Wikis/docs have been painfully static for ages. Keeping them updated is a huge pain and they're hardly ever in line with what's actually happening. But now, with these LLMs, we're finally getting somewhere. They're starting to get clever in ways that may actually help. From what I've gathered from the blog post about this new version, here's the cool stuff it does:

- Smart Addition : Adds bits of info (snippets) by pulling from various tools.

- Smart Modif : Figures out (insights) and fixes contradictions or repeats in the info.

- Smart Search : Lets you search using plain language.

Yes LLMs are an amazing technological breakthrough for knowledge base and document solutions.

You have a good understanding of what we are doing :)

With snippets:

* You can save Slack thread as a new document in GitBook; the document will be generated by a LLM (it's not a plain dump of the messages).

* You can record your work on VS Code (+ audio), and we'll translate it as a document (snippet)

With Insights:

* We analyze content (we have customers with more than 10000 pages in their docs) and identify contradictions or duplicate

* We don't auto-fix them yet, but it's planned ;)

With AI Search:

* We leverage LLM and a vector database to provide complete answers with sources to natural language questions