Do you think Reinforcement Learning still got it? [D]

Recently I've heard many people saying reinforcement learning itself hasn't shown any improvement in many years (maybe alphago was the last big thing).

Whereas other field of AI has seen many SOTA architectures like 'Transformers' for Sequence based tasks and 'ResNet', 'Diffusers' & 'VAE' like architectures for Computer vision tasks.

Thought I do believe, directly or indirectly, reinforcement learning still playing a crucial role behind LLMs like ChatGPT and Claude using 'RLHF' techniques. And in many other recent technologies including …

alphago architectures big computer computer vision diffusers improvement machinelearning people reinforcement reinforcement learning resnet sota tasks think thought transformers vae vision