Post by @kennu • Hey

I realized today that LLMs (like ChatGPT) are much closer to human behavior if you don't use them only as "snapshots" but also attach a continuous learning