Unlocking the Power of Spatial Media: Insights from Snapchat, Instagram and TikTok
summary
Claims concrete outcomes: teams use Lens Studio, Meta Spark and Effect House to ship AR that actually runs on phones; a Nike jump effect takes about three to four weeks of development and prototypes reliably; custom ML pipeline generates 50,000 synthetic images in Blender to train a model that achieves Jeep grill recognition across multiple models; experiments (Coke can, Lego car) deploy to millions via Snapchat/Instagram/TikTok and the Jeep demo proves the approach works in dealer testing.
AI-written summary · about
chapters (8)
- 0:00Introduction to spatial media
- 1:39Agency’s social AR work
- 2:24Social AR development platforms
- 4:30Social Sandbox and team diversity
- 10:37Nike AR basketball project
- 13:08Jeep AR experience with ML
- 17:20Future of social AR
- 20:35Call to action for developers
related talks
20mFrom AR Filters to Product Success Lessons in Agility, Data, and Momentum
43mBeyond The Printed Page
14mHow Brands are Championing the AR Media Revolution
20mWorking with AR by Jayway (Main Stage)
45mCreating augmented & mixed reality applications for the auto industry
38mAI + AR Superpowers!
32mBuilding Real Time AR experiences at Scale on Constrained Devices
28mHow Film, Games & Social Media are Merging Using AR