Mission statement
Teaching a machine what a good wing looks like
Humanity has mapped the genome, put a rover on Mars, and still cannot agree on whether a wing should be wet or dry. The Large Wing Model exists to settle it with data.
The LWM is a wing-rating model in training. It looks at a photo, forms an embarrassingly confident opinion across four axes, and then hands that opinion to you for correction. Your correction — not the model's guess — is the label we keep.
Every wing you rate makes the model marginally less wrong. Thousands of corrections later, the goal is a model that can look at a plate of wings and tell you, with a straight face, whether they were worth the drive.
01
Upload
A photo of a real chicken wing. Any sauce. Any state.
02
Be judged
The model guesses crispiness, sauce, meat and overall from the photo.
03
Correct it
Your scores overwrite its ego and become training data.
The plain-language data notice
Photos you upload and the scores you give are stored and used as training data for the Large Wing Model. Ratings appear publicly on the leaderboard. No account, name, or email is collected. Do not upload photos containing people, documents, or anything you would not want a wing researcher to see.