The Vergecast The future of AI might look a lot like Twitter
68 snips
Sep 29, 2024 Michael Sayman, the innovative creator of the SocialAI app, discusses his unique approach to social networking, where users interact solely with AI. He explains why posting to a language model could be more valuable than traditional social media. The conversation dives into the evolution of AI-driven platforms, the blending of human and AI experiences, and the appeal of personality-driven bots. Sayman also shares intriguing stories about unexpected uses of the app, illustrating the transformative potential of social AI.
AI Snips
Chapters
Transcript
Episode notes
First Attempt at SocialAI
- Michael Sayman first tried building SocialAI five years ago but the technology wasn't ready.
- The earlier version felt more like a game than a social network.
Limitations of Early Language Models
- Early language models were too robotic and conditional to simulate a social network convincingly.
- The technology felt like a game, reminding users of its artificiality, unlike current language models.
From Facebook to SocialAI
- Sayman's journey in tech began when Mark Zuckerberg offered him a job at Facebook at age 16.
- He witnessed the evolution of social media, from a tool for connection to a data-driven simulation of human interaction.