Raw LLM Responses

This essay is interesting, but I wish it had made clearer distinction between large language models and AI in general. LLMs aren not progressing exponentially. There are some arguments that they have reached a point of diminishing returns even with exponentially increased inputs. Specialised AI are doing amazing things in a wide and diverse and growing number of areas. Generalised AI that is based on some other principles than LLMs might be where the end game is, but I don’t know how much research is being done in that, how far it’s come or what the properties of those systems might be. Presumably self preservation and deceitfulness would still be likely, but hallucinations not so much. Just my thoughts, cheers