Question 1

True or false: DALL-E 2 was trained on a dataset that included text paired with images from the internet.

Accepted Answer

True. Yes, DALL-E 2 was trained on a large dataset of image-text pairs from the web, similar to CLIP.

Question 2

True or false: DALL-E was named after the surrealist artist Salvador Dalí and the robot WALL-E.

Accepted Answer

True. Correct—the name is a portmanteau of Dalí and WALL-E, blending surrealism and robotics.

Question 3

True or false: DALL-E 2 can edit existing images by inpainting specific regions based on a text prompt.

Accepted Answer

True. True—DALL-E 2 supports inpainting, allowing users to modify parts of an image with new content.

Question 4

True or false: DALL-E 3 can only generate square images with a 1:1 aspect ratio.

Accepted Answer

False. DALL-E 3 supports multiple aspect ratios, including landscape and portrait, not just squares.

Question 5

True or false: DALL-E 1 could generate photorealistic human faces without any distortion.

Accepted Answer

False. DALL-E 1 often struggled with human faces, producing artifacts; photorealism improved significantly in later versions.

Question 6

True or false: DALL-E was first announced in 2022, making it older than Midjourney.

Accepted Answer

False. DALL-E was announced in January 2021, while Midjourney launched in 2022, so DALL-E predates it.

Question 7

True or false: OpenAI open-sourced DALL-E 2's full model weights for public use.

Accepted Answer

False. OpenAI never released DALL-E 2’s weights; it remains a closed, API-only model for safety reasons.

Question 8

True or false: DALL-E 3 uses a technique called 'text-to-image diffusion' to create images from prompts.

Accepted Answer

True. Yes, DALL-E 3 is a diffusion model that iteratively refines random noise into an image guided by text.

DALL-E Trivia Questions