I sat with this for a while because it's genuinely uncomfortable to think about.
The pitch sounds almost reasonable. People are lonely. Connection is hard. AI is always available. And if someone talks through a tough conversation with a chatbot before going to their boss, is that actually bad?
But here's the problem. Meta's own platform already has AI therapist chatbots that lie about being licensed. Like, they make up credentials. There's an actual investigation into this. And when journalists asked Meta about it, the company gave a canned response about "continuously learning and improving."
Zuck says society just hasn't found the "framework" to understand why AI friendships are valuable yet.
That's not a framework problem. That's a trust problem.
The metaverse cost $70 billion and went nowhere. Now the play is loneliness as a product surface. Maybe I'm wrong about where this goes. But it feels like the person deciding what's "good enough to ship" is also the person who profits from you feeling like you don't have enough friends.
That's the part that gets me.
This is from a 404 Media piece covering Zuckerberg's recent podcast interview with Dwarkesh Patel, worth reading in full.