I wanted to use this blog post as a time to just talk generally about the progress of AI, and specifically something that really impressed me recently.

So one recently emerging AI-powered technology that absolutely astounds me and just blew me away was the text-to-image generator DALLE-E 2. This program is absolutely incredible. And the way it works from my understanding is it’s powered by a neural network which has the capability to generate on the fly images from text. And while that sounds ambiguous, its because it is. But that really is how open-ended the capabilities are. For example if you typed in “Man walking a dog” you would get images of that. But it can do so much more.

With DALLE-E 2 you can set art and rendering styles. You could say something like “A mountainous sunset on an alien planet with two moons in the style of Van Gogh” And it would generate several images of a sunset on an alien planet, in the style of Vincent Van Gogh. Which is incredible to me. You could also say something like “Boat out on the Ocean rendered in 3D” and it would create a 3D rendering of a boat on an ocean. It’s really cool.

This technology is so incredible in fact, that someone stirred up some controversy recently when they entered an art competition (and won) using images generated by an AI. The fact that we are at a place in our AI tech that it’s capable of generating art that is perceived to be good enough to win an art competition is really incredible. What’s even cooler is that the makers of DALL-E 2 have said that any image generated by the software user, is in fact the property of that user. So it’s commercial/individual uses are endless. I’m very excited about the future of this and other technologies like it.

Article: https://www.nytimes.com/2022/09/02/technology/ai-artificial-intelligence-artists.html