You can finally use DALL E 2, the AI ​​that generates realistic images from text

Image for the article titled You can finally use DALL E 2, the AI ​​that generates realistic images from text

Image: open AI

“A white cat wearing sunglasses on a pink background in studio light.” “A photo of a giraffe on a jet ski.” If you too have been waiting for weeks for OpenAI to give you access to DALL·E 2 to create this type of realistic and other more artistic images from a text description, I have two pieces of good news.

The first is that the DALL·E 2 artificial intelligence model is already available in beta versionwhich means that OpenAI will deliver a million invitations to those who want to test it, starting with those who are already signed up for its long waiting listand following by those who sign up from now on.

The second is that OpenAI has announced a credit system for DALL E users that, although it is payable after a certain limit, could revolutionize the industry of stock photography and commercial illustrations, since the Californian company will assign the rights of use of the photos to the users who generate them.

Each DALL·E user will receive 50 free credits during their first month of use, plus an additional 15 free credits in subsequent months. One credit allows you to generate a set of four images, so 50 credits equals 200 images based on 50 different text descriptions or “inputs”.

In addition to original images, generated from a description written in natural language, DALL E can make edits on an image uploaded as a source (the AI ​​​​changes or enlarges a photo or a work of art) and generate variations from one image (the AI ​​creates four new images inspired by the original). All of these actions cost one credit, but generate four results instead of one.

OpenAI has also announced the price of credits for this public beta phase. Users will be able to add balance to their account in increments of 115 credits for $15. That is, with 15 dollars they will be able to create 460 new images from 115 different tickets.

It seems like a reasonable price considering that, from now on, users will get all the commercial use rights of the images generated with DALL·E (including those generated in the previous research phase). This means that a person who has no idea of ​​photography, design or illustration can create images with DALL·E to illustrate a web project, create game characters, sell t-shirts and posters…

It is a before and after in the industry, but it has its limitations: you cannot create images of public figures with DALL·E, nor upload photos of realistic faces to edit them. Nor can violent or offensive images be generated, and the images generatedthey will not be monitored by algorithms or human reviewers to prevent abusive use of the tool.

Still the feeling is of that we are about to experience a major change, at least until other imaging models achieve similar results at a lower price or even for free.

You can sign up for the waiting list of DALL·E 2 on the OpenAI website, but be sure to take a look at its instagram account to see examples of realistic or artistic images generated with its artificial intelligence.