DALL‑E, the New AI Art Generator, Is Now Open for Everyone to Use

If you spend any time at all on social media, you’ll have glimpsed the work of DALL‑E, Ope­nAI’s now-famous arti­fi­cial-intel­li­gence engine that gen­er­ate images from sim­ple text descrip­tions. A veloci­rap­tor dressed like Travis Bick­le, Amer­i­can Goth­ic star­ring Homer and Marge Simp­son, that astro­naut rid­ing a horse on the moon: like any the-future-is-now moment, espe­cial­ly in recent years on the inter­net, DALL-E’s rise has pro­duced a host of arti­facts as impres­sive as they are ridicu­lous. Now you can try to top them in both of those dimen­sions your­self, since not just DALL‑E but the new, improved, high­er-res­o­lu­tion DALL‑E 2 has just opened for pub­lic use.

“How do you use DALL‑E 2?” You might well ask, and Cre­ative Bloq has a guide for you. “The tool gen­er­ates art based on text prompts,” it explains. “On the face of it, that could­n’t be more sim­ple. Once you’ve com­plet­ed the DALL‑E 2 sign up to open an account, you use the pro­gram in your brows­er on the DALL‑E 2 web­site. You type in a descrip­tion of what you want, and DALL‑E will cre­ate the image.”

Of course, some prompts pro­duce more visu­al­ly inter­est­ing results than oth­ers. The guide rec­om­mends that you con­sult the DALL‑E 2 prompt book, which gets into how best to phrase your descrip­tions in order to inspire the rich­est com­bi­na­tions of sub­ject, tex­ture, style, and form.

“Even the cre­ators of DALL‑E 2 don’t know what the tool knows and does­n’t know. Instead, users have to work out what it’s capa­ble of doing and how to get it to do what they want.” And indeed, that’s the part of the fun. DALL-E’s own inter­face rec­om­mends that you “start with a detailed descrip­tion,” and with a lit­tle exper­i­men­ta­tion you’ll dis­cov­er that speci­fici­ty is key. The ren­der­ings of “an eight-bit Nin­ten­do game designed by Hiroshige” and “a cyber­punk down­town Los Ange­les scene paint­ed by Rem­brandt” strike me as cred­i­ble enough for a first effort, but adding just a few more words opens up entire­ly new realms of sur­prise and incon­gruity.

Just above, we have two of DALL-E’s infi­nite­ly many pos­si­ble attempts to visu­al­ize “the cov­er of an old Ernest Hem­ing­way pulp nov­el about the adven­tures of David Bowie.” Though the designs look entire­ly plau­si­ble, the titles high­light the tech­nol­o­gy’s already-noto­ri­ous inabil­i­ty to come up with intel­li­gi­ble text. Oth­er lim­i­ta­tions of the new­ly pub­lic DALL‑E, accord­ing to Ars Tech­ni­ca’s Benj Edwards, include the require­ment to pro­vide your phone num­ber and oth­er infor­ma­tion in order to sign up, the own­er­ship of the gen­er­at­ed images by Ope­nAI, and the neces­si­ty to pur­chase “cred­its” to gen­er­ate more images after you’ve run through your ini­tial free 50. Still, there’s noth­ing quite like typ­ing in a few words and sum­mon­ing up works of art no one has ever seen before to make you feel like you’re liv­ing in the twen­ty-first cen­tu­ry. You can sign up here.

Relat­ed con­tent:

Dis­cov­er DALL‑E, the Arti­fi­cial Intel­li­gence Artist That Lets You Cre­ate Sur­re­al Art­work

An AI-Gen­er­at­ed Paint­ing Won First Prize at a State Fair & Sparked a Debate About the Essence of Art

Arti­fi­cial Intel­li­gence Brings to Life Fig­ures from 7 Famous Paint­ings: The Mona Lisa, Birth of Venus & More

Google App Uses Machine Learn­ing to Dis­cov­er Your Pet’s Look Alike in 10,000 Clas­sic Works of Art

Arti­fi­cial Intel­li­gence for Every­one: An Intro­duc­to­ry Course from Andrew Ng, the Co-Founder of Cours­era

Based in Seoul, Col­in Mar­shall writes and broad­casts on cities, lan­guage, and cul­ture. His projects include the Sub­stack newslet­ter Books on Cities, the book The State­less City: a Walk through 21st-Cen­tu­ry Los Ange­les and the video series The City in Cin­e­ma. Fol­low him on Twit­ter at @colinmarshall or on Face­book.