My long years of studying English (especially vocabulary) show two important factors for learning it:
1. Presence of context - sometimes, to remember a new word, it’s enough to encounter it just once, but in a memorable (especially emotionally) setting.
2. Periodic repetition - a fact that has been understood for a long time, and after collectively going through the course Learning How to Learn, it has also been scientifically confirmed. The more often (and with good intervals) a word is encountered, the more likely it is to stay in memory.
So, I’ve had an idea for a while to combine these two approaches using LLM: Analyze user communication (both email and Telegram allow for this) and offer quizzes, questions, and prompts based on the list of words they are currently learning.
Scenario 1: We see that the simple phrase “very exciting” is frequently used, we suggest adding the word “exhilarating” to their vocabulary.
Scenario 2: In various forms, we use LLM to create spaced repetition based on user context (for example, 4 sentences from a client using different words, and they need to choose the correct one).
What do you think?