that’s not Joel Haver’s rotoscope filter at all. That’s niche machine learning.
As for the exponential cost, that’s another con. Sam Altman just wants people to give him money.
Look up what it takes to train (say) Z Image or GLM 4.6. It’s peanuts, and gets cheaper every month. And eventually everyone will realize this is all a race to the bottom, not the top… but it’s talking a little while :/
True on most fronts except one. On a personal level, I do hate AI lol. The large language model itself. I just don’t think typing out or speaking out a series of instructions is that useful or efficient. If I want a computer to do something for me, I much prefer the more rigid and unnatural syntax and grammar of programming language. AI tools themselves just don’t produce a result that satisfies me.
Don’t produce a result that’s satisfies you yet. Early programming also was absolute dog s***.
Give it 20 years, and there’s bound to be new things that will replace the current concept of AI that do functionally the same thing just in a manner that actually does produce good results.
Just like we did with everything else computing related.
Hating a tool is the single stupidest f****** thing anyone can do.
That and chat prompting engineer b******* is one tiny tiny slice of the greater hole. It’s a footnote in the grand scheme of everything that the colloquial term AI represents. It’s just the most marketable one to end users so it’s the one that you see everywhere.
Again, they’re tools. Some of the most useful applications for LLMs I’ve worked on are never even seen by human eyes, like ranking, then ingesting documents and filling out json in pipelines. Or as automated testers.
Another is augmented diffusion. You can do crazy things with depth maps, areas, segmentation, mixed with hand sketching to “prompt” diffusion models without a single typed word. Or you can use them for touching up something hand painted, spot by spot.
You just need to put everything you’ve ever seen with ChatGPT and copilot and the NotebookLM YouTube spam out of your head. Banging text into a box and “prompt engineering” is not AI. Chat tuned decoder-only LLMs are just one tiny slice that a few Tech Bros turned into a pyramid scheme.
I have the same feeling about Kojima’s and Vincke’s latest comments on AI. Am I supposed to get mad at every single person who said they used/plan to use AI for something? I’d be as outraged as the average Fox News viewer, and it would be impossible to be taken seriously. I still won’t be using AI myself (fuck surveillance state AI) and I’d be making every effort to encourage others not to use it, but there’s no point in burning bridges and falling for rage bait.
They’re creative people who care about the craft and care about the teams in their employ, which gives their statements weight, where some Sony/Microsoft/EA executive making an identical statement has none.
Give it another 5 years maybe and local self-trainable models and alternative versions of it will be available that won’t have all the theft problems, surveillance problems and other issues. The tech is new and mainly controlled by giant companies right now.
It’s not like the tech is going to forever exist in a vacuum in the exact state. It’s in nothing ever does. Makes it doubly silly to get mad over a tool.
I understand the principle. Even if E33 is not slop, people should fear a road that leads to dependence on “surveillance state AI” like OpenAI. That’s unacceptable.
That being said, I think a lot of people don’t realize how commoditized it’s getting. “AI” is not a monoculture, it’s not transcending to replace people, and it’s not limited to corporate APIs. This stuff is racing to the bottom to become a set of dumb tools, and dirt cheap. TBH that’s something that makes a lot of sense for a game studio lead to want.
And E33 is clearly not part of the “Tech Bro Evangalism” camp. They made a few textures, with a tool.
When I give myself the leeway to think of a less hardliner stance on AI, I come back to Joel Haver’s video on his use of ebsynth:
Now my blood boils like everyone else’s when it comes to being forced to use AI at work, or when I hear the AI Voice on Youtube, or the forced AI updates to Windows and VS Code, but it doesn’t boil for Joel. He clearly has developed an iconic style for his comedy skits, and puts effort into those skits long before he puts it through an AI rotoscope filter. He chose his tool and he uses it sparingly. The same was apparently true for E33, and I have no reason not give Kojima and Larian the same benefit of the doubt.
On the other hand, Joel probably has no idea what I’m talking about when I say “surveillance state AI.” People Make Games has a pretty good video exposing its use case. There’s also…
Creatives may be aware of some, or all, or none of those things, which is why it’s important to continue raising awareness of them. AI may be toothpaste that can’t go back in the tube, but it’s also a sunk cost fallacy, you don’t have to brush your teeth with shit-flavored toothpaste.
You don’t hate AI. You hate Big Tech Evangelism. You hate corporate enshittification, AI oligarchs, and the death of the internet being shoved down your throat.
…I think people get way too focused on the tool, and not these awful entries wielding them while conning everyone. They’re the responsible party.
You’re using “AI” as a synonym for OpenAI, basically, but that’s not Joel Haver’s rotoscope filter at all. That’s niche machine learning.
As for the exponential cost, that’s another con. Sam Altman just wants people to give him money.
Look up what it takes to train (say) Z Image or GLM 4.6. It’s peanuts, and gets cheaper every month. And eventually everyone will realize this is all a race to the bottom, not the top… but it’s talking a little while :/
True on most fronts except one. On a personal level, I do hate AI lol. The large language model itself. I just don’t think typing out or speaking out a series of instructions is that useful or efficient. If I want a computer to do something for me, I much prefer the more rigid and unnatural syntax and grammar of programming language. AI tools themselves just don’t produce a result that satisfies me.
Don’t produce a result that’s satisfies you yet. Early programming also was absolute dog s***.
Give it 20 years, and there’s bound to be new things that will replace the current concept of AI that do functionally the same thing just in a manner that actually does produce good results.
Just like we did with everything else computing related.
Hating a tool is the single stupidest f****** thing anyone can do.
That and chat prompting engineer b******* is one tiny tiny slice of the greater hole. It’s a footnote in the grand scheme of everything that the colloquial term AI represents. It’s just the most marketable one to end users so it’s the one that you see everywhere.
Again, they’re tools. Some of the most useful applications for LLMs I’ve worked on are never even seen by human eyes, like ranking, then ingesting documents and filling out json in pipelines. Or as automated testers.
Another is augmented diffusion. You can do crazy things with depth maps, areas, segmentation, mixed with hand sketching to “prompt” diffusion models without a single typed word. Or you can use them for touching up something hand painted, spot by spot.
You just need to put everything you’ve ever seen with ChatGPT and copilot and the NotebookLM YouTube spam out of your head. Banging text into a box and “prompt engineering” is not AI. Chat tuned decoder-only LLMs are just one tiny slice that a few Tech Bros turned into a pyramid scheme.