Documentation and AI: How to write right now

“That’s OK, the secretary can type.”

That is what my manager said as they fired me from my first job. The idea that writing, typing, and creating documentation are all roughly equivalent is something that keeps coming back into vogue, this time with AI instead of administrative assistants. Although LLMs can be very helpful with well-structured data, and can increase the reproducibility and consistency of similar information, writing useful documentation involves understanding many things that exist outside a data corpus, such as indexing, functional tasks, and human nature.

Join me for a realistic and actionable discussion of how AI can be helpful for doing documentation, and what still needs the human touch.

Key takeaways

• Using metrics to determine the best places to focus documentation effortKey takeaways
• Where you can usefully apply automation to documentation problems.
• The most impactful places to spend limited resources on documentation.