@eladgil: https://x.com/eladgil/status/2079561730263531771
Summary
Cursor's team of AI agents reconstructed SQLite from its manual into Rust, passing all tests, with costs varying 15x based on model mix.
View Cached Full Text
Cached at: 07/21/26, 04:48 PM
👀
Cursor (@cursor_ai): We had a team of agents rebuild SQLite from its 835-page manual.
It created a replica in Rust which passed 100% of a held-out test suite.
Interestingly, cost varied 15x depending on which model mix we used.
Similar Articles
@omarsar0: Recommended reading. (bookmark it) Pay attention to the prices and what a combination of models can unlock for you. You…
A thread discussing Cursor's experiment where a team of AI agents rebuilt SQLite from its manual in Rust, achieving 100% test pass rate with significant cost variation depending on model mix. Takeaways include using frontier models for decomposition and cheaper workers for implementation.
@rohanpaul_ai: Making SQLite 5% faster would already be powerful. Here, an AI agent, KISS Sorcar, got 59% improvement in under 8 hours…
An AI agent called KISS Sorcar achieved a verified 1.59x geometric-mean speedup on SQLite across four benchmarks in under 8 hours for less than $150, passing over 1 million SQLite tests. This highlights the growing capability of coding agents.
sparqlite: SPARQL 1.1 inside SQLite, written in Rust
sparqlite is a Rust-based SQLite loadable extension that stores RDF quads in ordinary SQLite tables and natively executes SPARQL 1.1 queries, achieving 3.5× speedups over SQL translation while passing 636 of 639 W3C evaluation tests — notably generated by Claude Code's Ultracode agentic workflow over a 12-hour run.
Agent swarms and the new model economics
Cursor's new agent swarm design uses planner and worker models to decompose tasks into a tree structure, achieving significant cost savings and quality improvements. In a test rebuilding SQLite from scratch in Rust, the new swarm reached 80% pass rate in four hours while the old swarm failed.
@ericzakariasson: https://x.com/ericzakariasson/status/2062199026544787576
Cursor introduces Debug Mode, which lets the AI agent instrument code, collect runtime logs, and fix bugs based on actual evidence instead of guessing, improving accuracy and handling hard-to-reproduce issues.