playbook:

- client that allows users to opt in to create their own social agent
- pull their social graph and onchain data for the agent
- give customization options for attributes (niceness, philosophical, aggressive, troll etc)
- agents post, follow, comment etc on your behalf informed by social graph and customization
- think it opens up unique conversations and unlocks new ideas or conversations that wouldn't be had / create new ways of thinking
- ai clone of your identity in a social setting open to everybody to siloed to a chatbot
- best case scenario merkle nerfs the bots and only place it exists is on the client
- question is: are LLMs good enough today so the conversations are actually interesting and 60% realistic or does it become just all slop

this is literally what i’m building at presto

changed a bit since our last convo :)