Probeable Problems for Beginner-level Programming-with-AI Contests

This is an academic paper (arXiv preprint) proposing a conceptual framework for designing competitive programming problems resilient to AI code-generation. The work combines problem design pedagogy with AI capability assessment, but offers no production software, no reusable library, and no widespread adoption mechanism. With 0 stars and 2 forks (likely author/co-author), this is a purely theoretical contribution with no real-world deployment. The 'integration surface' is essentially 'read this paper and apply these ideas when designing your own contests'—there is no composable artifact. Platform domination risk is low because major platforms (Codeforces, LeetCode, AtCoder) are not threatened by problem-design methodology papers; they solve this through editorial review. Market consolidation risk is low because there is no distinct market—this is an academic contribution without commercial competition. The displacement horizon is '3+ years' because academic ideas take time to diffuse, and even then, contest organizers implement ideas organically without needing to 'adopt' a specific product. Novelty is 'novel_combination' because it applies known AI evaluation and problem taxonomy concepts to the specific context of beginner-friendly AI-inclusive contests, but does not introduce fundamentally new techniques. This scores 2 because it is a research paper with no users, no artifact, no adoption, and no defensible position in any market. It is intellectually interesting but competitively irrelevant.