LLMs are to intelligence as Brazilian butt implants are to butts. Worrying that the first is going to replace the second is a misunderstanding of what these things are.
That would be relevant if the OP were warning about the dangers of LLMs. There's plenty of research going on into machine-learning models that aren't LLMs. Also, the models we think of as LLMs aren't pure LLMs anymore: about 25% of the compute used to train them is spent on reinforcement learning. A lab is free to use any technique it wants in creating or improving a model.