Agents are great! But for true ongoing learning, the agents need to actually learn, ie update the model, not just remember. Been thinking on it for a bit, so here's Autotune_GPT, with recursive self-improvement. Every 10 prompts it updates its model! https://t.co/MKLXP3Jglg
— rohit (@krishnanrohit) May 3, 2023
from Twitter https://twitter.com/krishnanrohit
May 03, 2023 at 06:32AM
via IFTTT
No comments:
Post a Comment