Answers are grounded in this paper and cite their sources.
Two lower bounds for the waste W (uncovered area) of a packing of unit squares in a region {0 ≤ x ≤ L, 0 ≤ y ≤ w + ζx} whose ceiling rises by R = ζL ≥ 1. For every slope, W ≥ 4·10⁻⁵·L/√M with M = w + R. For slope at most 1, W ≥ 7·10⁻⁶·√(LR) whatever the width w. The second is Conjecture 1 of H. D. Bui, "Square packing with O(x^0.6) wasted area" (arXiv:2508.04603), with an explicit constant. The first comes from an inequality for regions under any concave ceiling, which also gives W ≥ 5·10⁻⁴·L/√(m + ½) for an L × (m + ½) rectangle, m ≥ 125. Given the decompositions described in the papers of Bui, McClenagan and Ryu, the trapezoid fillers in their constructions are optimal up to constants, so those frames cannot improve the exponents 3/5 for the waste of a square and 3/8 for c*(k). Nothing is claimed about packings that do not split into such pieces. Status: the proofs were found and written by Claude (Anthropic) working as a research agent under the author's direction. They have been reviewed only by language models (six blind passes, Claude Opus and OpenAI Codex) and tested by floating-point simulation. No person has checked them line by line and nothing is formalised. Please read the theorems as claims. Simulators and review reports: https://github.com/TheSnakeFang/squarepack-certs, folder wedge/.
No takes yet. Share an insight, caveat, or question.
Kevin Fang (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: