Full threadmaxignol·« LLMs can successfully author CODA kernels » That might speed up progress in this area thenView on HN