DALL-E 2 creates AI images of the future of San Francisco

“I keep waiting [Sutro Tower] roll down the hill and attack the Golden Gate Bridge,” joked longtime San Francisco Chronicle columnist Herb Caen.

It’s easy for any San Francisco resident to imagine the scene, our city’s two most important symbols engaging in a Godzilla-style battle. But what happens when you ask a computer to create an image of landmark-on-landmark combat? Or when you pose more complicated scenarios to the AI, like asking it for an image of how climate change would affect the city?

Thanks to software from a new San Francisco-based artificial intelligence company, we now know. The DALL-E 2 image generation web application allows users to enter a text prompt and receive an array of six stunningly realistic images. Beta access users produced everything from expansions to Michelangelo’s “Creation of Adam” at New York Cartoons. And a publicly available competitor dubbed DALL-E Mini (and recently renamed Craiyon) has become a go-to tool for generating memes.

AI imaging software DALL-E-2 created this visual rendition of “The Salesforce Tower, But Make It Beautiful”.

Courtesy of OpenAI

The app is the latest creation from OpenAI, a nonprofit Elon Musk and a host of other Silicon Valley heavyweights founded in 2015 with $1 billion in funding. It lost its nonprofit status when it received another billion from Microsoft in 2019. So far, the tech company has built technology used to teach machines to debate, help robotic hands to master movement, produce self-generated journalism and even create the first iterations of a very flirtatious chatbot app. Given the dangers inherent in AI advancements, the company’s publications are generally received with a mixture of wonder and dread, and OpenAI’s allegedly secretive nature and changing financial status have drawn suspicion from journalists. of investigation.

The deep and false dystopian implications aside (the software has safeguards against displaying real people’s faces and violence), the few users with access to DALL-E 2 are having a great time with the software. Currently, it’s in an invite-only beta phase, with the stated goal of onboarding up to 1,000 users from their waitlist each week.

Although OpenAI does not grant a beta account to SFGATE, we were able to provide a series of prompts. However, we learned that specificity is key, and without the ability to fine-tune the input, the experience was like telling a genius you wanted to fly and get a pair of mosquitoes in return. For example, asking to see the streets of San Francisco with gondolas, instead of cars, resulted in Venetian-style boats, rather than soaring ski lifts.

AI imaging software DALL-E 2 created this visual rendition of what San Francisco would look like with gondolas, in the style of a Norman Rockwell painting.

AI imaging software DALL-E 2 created this visual rendition of what San Francisco would look like with gondolas, in the style of a Norman Rockwell painting.

Courtesy of OpenAI

Danielle Baskin, concept artist from San Francisco (of Google Spirit of Halloween prank and BART Basel infamy) is an early adopter of DALL-E 2 and has spent hours playing with the software and exploring its limits. In April, she posted a Twitter thread showing what the Golden Gate Bridge would look like if there was a BART train crossing it, as well as a Sausalito full of high-rise buildings.


“My favorite thing about it isn’t necessarily doing anything with the results, I love the feeling of exploring endless combinations of images,” Baskin said. “It’s kind of like walking into a museum and being able to go down to different wings and see any particular part of history or an imaginary scenario.”

Although the software generates results in seconds, creating professional looking photos is not instantaneous. Baskin spent two hours refining his Golden Gate Bridge prompts. And some ideas, like showing what an accessory living unit would look like inside a San Francisco garage, were simply beyond the limits of the software’s conceptual imagination. Bike advocates will be sad to learn that it was also nearly impossible to create a bike-only version of the Bay Bridge, without any cars.

AI imaging software DALL-E-2 created this visual rendition of a corgi flying over the Golden Gate Bridge, like a stained glass window.

AI imaging software DALL-E-2 created this visual rendition of a corgi flying over the Golden Gate Bridge, like a stained glass window.

Courtesy of OpenAI

“Maybe from the outside it looks like the AI ​​is so good because you can just input whatever you imagined and it spits out a nice picture,” Baskin said. “But the first ones were all messed up, the bridges were mutilated. I had to be very specific.

When SFGATE provided DALL-E 2 with a prompt asking what San Francisco would look like after climate change, the result was surprisingly conceptual, akin to something a publication might ask an illustrator to do to accompany an article. This type of use of the software understandably sent shivers through the art world, lest DALL-E 2 put illustrators out of work.

AI imaging software DALL-E 2 created this visual rendition of San Francisco after experiencing the effects of climate change.

AI imaging software DALL-E 2 created this visual rendition of San Francisco after experiencing the effects of climate change.

