Artist fakes time-travel selfies with Stable Diffusion
An anonymous artist known as "Stelfie the Time Traveller" just used Stable Diffusion and DreamBooth to craft photorealistic "historical selfies." By fine-tuning the model on a fictional face, the creator inserted the character into scenes like building the Egyptian pyramids and Leonardo da Vinci's studio. It's a clever, playful demonstration of maintaining character consistency for visual storytelling.

If time travelers were real, they would almost certainly be taking selfies.
Throughout December, an anonymous creator posting under the handle Stelfie the Time Traveller has taken social media by storm with an inventive, AI-generated photo diary. The project features a fictional, photorealistic bald traveler popping up in some of the most famous eras in history.
So far, Stelfie's timeline is wild. He has snapped selfies while being chased by a woolly mammoth during the ice age, checked in at the construction site of the Egyptian pyramids, and stood outside the walls of Troy next to the Trojan Horse.
In other posts, he is hanging out with Leonardo da Vinci while the master works on the Mona Lisa, sailing across the Atlantic with Christopher Columbus, getting chased down by a Tyrannosaurus rex, and roughing it in the American Old West.
The tech behind the fictional traveler
Generating an ancient scene with AI is relatively easy, but keeping the exact same character across dozens of different prompts is notoriously difficult.
To pull this off, the creator used the open-source image model Stable Diffusion paired with a fine-tuning technique known as Dreambooth. Dreambooth allows creators to train an AI model on a specific subject so it can consistently reproduce that person inside newly generated environments.
Instead of training the tool on a real human being, the artist designed a consistent fictional character from scratch. Once the model learned Stelfie's face, the creator could drop him into virtually any era with natural lighting and matching aesthetics.
Why it matters
For digital creators and storytellers, Stelfie is an early blueprint for how generative AI moves beyond one-off images into character-driven narrative.
Traditionally, maintaining a consistent original character required 3D modeling pipelines or skilled illustrators. With Stable Diffusion and Dreambooth, a single person can build, direct, and maintain a consistent virtual influencer without hiring a production studio. It turns AI from an unpredictable sketchpad into a reusable digital actor.
We think this project worked because it leans directly into internet culture. Stelfie does not pretend to be a serious archival discovery; the presence of an iPhone-style selfie arm in ancient Egypt is an intentional joke that sidesteps uncanny valley criticism.
At the same time, it highlights an uncomfortable truth for the creative industry. The exact same technique used to drop a harmless fictional buddy into prehistoric jungles can just as easily be used to place real, living people into compromising or fabricated historical scenes. As personalization tools get sharper, identity provenance is going to become a massive headache.
Sources (2)
- cacm.acm.org(opens in a new tab)Originalcacm.acm.org
- arstechnica.com(opens in a new tab)Reportarstechnica.com