Personally, I don’t think LLM should be the focus so soon. Eventually, yeah; but definitely not any time soon. There are other things that should take priority that are actually needed by most users, not an LLM that for most people is just a nice to have.
I highly disagree with this. I did not find out about Notion, Tana, Obsidian, Capacities, Logseq, Roam, Amplenote, Coda, Kortex etc. because of their LLM capabilities (if they even have LLM implemented, which most don’t). I hear about these because of how they cater to the user’s needs via fundamental features amazingly. I would not have cared about Anytype even a single bit extra if it had LLM, I came to Anytype because of it’s privacy centric approach and built like Notion. Anytype having LLM will not make me any more excited about it.
If you want to see what actually turn heads toward an app (that has nothing to do with AI), Obsidian releasing their Canvas feature does just that. That is a good example. But you get there by first building the basic features which Anytype still hasn’t fully achieve.
Without link to blocks, without set collection functions, without synced blocks, without inline set collection view, dynamic dates, notifications, tabs, etc., there will be less people importing all their notes into Anytype. And if there aren’t as many people importing all their notes into Anytype, there will be less use case of an LLM. I truly just don’t think LLM is a priority at this point. I don’t think Obsidian even have an LLM (correct me if I’m wrong) when they have way more users than Anytype with more notes.
Focus on the features that actually cater to most of the current users better and makes people want to import all notes into Anytype first. If someone doesn’t think Anytype is good enough at the moment to decide importing all notes in (and I’d say there’s a big chunk of this users), then an LLM will have no purpose if there are no notes to based off of.