- MVP version for Telegram (since spamming is a part of their business model, it feels natural to start with it)
- More precisely, data pipeline for weights and measurements for word frequencies. Think of it as small-language models.
- More precisely, it is about morphological analysis of words across different languages. Unlike Meta with regex-based dictionaries[0], I am porting rule-based morphology analysis python library[1] into target programming language.
- More precisely, right now it is about understanding DAWG data structure by porting it from C++[2] to Haskell[3].
- Instead of introducing FFI I wanted to become more comfortable with LLMs, I am trying to approach their internals (or my possibly wrong vision of their internals) by building small language model based on a corpus of thousands of spam messages.
From data modelling point of view it's a challenge to wipe the user data since it will affect a social graph. And there're different strategies to handle corner cases (e.g. how to deal with reactions/replies on "deleted" comments or with reactions on your photos or your reactions on different news, mark as deleted and wipe the content or completely remove nested graph). And it actually makes user tracking much harder (please keep in mind, they're tracking users that have not register yet, in that case user profile might be converted from one user type to another if they are going to continue track you (why didn't want that?)).
It might be much easier to extend account entity with something like:
- Maybe some interaction with 3rd-party websites triggered
cancellation of the process, you thought.
- Then, you'll implement blacklist just to avoid any interaction with facebook, something similar to: https://pastebin.com/FAV2f9eA and try to repeat the flow again.
- Then another 2+ years later situation will repeat again. Deja-vu. And again. And again.
There's no way to delete facebook profile if facebook didn't really care about its users.