Everything about DALL-E 3: the A.I. text-to-image generator

DALL-E 3, an artificial intelligence model capable of generating images from text, is now available to all users. We tell you everything you need to know about it, how to get credits, the differences with DALL-E 2, how to use DALL-E Mini, and what implications this tool could have in the future.

What is DALL-E?

The DALL-E is a new A.I. software that turns your words into images and works of art. Users can simply insert a text in English, and the neural network will generate an image out of it. The tool is developed by OpenAI, a startup backed by Microsoft, and Google's Imagen technology. So far, the description of DALL-E on the OpenAI website says it's "a new AI system that can create realistic images and art from a description in natural language", which is exactly what it does.

It can create even the most improbable pictures: such as "Teddy bears shopping for groceries in ancient Egypt". In January 2021 OpenAI introduced DALL-E, an older version of the software. Now it's a smarter and newer system – DALL-E 2.

Its developers say that "DALL-E 2 will empower people to express themselves creatively. DALL-E 2 also helps us understand how advanced AI systems see and understand our world, which is critical to our mission of creating AI that benefits humanity".

How does DALL-E work?

The DALL-E tool is a neural network that is trained by OpenAI to generate images from text. It can combine unrelated concepts and create anthropomorphized versions of animals and objects. DALL-E works together with CLIP, a computer vision system that OpenAI has already introduced last year. The neural network translates human language into its own language and only after that turns them into images. 

Giannis Daras, a Ph.D. candidate in computer science, posted examples of what he called "AI's own language" on Twitter:

"Apoploe vesrreaitais means birds," Daras wrote on Twitter. "Contarra ccetnxniams luryca tanniounons," means bugs or pests".

DALLE-2 has a secret language. "Apoploe vesrreaitais" means birds. "Contarra ccetnxniams luryca tanniounons" means bugs or pests. The prompt: "Apoploe vesrreaitais eating Contarra ccetnxniams luryca tanniounons" gives images of birds eating bugs. A thread (1/n) 

DALL-E's interface is simple: there is a text box where the user can insert text, a button to start the conversion, and images just below it. Its developers explain how DALL-E 2 works: "It uses a process called "diffusion," which starts with a pattern of random dots and gradually alters that pattern towards an image when it recognizes specific aspects of that image".

For now, the new version of DALL-E is only available to a small group of people to avoid any hate, nudity, and inappropriate statements that may harm users. Although, you can join the waiting list by explaining why you'd like to test the program and maybe the developers will grant you early access.

What is the difference between DALL-E 3 and DALL-E 2?

DALL-E 3 is mainly an improvement of the algorithm previously released under the name DALL-E 2. In its development, several fundamental improvements have been implemented and the image base to feed the machine-learning model has been extended. In addition to handling textures, shadows, and perspectives, it can realistically edit existing images from a more complex textual description than before. You can also "draw inspiration" from an image to create variations while retaining the same style. 

How to use DALL-E 3 with ChatGPT

The latest version of DALL-E is now coupled with ChatGPT to create even more sophisticated images. That means that from now on you can type a prompt or command in ChatGPT to make it create an image with more complex features. From the first result, you can continue to make improvements by launching new specifications in the chat bar of this tool. This leap represents a major advance for OpenAI over other artificial intelligence algorithm developers, as the richness of ChatGPT's collection is capable of transmitting more complex commands to the DALL-E 3 image generation tool. 

Using these tools combined is like giving precise indications to a painter based on the progress of his work. You can be as specific as you require, asking for example for "an image of a rose beginning to wilt in a glass vase, placed on a table in front of a window, where the rays of sunset light are coming through". As a comparison, in earlier versions of DALL-E, the prompt would have had to be simpler: a rose in a vase on a table, and the results might not have been as pleasing to you.

You may wonder how you can use DALL-E 3 with ChatGPT. This option is included in ChatGPT Plus, Open AI's subscription-based service. It gives you access to ChatGPT 4, with faster response times and priority access to new features. All this costs $20 per month.

How to use DALL-E 3 for free with Bing

Now, if you're looking for a free way to take advantage of all the possibilities of DALL-E 3's generative AI, we have good news for you. Microsoft has taken a big step to position its Bing search engine as one of the most influential in the market, and to do so it has implemented OpenAI technology in its search engine. The reason is very simple: it is Microsoft itself that has invested more than US$1 billion in OpenAI since 2019. The implementation of ChatGPT and DALL-E in Bing Chat will be rolling out in the coming weeks, with the goal that you will soon be able to make faster queries or search for images as you were thinking of them.

• To use it, simply login to Bing, sign in with your Microsoft account, and click on the Chat icon at the top of the screen.
• You can then type in the description of the image you want in the text bar, even putting specific information such as a certain style, color range, etc. In this example, we have given the command "Create an image of a cat eating in the style of Lucien Freud". Also, one of the new features of this integration is that you can now write your prompts in other languages than English, although many times the link suggestions appear in English.
• After a few seconds, you will see an answer with additional information and useful links according to the topic, and finally, you will see the generated images. This tool is already available with DALL-E 2, so you will immediately see an AI-generated image, but once the new version is implemented, the results will be even more impressive.

Is DALL-E 2 available to everyone?

When Open AI released DALL-E 2 it allowed all users to create an account to experiment with this image creation tool. However, it has now moved to a paid model whereby you are required to purchase credits to "spend" creating and downloading images. Knowing this, the only thing you have to do to use DALL-E 2 is:

• Access the Open AI DALL-E page.
• Create an Open AI user account or log in with your username and password if you already have an account.
• Once inside, in the central window, you will see a search bar where you can type the combination of words you are interested in to generate your images. Then click the Generate button. 
• Once created, you can edit the images, change specific parts, the style of the images, etc.
• You can also simply upload an image to edit it with DALL-E. To do this, click on the Upload image to edit option, which appears just below the bar to generate new images.
• You can only generate new images if you have DALL-E credits. 

Is DALL-E 2 free: how to get credits

DALL-E offered free monthly renewable credits to everyone who registered before April 6, 2022. These credits were renewed monthly on the day the user created the account and, if not used, expired with a validity of 30 days. Unfortunately, if you are creating your account after this date, you will no longer be able to receive free monthly credits and in order to use DALL-E 2 you will have to buy credits by clicking on the Buy credits option. Just above this button, a number will appear with the amount of credits you have available.

Currently, the cost of credits to use DALL-E 2 is $15 for a package of 115 credits, each of which allows you to make a query that generates 4 images, having a total of 460 files for that price.

What is DALL-E Mini and where to try it?

DALL-E Mini is the simplified, open-source version of DALLE-2, available on GitHub and with a less advanced infrastructure than its big brother, which makes it less resource-consuming. While it is easier to use, its results are less accurate, which has attracted an audience interested in playing with this eccentricity of the DALL-E Mini. For the programmer, Boris Dayma, this interest is extremely useful because, although users only want to have fun, their interaction with the program is already feeding the algorithms that compose it. In DALL-E Mini you will be able to experiment with objects, painting styles, places, and even TV characters, singers, politicians, and other celebrities. Although you will be able to see for yourself the limitations in the definition of the images, you will get an excellent idea of how this technology works.

To start using DALL-E Mini go to this page, type the words you want to use to compose your image in the central blank bar, then click the Run button to display the results.