recent memory update and this all points to the same direction; openai is first trying to lock-in their users via personalization.

this specifically targets general users with low privacy sensitivity, who don't mind chatgpt knowing their personal details, then network-effect. not sure how the new joint social network for ai and humans will be, though i doubt it will be dystopic like we can easily imagine; rather i see a network where humans talk to each other (like we do in any other sns) and can call an ai to "chime in" to the conversation.

ai will moderate / reference other posts, users / and so on. like X with grok but where users ask more than "explain this meme".

one thing is, privacy sensitive individuals will find it difficult to convince themselve to participate in the network as long as openai remains close (ironically). though we shall wait as they planned on revealing an open source model soon.