Hacker News Reader

Has anyone thought of using embeddings to solve Little Alchemy? #sample-use

I thought it was the other way around. First little alchemy, then they used an LLM to create a better version of it.

That's possible- I may have assumed incorrectly