Dall-E 2, the amazing AI artist

OpenAI team created an AI that generates images with English text as its input. This AI is called Dall-E. The responses it generates are so amazing that you can consider Dall-E as your personal AI concept artist.

An AI artist robot creating amazing artwork, a digital art

In this article, I will share a few capabilities (with illustrations) of this amazing AI artist, Dall-E.

All the images used in this article were generated by Dall-E based on my text prompts. Including the above cover photo of this article, where I asked it to generate “An AI artist robot creating an amazing artwork, a digital art”.

How to get yourself access to Dall-E?

At the time of writing this article, they’ve opened up Beta access and plan to provide access to over a million individuals over the next few weeks. So, apply for access on their page.

I got my access earlier during their private beta testing. So I’d say I’m grateful to be an early member to try out Dall-E.

Limits on the number of Dall-E requests

Once you get access to Dall-E, you will get 50 free successful requests for the first month and 15 free credits every subsequent month. Free credits get expired at the end of each month if unused. You can also buy 115 additional credits for $15. These paid request credits have an expiry of 12 months.

Usage of the images generated

With the beta release, OpenAI announced that you may use these image generations for any legal purpose, including for commercial use, as long as the images that are generated comply with their policies. You can read more about their terms and policies here.

Capabilities of Dall-E 2

I tried to list down a few different styles of images Dall-E is capable of generating. And these are just a few styles I wanted to show and it's capable of way more than what you see below.

I suggest, don’t just scroll through, but observe the details in each image along with the text I used for the input which I mentioned in each image’s caption.

Concept Art

Planning to make a movie? Say, a story of a monkey sage in Stonehenge.
Dall-E helps you to generate amazing concept art given a good prompt.

A landscape view of a huge monkey man, glowing in the divine yellow aura, meditating in the Stonehenge area on a starry night, a high-quality digital painting

Mimic artist styles

How about mixing the art and art styles of legendary artists?
Say, a mixup art piece of Leonardo Da Vinci and Vincent van Gogh

A painting of Mona Lisa in the style of Starry Night

Photography

Realistic photos with a specific shot type and even a specific focal length.
OH, MY DSLR!

a teddy bear with sunglasses resting on a beach chair under the sunlight on the beach, a wide-angle shot from a 24mm lens

Creative Digital Art

You use the words to your imagination and let Dall-E take care of the artwork.

Digital art of Sadhguru saving the soil to restore soil health and rekindle our bond with Mother Earth

Pencil art style

Imagine the impossible and put it into words for the fastest pencil artist ever.

Color Pencil art of two red ants carrying a clown fish on a leafy tree branch

One-line drawings

Me: Hey Dall-E! Don’t lift the pencil off, while you draw this.
Dall-E: Challenge accepted!

One line drawing of a cat chasing a mouse in a street

Synthwave-style

Beautiful usage of synth-wave colors in an artwork.

A Lamborgini car racing towards the sunset on a deserted road, synth-wave style

3d Renders

A 5-second quick 3d render is generated just using words in your mind.

A squirrel sneaking in through a grilled window into a kid’s room (with photos and posters on the wall and a DSLR camera hanging), trying to steal a piece of cornflakes from a bowl of cornflakes, 3d render

Pixelated art

Create your own pixel art sprites for your 2D games

Pixel art of a clown fish in an aquarium bowl

Upload your own photos and fill in what you want

You can even upload your image to Dall-E. It allows you to erase a part of the image and then let you enter a prompt and Dall-E fills in the erased part based on your text prompt.

A butterfly flying on the flower, a high-quality image

Limitations

Dall-E is not completely perfect. The images it generates are sometimes not accurate. The images are not always razor sharp. Below are a few more notable limitations.

Text in the image

Although it accepts text as input to generate an image. It somehow cannot put up words in the image. It shows some gibberish text.

A man photographing a sign that says “No Photography”

Faces

If you try to generate human faces, they are most of the time blurred out or distorted. But this limitation is understandable.

Close-up shot of a man fishing in a river dock along with his daughter during a sunset

And you may have observed too that the face of the Mona Lisa — Van Gogh image you saw up there was also not super accurate. But, at least it tried.

There are a few more such limitations. But overall it's an amazing tool. Apply to access it and try out your own creations.

Conclusion

What we saw are just a few simple prompts that I was able to come up with in a matter of a few minutes. Companies out there, are even creating magazine covers using this tool.

As for me, I always wanted to learn to create animated movies. Coming up with good concept art always seems to be a starting trouble. Now, I found my personal AI concept artist to draw whatever I want in a few seconds.

Believe me! The results are so amazing when you get the prompts right.

I can’t imagine what the future looks like with a tool like this in every creator's hands. It speeds up a lot of creative tasks.

Like Dr. Karoly always says, What a time to be alive!

If you liked what you just read, follow me on medium for more and go through the journey with me on topics like Technology, Photography, Self Development, and Personal Finance.