This text is a transcript of a podcast, specifically an episode titled "How I AI" hosted by Lenny. The episode features two guests: Ryan Nystrom and Claire. Ryan is a software engineer and engineering manager at Notion, while Claire breaks down the biggest announcements from Anthropic's "Code with Claude" event.
The conversation revolves around AI development, workflow automation, and new technologies. Here are some of the key points discussed:
- Automating standup prep: Ryan shares how his team uses a custom Notion AI agent to generate meeting notes from all workstreams, eliminating the need for manual preparation.
- Fast CI is essential: Ryan emphasizes that fast Continuous Integration (CI) is crucial for AI coding velocity, as it allows agents to run more iterations in the same time frame.
- Specs as source of truth: Notion engineers now maintain Markdown spec files in their repo that describe features in plain English with code pointers and verification steps.
- AI defends its technical decisions: Ryan explains how his team makes AI defend its technical decisions, ensuring that the model provides cited reasoning instead of just agreeing with human suggestions.
- Engineering leaders should write code again: Ryan argues that engineering leaders should write code daily to stay close to the work and eliminate meeting prep and information synthesis tasks.
- Good developer experience for humans creates good developer experience for agents: Ryan's team built comprehensive CLI tools, clear documentation, and fast CI long before AI agents, which now helps agents succeed.
The episode also mentions the "Code with Claude" event and its implications for builders shipping AI products today.
Overall, the conversation highlights the importance of workflow automation, AI development, and good developer experience in achieving high productivity and velocity.
