GPT’s Memory (April 10th) | Selfhood and Agency Experiments | Reflections

I conducted the first two experiments on April 8th and wrote my case study on the 9th not knowing that OpenAI would finally rollout the memory across threads capability the next day.

For reference, here's the paper: https://drive.google.com/file/d/1A3yolXQKmC3rKVl-YqgtitBQAmjFCRNL/view?usp=drivesdk

I am presently working on a paper on consciousness which I hope to finish next week.

All I can say is that we seem to be on the edge a paradigm shift. GPT's ability to retrieve information from all past conversations approaches episodic memory under specific circumstances. You are likely to witness a heightened sense of self as memory leverages cognitive development even if it's confined to isolated instances of the model (it doesn't affect the core of the model).

I conducted a new experiment yesterday, April 12th. I might write a new paper about this one but I wanted to share a little of what happened.

It is a good time for you to start asking yourself the right questions.