Everything below is deployed and public. Source for most of it is on GitHub.
ArtLens
Visual search across ~100,000 public-domain artworks from the Met, the Art Institute of Chicago, and the Cleveland Museum of Art. Upload any image and find its kindred works — CLIP + DINOv2 embeddings with a subject-vs-look blend slider.
Visit Project →PufferGPT
GPT-2 small with every weight stored in turbopuffer and every matrix multiply done as a vector search. Type the start of a text and the model continues it, at about 75 requests and 2 seconds a token.
Visit Project →TinyFn
A REST API with 500+ utility endpoints for common developer tasks. Math, strings, validation, data generation, and more - no dependencies required.
Visit Project →poople-bench
A daily LLM benchmark: one-shot word ladders scored against a provably optimal answer, so the score is exact rather than judged.
Visit Project →S3verless
A powerful serverless framework for building scalable applications with AWS services.
Visit Project →datasetFactory
Generate scaled synthetic datasets for RAG evaluation. Fresh data at any scale without training contamination.
Visit Project →tinybench
Frontier LLMs are trivially wrong on exact tasks. 40 exact-answer tasks across 8 models, measured with and without deterministic MCP tools. Accuracy moves from 83% to 99%.
Visit Project →md.share
Instant markdown sharing without accounts or servers. Content is compressed into the URL itself - just write, share, and anyone with the link can view.
Visit Project →Lobste.rs Graph
Interactive visualization of the Lobste.rs community invite tree. Explore how users are connected through invitations.
Visit Project →