2 ms·Using semantic benchmarks to build a self-improving text-to-query agent3 points by jimmyl02 1mo ago