Conditions have been absolutely dramatically utterly bad over the last few months. It sucks because September to December is usually when we get our best visibility here around Laguna Beach, because  cool waters coupled with Easterly warm winds  knocking down the surf usually means awesome visibility. Well so far, not this year... Viz has been steadily reported by a few courageous locals to be stubbornly stuck around 10ft. As the good peeps at South Coast Divers would say: not worth it. 

So to pass the time, I've decided to take pictures in the metaverse instead. 

Well, not exactly. 

More like I asked the metaverse to take pictures for me. 

It's fun and addictive. It's disruptive. It's scary. 

But let me get a few things straight first: one, nobody knows what the metaverse really is today. So *my* metaverse is a place where cool Artificial Intelligences (AIs) live. And in my metaverse, there's an AI called GPT-3 and an AI called Stable Diffusion. GPT-3 understands me when I talk/type and I can ask them to tell me a story. I tried this a year or so ago and it was quite bluffing (a little bit on this here).  Now GPT-3 is two years old. An eternity in AI (remember Skynet took over the world in like 3 months). Stable Diffusion is way younger. Just a few months old. And in its current iteration, it's only a few days old! Just like her cousin GPT-3, Stable Diffusion is a generative AI, meaning it creates new stuff based on a prompt. Where GPT-3 will tell you a story based on a text you input, Stable Diffusion (SD) will draw you a picture. And it's bluffing. Disruptive. Scary. 

You can ask SD for a picture of fluffy kitteh (everyone does). It complies: 

 

You can ask for a painting of a kitty in the style of Monet and it will also comply:   

       

Bluffing.   

So let's ask for turtles. Here they are:                                                                                            

Where does that leave us photographers, artists and other creative people?

Well if you're making money selling your images today, your days are numbered. Or not. SD can actually be an opportunity for those who embrace it. Generative AI is here to stay so we might as well take advantage of it. One of its great features is it's able to learn a new style. An artist can easily train it to imitate their own style and generate infinite amount of new art. Since this new trained instance of SD is unique to that particular artist, they are still in control. Another great feature is image to image: you feed it an image and tell it what you want that image to become, and it does it. You can start with a sketch and iterate, or you can start with one of your award winning shot and it will adapt accordingly. The kelp and reef fish where all based off real shots I took this year. That's why composition stayed consistent. SD can generate a ton of things but true artistry will still be in the eyes of the artist. Sure, I can't draw anything even if my life depends on it and now I can paint like Picasso, but that does not mean that what I ask SD to paint for me will be systematically an award winning piece of art. It might, but since everyone have the same access to the technology, chance are true artists will still be able to differentiate.

However, if you are making money selling stock photos. Well good luck with that. Why would I pay a subscription to Shutterstock or Getty images to illustrate the 10 public PowerPoints I do each year when I can  generate anything rights-free in a few minutes? Really anything. Even new species of nudibranchs.