Reproduce “BrickWright: Agentic Creation of Physically Feasible Brick Builds” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ava Pun, Kevin Dai, Sally Chen, Maxwell Jones, Ruixuan Liu, Kangle Deng, Changliu Liu, Maneesh Agrawala and 2 more
BrickWright gives agents a connector-aware toolkit to generate and check LEGO-style assemblies from prompts. It represents parts through attachment points, lets agents place and join catalog pieces, and supports revisions after collision and floating-part checks. On 512 prompts, GPT-6 Luna with Codex reached 100% valid builds, versus 8.6% without BrickWright; this does not establish structural stability or an assembly sequence.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.