You’re scrolling through your social feed at 2 AM, and once again, the algorithm has decided you need to see engagement bait and viral drama instead of the niche robotics discussions you actually care about. You close the app, frustrated. What if you could just… build your own feed algorithm?
That’s exactly what Bluesky is betting on with Attie, their new app for creating custom feeds. And as someone who spends their days building bots and wrestling with APIs, I’m genuinely excited about what this means for the future of social platforms.
The Feed Problem Nobody Talks About
Here’s what most people don’t realize: every social media feed is a bot. It’s an algorithm making thousands of micro-decisions about what you see and when you see it. The difference is that until now, you’ve had zero control over how that bot thinks.
Bluesky’s approach with Attie flips this model. Instead of one monolithic algorithm serving billions of users, they’re enabling anyone to create custom feed algorithms. Want a feed that only shows posts with code snippets? Build it. Want to filter out all political content after 8 PM? You can do that too.
This isn’t just a feature—it’s a fundamental rethinking of how social platforms work.
What Attie Actually Does
From a technical standpoint, Attie is a feed generator builder. You’re essentially creating a bot that crawls Bluesky’s firehose of posts and applies your custom logic to decide what makes it into your feed.
The beauty is in the accessibility. You don’t need to be a developer to use it, but if you are one, you can get as technical as you want. Think of it as the difference between using IFTTT’s visual interface versus writing your own webhook handlers—both work, but they serve different skill levels.
For bot builders like us, this opens up fascinating possibilities. We can experiment with different ranking algorithms, test content filtering strategies, and even build feeds that respond to time of day, trending topics, or user engagement patterns.
Why This Matters for Bot Architecture
What Bluesky is doing with Attie mirrors a trend I’ve been watching in bot development: the shift from monolithic systems to modular, user-controlled components.
Traditional social platforms treat their algorithm as a black box—proprietary, unchangeable, and optimized for engagement metrics that may not align with user wellbeing. Attie treats the algorithm as a tool that users can modify, fork, and share.
This is the same philosophy driving modern bot frameworks. Instead of building one massive bot that tries to do everything, we’re moving toward smaller, specialized bots that users can combine and customize. Attie is applying this principle to social feeds.
The Technical Implications
For developers, Attie represents a real-world testing ground for feed algorithms. You can prototype an idea, deploy it, and get immediate feedback from actual users. This is huge for anyone working on recommendation systems or content filtering.
The platform also forces you to think about edge cases. What happens when your feed algorithm encounters a post type you didn’t anticipate? How do you handle rate limiting? What’s your fallback when the firehose is moving too fast?
These are the same questions we grapple with when building any bot that processes real-time data. Attie just makes the feedback loop much tighter.
The Bigger Picture
Bluesky’s move with Attie is part of a larger conversation about AI and user agency. Recent news about the dangers of asking AI chatbots for personal advice highlights a real concern: when AI systems make decisions for us, we need to understand and control those decisions.
Attie doesn’t solve every problem with algorithmic feeds, but it does something important—it makes the algorithm visible and modifiable. You’re not just consuming what an AI decides to show you; you’re actively participating in how that AI thinks.
For those of us building bots and AI systems, this should be the standard. Transparency isn’t just good ethics—it’s good design. When users understand how a system works, they can use it more effectively and trust it more deeply.
What’s Next
I’m planning to build a few experimental feeds with Attie and document the process on ai7bot.com. I want to see how far we can push the customization, what the performance limits are, and whether we can create feeds that genuinely improve the social media experience.
The real test will be whether other platforms follow Bluesky’s lead. If they do, we might be looking at a future where every user has their own personalized algorithm—not because a company decided what’s best for them, but because they built it themselves.
That’s a future worth building toward.
đź•’ Published: