Because who has the time ? (also part 1 because it took me further than I expected 😬)
Couchbase recently introduced support for Vector Search. And I have been looking for an excuse to play with it. As it turns out there was recently a great Twitter thread about Developer Marketing. I can relate to most of what’s in there. It’s a fantastic thread. I could summarize it to make sure my teammates can get the best out of it in a short time. Like, I could write that summary manually. Or that could be the excuse I was looking for.
Let’s ask an LLM, Large Language Model, to summarize this brilliant thread for me, and for the benefit of others. In theory, things should go as follow:
-
- Getting the tweets
- Transforming them in vectors thanks to a LLM
- Storing the tweet and vectors in Couchbase
- Creating an index to query them
- Ask something to the LLM
- Transform that …