Stable Diffusion

And what I observed was like a lot of like frontier algorithmic ideas were developed by professors or researchers in Germany.

For example, stable diffusion was developed in Germany, invented and open sourced.

And they also wanted to turn it into a company.

But by the time they finished basically paperwork, Already mid-journey and open eye turned it into products on the other end of the world.

And you had Delhi and mid-journey and no one was interested anymore in buying stable diffusion as a commercial product.

So with inflammation speed is something we still lack.

But what I'm wondering maybe is now basically, okay, we export a lot of knowledge and they turn it into AI.

Meanwhile, two computer scientists managed to fool a bunch of Redditors into giving them money.

They used stable diffusion to generate fake images of an attractive female and convinced the men to give some money if they wanted to see more pictures.

I'd imagine that this was one of the easiest scams to pull off, and unfortunately, there's going to be plenty more like it.

An AI version of Drake and an AI version of The Weeknd collaborated in a fake song called Heart on My Sleeve, and it was doing the rounds on social media and Spotify, gaining hundreds of thousands of streams on Spotify alone.

We're all used to it.

It's almost like this has always been possible.

From mid-journey to stable diffusion and dream up, any user can now create stunning and surreal images from text prompts in under a minute.

It's amazing when you think about it.

So that man was Stephen Andring of Wisconsin, who was a software engineer.

And what he had done was he had installed something called stable diffusion, which is some kind of AI program on his laptop.

He added software components that allowed the program to create depictions of realistic human private parts.

And then he used curated text prompts and all kinds of things to create photorealistic images of minors in sexually explicit positions.

Um, so the paper which you posted yesterday, and you've been, um, you've been probably the, the most important, you know, kind of, uh, commentator, as also professional in this area.

I'll remind people you co-founded Stable Diffusion, Stability AI.

Uh, you have a master's in mathematics from Oxford.

And, and you've been thinking about these problems for many, many years, and you and I have been talking for many months now.

The first was image generation.

Now, this is a little bit borderline because obviously even before things like Midjourney and Stable Diffusion and later NanoBanana and GPT Image, people were interacting with images.

But still, before it was really easy to generate them, interacting with images or at least creating images remained a more specialized domain.

If you look at the big generative AI releases from November 2022 as a proxy, it's not surprising that ChatGPT was so quickly so much more ubiquitous than Midjourney's version four, which also came out that month.

So the Minimax model, the video model, which was released as open source, we covered it, but it is making so many waves.

It's like a stable diffusion moment where they throw this out and people-

Yeah

... grab it and make it faster and make it render continuous, as we'll cover later.

But this is, yeah, a milestone I think.

Very much a milestone.

This definitely feels like the stable diffusion moment for video.

For those of us who are new to AI and didn't join the stable diffusion moment, once the open weights dropped for stable diffusion, multiple things happened in parallel.

they couldn't be linked to any actual they were they were computer generated so the children in these pictures were not real children but prosecutors argue that andrag did more than just create these images and keep them inside of his house okay so this investigation actually started after instagram reported that images had been sent to an account belonging to a 15 or 16-year-old boy.

According to the court record, Anderegg then explained to that boy how he made the pictures that he was sending, which was by using this platform, Stable Diffusion.

So Anderegg eventually gets charged with four federal crimes, producing obscene images depicting minors, distributing those images, sending obscene material to someone under the age of 16, and possessing the images inside of his home.

Only that last charge, possessing the images inside of his home, has been dismissed, okay? The other three charges are still very much moving forward.