Pre-market chit-chat, about my vibe coding

Recently, I deployed a personal service related to "short-term fragmented memory" through AI. It's about those temporary things that can be noted down and forgotten. I use a Telegram Bot to aggregate them into a vector database, and then retrieve the memories through LLM using natural language at any time, instead of the traditional keywords mode. For example, a sample:

This is a manuscript of today's internal seminar at HSBC Private Bank that a friend just sent me. I directly stuffed it in as part of my short-term memory. Generally, I would forget such things after reading them and never recall them again. But with this tool, I wonder how much will be stuffed in after running it for half a year, and what interesting things might be there. What will it answer me then? 😂

$XTALPI(02228.HK) $NVIDIA(NVDA.US)