99.9% of emails are spam or at the very least unsolicited. 99.9% of AI content is useless, wrong or disturbing, thus, slop. And both these numbers are generous estimates, it’s probably closer to 99.99% slop and spam.
I don’t think people are calling all AI slop. But if most actual use of AI is creating slop, people will associate the two. Also, it shows the technology isn’t as groundbreaking or universally applicable as advertised…
Some people are unambiguously calling all AI slop.
Advertising shits in your brain, but we can’t let the obvious lies of obvious liars obscure how neural networks are a whole new kind of software. Even these stupid chatbots let anyone write the old kinds of software. There’s a guy on Youtube who built a camera to visualize the speed of a laser in flight. Halfway through the video about integrating a hilariously sensitive photodiode and a high-precision motor system, he completely hand-waves the code for everything you actually see.
None of this is going to disappear when the bubble bursts. The dotcom bubble didn’t kill the web. We’ve demonstrated that a DVD’s worth of linear algebra can turn plain English into amateur Python, and suggesting that will soon be lost to history is absurd. Your IDE’s gonna have an autocomplete where Clippy does what you fucking tell him. It’s just not going to be the as-a-service remote computing bullshit these vultures are betting on, because remote computing has been a stupid idea for at least half a century. Spicy autocomplete will be another tool in the menu… like normal autocomplete. We can sneer at people for using either one, but rough standards and working code move the world.
Diffusion almost lives up to its hype. It is CGI for dummies, and will produce photorealistic video with less effort than hand-drawing a stick figure. LLMs might disappear entirely, with every aspect of their design replaced, but ‘remove all the pixels that don’t look like Hatsune Miku impregnating Goku’ actually works. God help us all.
The practical future is in “diffusion forcing,” where a human artist can draw however many frames they like, and the robot can only fill in the gaps. If the robot does something wrong… draw more frames between stuff. One frame per second and a say-what-you-see description will probably suffice.
We can presumably also expect variations that finish sketchy animatics, but that’s always going to be less art-driven than artists would like. Absolute maniacs like James Baxter can feed in a pencil version of a camera orbiting a ballroom dance, but the robot will emit a broadly similar motion in finished quality. It’s better-off being used to turn on-fours into on-ones.
Calling all AI use slop is like calling all e-mail spam.
99.9% of emails are spam or at the very least unsolicited. 99.9% of AI content is useless, wrong or disturbing, thus, slop. And both these numbers are generous estimates, it’s probably closer to 99.99% slop and spam.
Right, the vast majority by volume.
Yet you would not say ‘my mom sent me spam of Christmas photos.’
I don’t think people are calling all AI slop. But if most actual use of AI is creating slop, people will associate the two. Also, it shows the technology isn’t as groundbreaking or universally applicable as advertised…
Some people are unambiguously calling all AI slop.
Advertising shits in your brain, but we can’t let the obvious lies of obvious liars obscure how neural networks are a whole new kind of software. Even these stupid chatbots let anyone write the old kinds of software. There’s a guy on Youtube who built a camera to visualize the speed of a laser in flight. Halfway through the video about integrating a hilariously sensitive photodiode and a high-precision motor system, he completely hand-waves the code for everything you actually see.
None of this is going to disappear when the bubble bursts. The dotcom bubble didn’t kill the web. We’ve demonstrated that a DVD’s worth of linear algebra can turn plain English into amateur Python, and suggesting that will soon be lost to history is absurd. Your IDE’s gonna have an autocomplete where Clippy does what you fucking tell him. It’s just not going to be the as-a-service remote computing bullshit these vultures are betting on, because remote computing has been a stupid idea for at least half a century. Spicy autocomplete will be another tool in the menu… like normal autocomplete. We can sneer at people for using either one, but rough standards and working code move the world.
What’s the most concrete use case for AI that justifies its existence?
Diffusion almost lives up to its hype. It is CGI for dummies, and will produce photorealistic video with less effort than hand-drawing a stick figure. LLMs might disappear entirely, with every aspect of their design replaced, but ‘remove all the pixels that don’t look like Hatsune Miku impregnating Goku’ actually works. God help us all.
The practical future is in “diffusion forcing,” where a human artist can draw however many frames they like, and the robot can only fill in the gaps. If the robot does something wrong… draw more frames between stuff. One frame per second and a say-what-you-see description will probably suffice.
We can presumably also expect variations that finish sketchy animatics, but that’s always going to be less art-driven than artists would like. Absolute maniacs like James Baxter can feed in a pencil version of a camera orbiting a ballroom dance, but the robot will emit a broadly similar motion in finished quality. It’s better-off being used to turn on-fours into on-ones.