I signed up to try out DALL·E as a tech blogger from Trinidad and Tobago some time ago and today I was accepted to try it. What is DALL·E? It is an AI app developed by Open AI that creates images from textual descriptions. According to their blog, they have trained a neural network called DALL·E that creates images from text captions for a wide range of concepts expressible in natural language.
For this blog post I am sharing some of my testing. I used 6 phrases and for each phrase I am sharing what I think is the best image. It takes about 5 seconds to generate 6 images after I submit my description. I am impressed by what this app can do and to me this is tangible proof of the power of AI but I think it can be greatly improved. It returns nonesense for something like "the year 2050" and fails to depict correctly something like "A blue sign that says 'we are open'" or just simply "5000". Share your thoughts with me in the comments below. The DALL·E name is a mashup between the movie robot named WALL-E and Dali the surrealist painter.
|
A monkey eating doubles in Trinidad |
|
A panda playing the steelpan on dimache gras |
|
The meaning of life |
|
The center of the universe |
|
Mango ice cream melting on the beach |
|
A fish singing karaoke |
Comments