Very cool! But for hard enough problems, prompt engineering is kind of like hyperparameter tuning. It's only a final (and relatively minor) step after building

Very cool! But for hard enough problems, prompt engineering is kind of like hyperparameter tuning. It's only a final (and relatively minor) step after building up an effective architecture and getting its modules to work together. DSP provides a high-level abstraction for building these architectures—with LMs and search. And it gets the modules working together on your behalf (e.g., it annotates few-shot demonstrations for LM calls automatically). Once you're happy with things, it can compile your DSP program into a tiny LM that's a lot cheaper to work with. https://github.com/stanfordnlp/dsp/ https://github.com/stanfordnlp/dsp/