Meta Unveils 'Friend Bubbles' for Reels: The Engineering Behind a ‘Simple’ Social Discovery Feature
Meta Unveils 'Friend Bubbles' for Reels: The Engineering Behind a ‘Simple’ Social Discovery Feature
In a major update to its short-form video platform, Meta has launched 'Friend Bubbles' on Facebook Reels. This new feature surfaces Reels that friends have watched and reacted to, but its simplicity belies complex engineering.

"On its face, Friend Bubbles looks simple enough, but sometimes the most straightforward features require the deepest engineering work," said Pascal Hartig, host of the Meta Tech Podcast, in an exclusive interview with the engineering leads behind the feature.
The feature marks a significant step in social discovery, aiming to connect users through shared video experiences. Meta engineers Subasree and Joseph detailed the challenges in a recent podcast episode.
Engineering Breakthrough: Machine Learning and Cross-Platform Behavior
The team revealed the evolution of the machine learning model that powers Friend Bubbles. The model had to account for different behaviors between iOS and Android users. "We found that engagement patterns varied dramatically between the two operating systems," Joseph explained.
A surprising discovery finally made the whole feature click. "The breakthrough came when we realized that the model needed to prioritize recency and mutual engagement over raw watch time," Subasree said.
The feature scales to billions of Reels watched daily, requiring real-time processing of friend interactions. "It's not just about showing what friends watched; it's about showing what's relevant and timely," Joseph added.
Background
Friend Bubbles is part of Facebook Reels, Meta's answer to TikTok. The social media giant has been investing heavily in video discovery. The feature initially launched in limited testing in early 2023. The engineering team faced challenges in making the social signal useful without being intrusive.

The Meta Tech Podcast episode, released today, gives an insider look at the development process. Hartig, a software engineer himself, probed the team on technical trade-offs. The episode is available on Spotify, Apple Podcasts, and Pocket Casts.
What This Means
For users, Friend Bubbles transforms Reels from a solo experience into a shared one. It could increase time spent on the platform by making discovery more social. For Meta, it's a competitive differentiator—blending social graph data with algorithmic recommendations.
"We think this creates a more connected experience," Subasree said. "When you see a friend's reaction, it adds a layer of trust and curiosity." Early tests show higher engagement among friend groups.
The feature also raises privacy considerations. The team emphasized that users can opt out of showing their activity. "We built in controls from day one," Joseph noted.
Where to Listen
The full podcast episode titled "Reel Friends: Building Social Discovery that Scales to Billions" is now streaming. Find it on Spotify, Apple Podcasts, or Pocket Casts.
Meta is accepting applications for engineers interested in working on such features. Visit the Meta Careers page for opportunities.
Related Articles
- Apple's Upcoming Lineup: MacBook Ultra, Foldable iPhone Ultra, and Vision Pro's Decline
- Anthropic Brings Claude’s Persistent Context to Outlook, Word, Excel, and PowerPoint
- How to Demand Transparent Edtech Vetting for Your Child's School
- Modeling Complex Systems Made Easy: A Guide to HASH Simulations
- 10 Key Insights from May 2026's Patch Tuesday: AI's Dual Role and Record Fixes
- Kubernetes SELinux Mount Optimization: What v1.36 Means for Your Cluster
- How to Enable User Namespaces in Kubernetes v1.36 for Enhanced Container Security
- The USB Drop Heard Around the Security World: How a Simple Pen Test Became Legendary