LLMS AS UNBIASED ORACLES

Summary from the TLDR newsletter

Large language models can be used in automated black box testing, not just as a post-hoc verification step, but as a foundational driver for the code generation process itself.