Remnote, a note-taking and spaced repetition app, has possibly the best integration and application of AI I've experienced, on par with GitHub Copilot:

https://www.remnote.com/

I'm a happily paying subscriber. Totally recommend it. They don't use AI simply for its own sake, as a primary selling point, but as a tool to make the base product better -- I'd certainly be using it (and often do) without the AI integration anyway.

Just a few examples in my case:

- I'm studying spanish. During tutoring sessions, I can create a flashcard for a new word or phrase by typing in the editor the spanish word, followed by ==, followed by tab, to generate an english translation; or vice versa, from english to spanish. I can make it double-sided by typing ==< instead of ==. I get much more out of my tutoring sessions since it's so easy to create these flashcards, which are scheduled at appropriate intervals. I could even do mass translation/generation of cards, but I prefer to select the terms I want to learn on my own as they occur in my lessons and studies.

- I'm studying ML papers. I can upload a PDF, or just a link to a PDF, and it's stored. I can highlight, annotate, make notes about it. I can generate summaries (if I'd like, though I prefer not to), or ask the integrated AI for more information about parts of the paper, or to explain things I don't understand.

- I make math flashcards for certain concepts and equations. I can write something like "partial derivative of the cross-entropy loss function == $$", then press tab, and it generates and renders a latex representation of the concept. I can very easily create "fill in the blank" (called "cloze deletions") spaces, even in the latex in the equation.

I also recommend the youtube videos as well. The release notes are really well done, and they develop new features extremely quickly and with great quality: https://www.youtube.com/@RemNote

Just an amazing tool perfectly augmented by AI. (I swear Remnote isn't paying me or rewarding me for this -- just a huge supporter.)

You could try my project Plandex (https://github.com/plandex-ai/plandex) — you can use it to explore/understand/chat with just about any codebase, including massive ones up to 20M tokens.

Here's an example, chatting with the SQLite codebase to understand how transactions are implemented: https://plandex.ai/_next/static/media/plandex-sqlite.0ee6cb2...

It’s not really a specific tool, but I came across another comment on HN where someone used an LLM to generate Anki cards from textbooks. I’ve been doing this with Claude Sonnet 3.7 and it’s been very effective.

[dead]