We might see another AI winter first, is my assumption. I believe that LLMs are fundamentally the wrong approach to AGI, and that bubble is going to burst until we have a better methodology for AGI.

Unfortunately, the major players seem focused on 'getting to AGI pretention through LLM'.