DALL-E 3 Is Now Free For Everyone!

Great news, the new image generator AI, DALL-E  3 is now available for everyone for free. I’ve   put a link to it in the video description,  make sure to check it out. And it has some   surprises for us, so we are going to  have a look at 6 unexpected results. First, I read that supposedly, it is excellent  at visualizing proverbs. For instance,   “laughter is the best medicine”. What could  that look like? Oh. Do you remember when we   tested Photoshop’s image generator, and I asked  for an image with caustics, colorful patterns of   reflected light that should look like this, and we  got…this instead? Well, similar phenomenon. Hope   it gets sorted out. Now, luckily, just a week  ago, someone ran the same prompt successfully,   and this is truly excellent. Note that it did  not get detailed instructions on how to do this,   just “laughter is the best medicine”. You  figure out the rest, little AI. Plus points   for the old comic book style. So good!   Midjourney can’t do this. Not even close. Two, what about character consistency? We got a  taste of this with the amazing Larry the Hedgehog,   however, that was in the promo materials.   Does it also work well in practice? Oh my,   yes it does! This is the same family in four  different settings. Once again, this requires   the implementation of research papers, or at the  very least, a series of tricks in most other text   to image systems. I have to note that I tried it  too, and in my case, it was a little copy-pasty,   you know, it’s easy to keep consistency if you  just copy the same thing over and over, the images   are of somewhat lower quality, so my impression  is that this can be done, but requires a little   cherry picking. That is a huge step forward.   Next step will be that we take it for granted. Especially that three, this consistency can be  requested in more advanced ways. For instance,   the same character with different ages. This  is one of the best text to image AI results   I have ever seen. Goodness. And hold on to your  papers Fellow Scholars, because it gets better.   If you feed these images into  Runway’s interpolation technique,   you get a fantastic video. I put a link  to these results in the video description,   many of which have prompts available, which  will help you in your own Scholarly experiments. Four, I was surprised by how good it is at  sketches. This one is incredible. And, of course,   we will get sketches of you Fellow Scholars as  well. Always. But you know what is missing here,   Fellow Scholars is great, but Fellow Scholars  holding on to their papers is even better. Oh,   look at that! Five. Text! Text is finally  working! A historic moment! Glorious! And finally, six. It is surprisingly imaginative.   It can cook up an image of a relaxing parrot.    Easy. There is room for the tail, and the legs  are up in the air. Really cool! Asking for a   base on the Moon? Not a problem. Paintings are  also incredible. This one just radiates with   joy. Love it. Digital art? It can not only  do it, but it is super imaginative. And now,   what you all have been waiting for. As is  tradition, our legendary stern fox librarian   from DALL-E 2 can now be created again.   Will it get better? It is really good,   and he has gotten a lot more stern this  iteration, I can tell you that right now. So, yes, DALL-E 3 is really good. No, it is  not perfect. But, really good. Yes, it might   require a little cherry picking. And one more.   And one more. There we go. Thank you, little AI! And just one more thing. We are in the age of AI,   and I hear more and more rumors about myself  being replaced by an AI who is now narrating   these videos. So let it be known that every  episode is being recorded by my own voice.