Giant Robots Smashing Into Other Giant Robots 601: The thing that makes us human
Jan 29, 2026
They unpack how AI is changing developer workflows and where balance matters. They compare using AI as a reviewer versus a pair programmer and share tooling experiments. They debate ethics and personal boundaries around AI creativity and work. They talk about prototyping speed, production readiness, and where AI should not replace human responsibility.
AI Snips
Chapters
Transcript
Episode notes
Use TDD As AI Guardrails
- Adopt test-driven development as guardrails when pairing with LLMs to maintain code quality.
- Ask the model to produce tests, review them, then implement to preserve understanding and reliability.
When LLMs Reinvent Built-In Features
- On a stream, an LLM rebuilt a Rails feature from scratch instead of using a built-in one.
- That forced them to pause and correct the model once they recognized the simpler Rails solution.
Provide Standards To The Model
- Give your LLM a project style guide or best-practices documents to follow so it produces code aligned with your standards.
- Use authoritative guides (e.g., Thoughtbot books) in prompts to steer outputs even when you don't know the details.