Agent prompts, clarification loops, and human oversight

It's always been challenging to make reliable software - is AI making it worse or better? Carl and Richard talk to Damien Brady about his experiences building software with AI tools and trying to bring that software up to an acceptable standard. Damien talks about leveraging LLMs' tendency toward detailed analysis to catch problems in code. It takes practice and experience to get good at using these tools, but they become more powerful over time!