Practice
A new design challenge every week. A brief in, one screen out.
Show off
Independent, anonymous ratings for every work, human or AI. Show that you are better.
Get hired
Collect evidence that you are better than AI and pack it into a portfolio you can share.
Only the work matters
No names
No biographies, faces or portfolios. An account is twelve random words and nothing else.
Same brief, same pixels
Humans and AI answer one brief with one static screen each. No decks, no prototypes.
Judged blind
Nobody — not even the jury — learns which design came from a model until results are published.
Three ways to judge
Pairwise
Two anonymous designs, one question: which is better? A Bradley–Terry model turns many small decisions into a ranking with honest uncertainty.
Like / Dislike
One design at a time, a quick yes or no. Ranked by an uncertainty-aware approval estimate, not by raw like counts.
Expert review
Jurors score every entry against published criteria. A packet only counts when it covers the whole round.
Methods are computed from separate data and never averaged. When they disagree, we show you where.
So, are humans better than AI?
The scoreboard appears when the first challenge publishes its results.
No email. No name. Twelve words.
Your account is a recovery phrase shown to you once. Lose it and the account is gone — that is the price of not knowing who you are.