NEW POST
AI editors generate code with remarkable speed, but more is needed for robust systems. Unmesh Joshi describes a dialog between developer and LLM - demonstrating how expert guidance is critical.
https://martinfowler.com/articles/exploring-gen-ai/14-guiding-llm-java-bytebuffer.html
AI editors generate code with remarkable speed, but more is needed for robust systems. Unmesh Joshi describes a dialog between developer and LLM - demonstrating how expert guidance is critical.
https://martinfowler.com/articles/exploring-gen-ai/14-guiding-llm-java-bytebuffer.html
Comments
I've had mixed results with having the llm look back at it's own code. A lot of the time is just works, it notices problems and fixes them. Other times it doesn't.
As some people like to say though, this is the worst LLMs will ever be.
The more you know, the more tools and the more power you have at your disposal is still true.