I'm not really talking about LLMs specifically. I'm talking about AI as a field. Although neural networks are capable of any form of reasoning or understandin

I'm not really talking about LLMs specifically. I'm talking about AI as a field. Although neural networks are capable of any form of reasoning or understanding and human language is a very strongly diverse field so I do believe LLMs are a strong contender for the production of general AI. Lots of data. Easy to train thanks to text being very simple. Easy to express just about any concept or information. Requires some amount of domain knowledge to make good prediction of what comes next.