Bluesky launches Attie, an AI app for custom feeds. This tech bet arrives as social platforms seek new business models.

The Big Picture Social media faces regulatory and user pressure over opaque algorithms. Twitter-now-X and Meta have faced criticism for their recommendation systems. Bluesky, built on the open AT Protocol, always promised transparency. Attie represents the next phase: giving users complete control.

The app uses Anthropic's Claude model, one of the leading AI companies. Users can request feeds like "posts about folklore, mythology, and traditional music, especially Celtic traditions." Initially confined to a standalone app, the plan is to make them available in Bluesky and other atproto apps.

“A bet on returning algorithmic control to users.”

Why It Matters This move comes at a crucial time for social media economics. Traditional platforms monetize through advertising driven by proprietary algorithms. Attie challenges that model by putting personalization in user hands. If successful, it could create a new paradigm where value lies not in the central algorithm, but in tools to build it.

For investors, this represents a bet on social tech's future. Bluesky hasn't disclosed user or revenue numbers, but its focus on open protocols and now personalized AI marks a clear divergence from established giants. In a market where Meta dominates with closed models, openness could be the competitive edge.

The integration with the AT Protocol is key. It allows any developer to build on the same infrastructure. This could accelerate innovation in content discovery tools, creating an ecosystem around atproto. In 2026, as AI and social media regulation intensifies, algorithmic transparency might become requirement, not option.

The Bottom Line Watch how Attie adoption evolves and whether other developers create similar tools on atproto. Success will depend on whether users actually want to control their algorithms, or prefer the convenience of automated systems. For Bluesky, this is its boldest bet yet: competing not with more content, but with better discovery.