PIATRA . INSTITUTE
prompts
the picobreakthrough prize
You don't have to solve the problem. Move it.
Piatra Prompts awards prizes for small but genuine advances in mathematics, science, engineering, and scholarship made through conversations with AI.
picobreakthrough
a small, novel, verifiable advance beyond what was previously known
A new counterexample. A tightened bound. A simplified proof. An error found in published data. A result reproduced for the first time. A useful conjecture. Research progress is usually not solving the problem, it is moving the frontier by an almost absurdly small, checkable amount.
the submission
We do not award prizes for getting an AI to say something impressive. We award prizes for discovering something true. The conversation is your laboratory notebook; every submission contains five things:
question · what were you trying to discover?
conversation · the complete model conversation, exported, not paraphrased
result · what exactly did you discover?
verification · independent evidence that the result is correct
reflection · what did you do, what did the model do, where did the crucial insight occur?
the standard
Is it new? Is it true? Can you show how you got there?
Model agreement is not independent verification: proofs are checked, code is run, sources are read. Finalists defend their result in a short oral discussion: if you cannot explain it without the AI, it is not yet yours.
the prizes
Piatra Signal · $100–200
a small, novel, verified observation
Piatra Result · $250–500
a substantial result with strong independent verification
Piatra Breakthrough · $1,000+
an unusually important result
Piatra Negative Result · $100–250
compelling evidence that a plausible approach cannot work
The jury may award several prizes, one, or none. There is no minimum breakthrough size. Entry is free, any model is allowed with disclosure, and winning conversations are published in a permanent open archive.
how it works
1 · Submit a short claim: the question, the result, why it might be novel, the evidence, the conversation.
2 · Every claim is read. Promising ones are invited to the full review: research note, transcript archive, verification package.
3 · Finalists defend the result in a short oral discussion; prizes follow.
2 · Every claim is read. Promising ones are invited to the full review: research note, transcript archive, verification package.
3 · Finalists defend the result in a short oral discussion; prizes follow.