Living in the future: AI image generator DALL-E is now public

Hey OF,

Not sure if this will resonate with any of you, but I think this is pretty cool. And I’ll try to keep this simple. So… y’know that predictive text stuff you see every now and again on your iPhone or Google service. Well it doesn’t stop there, and this group called OpenAI have used their skills to create DALL-E 2. Simply put, it’s a language prompted Artificial Intelligence engine that can create entirely unique images based on your inputs. And it’s free. And no, I have no reason to promote it other than the fact that I think it’s super neat.

Examples that I tried: “Monet painting of a miniature poodle water skiing”
Result:


Another: “cartoon koala jumping over a fault line in the desert”
Result:


Or something kinda bonkers. “Marc Chagall oil painting of a timepiece on a forest floor surrounded by bananas”
Result:

Can you try “My uncle left me this watch, I’m not trying to sell it, but tell me what it’s worth”?

My understanding is that all the images are public domain as these are entirely unique creations. They aren’t repurposing existing images of Speedmasters in the examples above (which could/would be a copyright violation).

And it works for photorealistic images as well, so in addition to being a tool for graphic designers it would also be beneficial for people who would otherwise have to pay for stock images.

The tool that powers Dall-E also works for complex writing compositions. I don’t think it’s public yet, but apparently you can give it a few lines of text and then tell it to, say, write a term paper on the history of watchmaking in the style of Hemmingway… and it will! And that could be an entirely unique creation as well.