Just for fun · AI experiment
Same palms. Two models. A little curiosity.
Palm Perspectives is a playful experiment in AI-generated palm readings, and in how two models turn the same input into different stories.
Try Palm Perspectives(opens in a new tab)Context
I used AI-assisted development to explore a product idea from input to report: independent readings, a comparison view, and choices about what to save or share. The interesting design question is how to make the experience engaging while keeping its uncertainty visible.
The problem
A confident story can feel like a fact. In a playful experience, the interface still needs to make clear what is observed, what is interpretation, and what the system cannot know.
The design decision
Keep the readings independent before comparing them. Separate visible features from symbolic interpretation, and make uncertainty part of the report rather than hiding it behind confident language.
- Provide the same image inputs to two models for independent responses.
- Organize each response into a consistent report structure.
- Compare agreements and differences without treating agreement as proof.
- Give users control over which report content they share.
- Keep the entertainment framing visible throughout the experience.
Explore the example
Just for fun · Illustrative report layout
These invented snippets show how a comparison could read. They are not real readings, model outputs, or claims about a person.
Interpretation A
“A story about exploring new paths.”
Interpretation B
“A story about finding a steady rhythm.”
The useful comparison: different narratives can emerge from the same input. Neither interpretation establishes something true about the reader.
Reflection prompt: which story sparks a thought for you, and which would you happily ignore?
What the work demonstrates
The implementation combines independent model responses, structured reports, and selective sharing, with share snapshots that omit photos and account identifiers.
Scope and limits
Just for fun. This project offers imaginative reflection, not factual assessment or prediction. The example below is editorial copy written to explain the interface, not output from an actual model run. This case-study page collects no photos and makes no AI calls.