I've been thinking about AI-powered auto organization recently. I'm not a fan of fully automated categorization that requires no effort from the user. It often results in generic tags/categories that LACK YOUR PERSONALITY.
Instead, what if users still had to create folders (collections) and manually place bookmarks in them?
Over time, the AI would learn the user's preferences and criteria for organizing their stuff. If the AI makes a mistake, the user can simply move the bookmark to the correct folder, further teaching the AI.
I've made some initial attempts with my own bookmarks, and 8 out of 10 were placed in the correct folder automatically (screenshot below).
Here's how I envision it working for new users:
1. Register for a Raindrop account.
2. a. Import your existing bookmarks/folders, allowing the AI to understand your organization style.
b. Alternatively, start from scratch by creating folders and adding bookmarks. The AI will quickly catch up.
3. When you add a bookmark, Raindrop will automatically place it in the most suitable folder or in an "Unsorted" folder.
4. Later on, you can visit the special "Organize" page where the AI will provide advice on how to reorganize your collection, including suggestions for the bookmarks in the "Unsorted" folder.
The same process applies to existing users. As you can see, this solution requires some effort from the user, such as occasionally changing the auto-selected collection and performing reorganization.
What are your thoughts? Do you think this is the right approach?