Capture before classification.
Prototype a spoken object that remains useful without allowing the transcript to replace the source audio.
Open brief →Challenge environment
Pick a problem.
These are bounded slices of real ThumbThink concepts. Choose the problem that best demonstrates how you think and build.
Disclosure boundary in review.
The next problem will open after its public challenge slice is separated from protected project architecture.
Request notice →Third slot reserved.
The challenge environment stays small enough to review seriously. New projects replace completed or paused challenges.
Enter roster →Challenge standard
Artifact first. Explanation second.
We look for comprehension, judgment, architecture, execution, communication, resourcefulness, intelligent AI leverage, and integrity. Fast code alone is not the test.
Substantive build auditions that create usable product value should be compensated or separately agreed before work begins. Challenge submissions remain the builder's work unless a written agreement says otherwise.