Courtesy of OpenAI

On that note, Baskin has received several private requests from publications asking him to run prompts for them. She postulated that in these situations, media companies had no budget to commission anything and would have ended up using stock images or creative commons.

“Any kind of powerful technology kills jobs and creates new jobs,” she said, noting how useful the tool would be for illustrators looking for inspiration or helping to refine their art to meet customer needs.

AI imaging software DALL-E-2 created this visual rendition of the cast of

AI imaging software DALL-E-2 created this visual rendition of the “Star Trek” cast enjoying a day at Fisherman’s Wharf.


Courtesy of OpenAI
AI imaging software DALL-E-2 created this visual rendition of the location of buried treasure in Golden Gate Park, as described in the book "The secret."

AI imaging software DALL-E-2 created this visual rendition of the location of buried treasure in Golden Gate Park as described in the book ‘The Secret’.


Courtesy of OpenAI

Three images created by AI software DALL-E-2 showing the ‘Star Trek’ cast at Fisherman’s Wharf and the location of the Golden Gate Park buried treasure described in the book ‘The Secret’. Images courtesy of OpenAI.

Besides exploring silly possibilities like what the cast of “Star Trek” would look like while enjoying a day at Fisherman’s Wharf, Baskin sees the potential of using DALL-E to help stretch people’s imaginations; perhaps it could even be used as a tool for activism. As the old saying goes, she discovered that a picture is worth a thousand words.

“A friend commented on the political power of DALL-E to show people very realistic alternative futures that they might otherwise object to because they can’t imagine it,” Baskin said. “Many architectural renderings look more like blueprints or sketches than a striking realistic photo.”

AI imaging software DALL-E 2 created this visual rendition of a picture-perfect day at Dolores Park, in the style of a Renaissance painting.

AI imaging software DALL-E 2 created this visual rendition of a picture-perfect day at Dolores Park, in the style of a Renaissance painting.

Courtesy of OpenAI

When it came to prompts provided by SFGATE about our fair city, DALL-E’s results were mixed. The aforementioned gondola prompt (in the style of a Norman Rockwell painting) and the one asking for “a renaissance painting of a perfect day in San Francisco’s Dolores Park” looked like paintings you might see on the wall of a room. a university cafe. An admittedly morose request to show the city after an 8.5 magnitude earthquake sounded like something out of a disaster movie trailer. Asking for help identifying the location of long-buried treasure in Golden Gate Park yielded no clues worth investigating, and our only attempt to recreate Baskin’s BART experience failed quite miserably. .

AI imaging software DALL-E 2 created this visual rendition of downtown San Francisco after an 8.5 magnitude earthquake.

AI imaging software DALL-E 2 created this visual rendition of downtown San Francisco after an 8.5 magnitude earthquake.

Courtesy of OpenAI

And as for the old Herb Caen quote about Sutro Tower attacking the Golden Gate Bridge? Well, sometimes a picture just can’t compete with a few well-chosen words.




AI imaging software DALL-E-2 created this visual rendition of Sutro Tower attacking the Golden Gate Bridge, in the style of a Godzilla movie poster.

AI imaging software DALL-E-2 created this visual rendition of Sutro Tower attacking the Golden Gate Bridge, in the style of a Godzilla movie poster.


Courtesy of OpenAI
AI imaging software DALL-E-2 created this visual rendition of Sutro Tower attacking the Golden Gate Bridge, in the style of a Godzilla movie poster.

AI imaging software DALL-E-2 created this visual rendition of Sutro Tower attacking the Golden Gate Bridge, in the style of a Godzilla movie poster.


Courtesy of OpenAI
AI imaging software DALL-E-2 created this visual rendition of Sutro Tower attacking the Golden Gate Bridge, in the style of a Godzilla movie poster.

AI imaging software DALL-E-2 created this visual rendition of Sutro Tower attacking the Golden Gate Bridge, in the style of a Godzilla movie poster.


Courtesy of OpenAI
AI imaging software DALL-E-2 created this visual rendition of Sutro Tower attacking the Golden Gate Bridge, in the style of a Godzilla movie poster.

AI imaging software DALL-E-2 created this visual rendition of Sutro Tower attacking the Golden Gate Bridge, in the style of a Godzilla movie poster.


Courtesy of OpenAI

AI imaging software DALL-E-2 created these visual renditions of Sutro Tower attacking the Golden Gate Bridge, in the style of the Godzilla monsters. Image courtesy of DALL-E-2.

Comments are closed.