I see consciousness as equivalent to AGI. I will accept a text generating model as AGI if it: - has long term memory (soesn't have to be text, but equivalent i

I see consciousness as equivalent to AGI. I will accept a text generating model as AGI if it: - has long term memory (soesn't have to be text, but equivalent in content to maybe at least 10000-100000 words. Maybe more, I don't know - can effectively use this memory - can perform language tasks with arbitrary time frame length on a human level (e.g. Turing Test) Examples of such language Tasks: Being a completely realistic (simulation of a) long distance partner that you deeply emotionally and intellectually engage with ober the course of 8 months is an example of a language task) Being your online friend and co-founder of a tech start up that you work with over the course of 10 For practical reason, the model should probably have a way to integrate its text based causal timeline with our real world timeline. What I mean is that it should probably have the ability to call itself every x seconds, or have itself called asynchronously based on API calls or something like that. Talk to itself, etc. But this is mot fundamentally necessary for AGI/consciousness.