DSPy: The End of Prompt Engineering as We Know It
summary
Open-source framework DSPy (short for declarative self-improving python) replaces manual prompt crafting with a declarative program of signatures, reusable modules (retrieval with ColBERT, RAG + chain-of-thought, tool calls), assertions, metrics and examples; optimizers perform few-shot selection, instruction optimization (hill-climbing style edits) and model distillation to smaller models. Typical runs use ~500 training examples but can succeed with 10–20; a red‑teaming case rises from 10% to 44% success after optimization. Model‑agnostic, experimental but demonstrably reduces manual tuning and supports continuous improvement.
AI-written summary · about
chapters (8)
- 0:00DSPy introduction
- 1:07Dark arts of AI prompts
- 6:48What is DSPy
- 12:15How DSPy optimizers work
- 17:44Real-world examples
- 20:13DSPy benefits
- 21:59DSPy limitations
- 23:48Q&A
related talks
01h 13mDSPy: The End of Prompt Engineering
26mDSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines
19mOn Engineering AI Systems that Endure The Bitter Lesson
30mIs Prompt Engineering Dead?
30mIs Prompt Engineering Dead? How Auto-Optimization is Changing the Game
29mIs Prompt Engineering Dead? How Auto-Optimization is Changing the Game
39mOptimizing a Prompt for Production
02h 01mPrompt Engineering and AI Red Teaming