What is the philosophical implication of the current LLM regime? Can we create an alternative writing algorithm to the big tech based LLMs? These are questions that precede the numerical demonstration of the weak/old-ness of the "baseline model" (there should be one, but for later!)

If, for example, we use CNN to experiment with the meaning of perception in the AI agent, is that outdated research because we don't compare it with ChatGPT (which has no sensory input - Please don't confuse it with "multimodal") in terms of numbers?