And they still are. OpenAI consists of two parts; a non-profit entity which owns the IP, along with the obvious commercialization-focused subsidiary of the company.

My question is: what was stopping both parties here from pursuing parallel paths? — have the non-profit/research oriented arm continue to focus on solving AGI, backed by the funds raised on from their LLM offerings? Were potential roadmaps really that divergent?

I had always assumed this was their internal understanding up until now, since at least the introduction of ChatGPT subscriptions.