Hey Guys!
I recently came across an interesting podcast with our new CEO, Kepano. During the conversation, it became apparent that Kepano was not aware of Napkin’s existence. This surprised me, as I’ve been eagerly anticipating AI integration in Obsidian, especially auto-linking features similar to those in Napkin. However, Kepano announced that Obsidian would NOT venture into AI for another two years.
I use Napkin, an app that leverages AI to link ideas, and it has significantly enhanced my PKM experience. Yet, I still prefer Obsidian for its flexibility. As a result, I supply my data to both Obsidian and Napkin using Readwise sync, relying on Napkin to browse through my Garden until Obsidian introduces its AI. Kepano’s words were disheartening, as I have been blindly dumping data into Obsidian, waiting for the day when AI magically organizes it all at once.
MyMind, another app, is tempting as it offers a seamless web-clipper, auto-indexing, and has announced a graph view similar to Obsidian’s. Its auto-indexing feature uses AI to read through articles and even analyze pictures to identify and tag their contents. Essentially, it combines my Obsidian, Napkin, and Readwise workflows into one and manages to offer much more. The only issue with MyMind is the lack of metadata in its export feature, trapping users’ data within the platform. Moreover, it’s not as flexible as Obsidian. Taking complete ownership of data and privacy is essential to me, and I’m sure many of you were drawn to Obsidian for the same reason.
AI auto-tagging alleviates much of the friction in tracking our learning progress through Obsidian, making it invaluable. It allows us to focus on learning while the app takes care of the rest. Many platforms have already started taking initiatives, and I believe it’s time for Obsidian to embrace AI capabilities as well. If you agree, please leave a like for visibility.
For those concerned about AI integration compromising privacy in Obsidian, keep in mind that Obsidian is built upon strong principles. I’m confident that the Obsidian team can find a workaround to integrate AI without compromising privacy. A potential solution could involve an isolated repository of existing tags that the AI can reference and create similar tags or use existing ones. This way, AI will not have direct access to all the files in your vault, and users can manually summon the AI as needed.
PS: Just to clarify, this request is solely about auto-linking and not AI writing.