About this talk
I don't like updating documentation. Most dev teams I have worked with have outdated docs. The documentation was correct when someone wrote it. Then the code changed, and the docs did not. A year later, nobody really trusts it anymore. Humans and agents alike try to read it and get confused.
We have been using the LLM wiki pattern to solve this.
You do not write the documentation pages yourself. You write instructions for the documentation: what kinds of pages you want, who reads them, and how the pages should link to each other. Then you give your data and the instructions to an LLM. Every time the code changes, the documentation that depends on that code gets rebuilt. So, documentation is no longer a separate manual task. It gets rebuilt automatically, as tests run automatically.
In this talk, I explain what the LLM wiki pattern is and how to get started, and show a real-world example of how we solved the age-old problem of project documentation.
Speaker
Surprisingly Human AI Engineer.
Timo Tuononen works as a Senior AI Engineer at Nitor. He has spent the last decade in the IT industry, both building software and leading teams. With the AI shift, he moved to building and managing AI agents and agentic solutions. On the human side, he has surrounded himself with psychologists for the past 10 years, in relationships ranging from marriage to co-founders.