AI Librarian

Patrick Rachford
3 min readApr 1, 2023

A new push for the technical writing world

The article discusses the changing role of Technical Writers and how they need to adapt to new technologies and tools like AI.

Technical Writers have traditionally written and maintained documentation, meeting developers where they’re at, and providing documentation through various channels such as Web UI prompts, --help messages, and product documentation.

This has allowed developers to use search on the documentation website to find answers or use sites like Stack Overflow to ask questions and have the community respond with links to clarifiy details and answers.

However, with the advent of AI-powered tools like ChatGPT, developers are no longer searching for solutions on Stack Overflow or using Google, but instead, asking for direct answers to their questions with generative AI models like ChatGPT.

To keep up with these changes, Technical Writers need to start writing documentation that can be easily consumed by large language models like ChatGPT.

Technical Writers should start writing their documentation to be easily consumed by large language models like ChatGPT

Technical writers should start writing their documentation to be easily consumed by large language models. This involves testing out prompts and reworking answers to provide world-class documentation that can be readily used by AI models.

Companies like Supabase and Stripe have already trained their documentation to be used by generative AI models like ChatGPT.
This makes the developer experience easier as it removes clicks and saves time, as developers just want to get stuff done.

Technical Writers can train AI models on their engineered prompts

In the not to distance future, the role of the Technical Writer may evolve into that of an AI librarian.

Where Technical Writers train AI models to understand the needs of developers for a company’s product. The AI librarian will collaborate with data scientists and machine learning engineers to ensure that the AI models are accurate and effective in providing requested information.

This will result in a lower bar for getting started on many projects, with developers being able to ask their favorite SDK to set up a client, create a connection, and perform some business logic without having to sit through pages of tutorials.

Simon Wilson had a great post on this very developer workflow, on how “AI-enhanced development makes me more ambitious with my projects”.

ChatGPT (and GitHub Copilot) save me an enormous amount of “figuring things out” time. For everything from writing a for loop in Bash to remembering how to make a cross-domain CORS request in JavaScript.

Wilson emphasizes that for developers, the biggest win is lowering the bar on getting started for many projects.

The role of the Technical Writer is about to change into a custodian or a librarian for AI-powered tools. The role of the AI librarian will involve training AI models to understand the needs of developers for a companies’ product, similar to their role now; but with the understanding of feeding information to the generative AI.

AI librarians will collaborate with data scientists and machine learning engineers to ensure that the AI models are accurate and effective at providing requested information.

AI is here to stay, and the role of the AI librarian is to leverage AI to enhance the developer experience.

--

--

Patrick Rachford

Currently at Temporal, previously at AWS. I enjoy documenting my thoughts and sharing ideas with others.