Building the best maps in the world: A conversation with leaders from Mapbox and Tripadvisor
summary
Mapbox ships procedural 3D building generation that creates roofs, facades and entrances at runtime using ML and sensor-derived road data, and a Navigation SDK now running in BMW and DoorDash fleets. Public releases include MCP Server, MCP DevKit and a Location Agent for conversational map reasoning, plus a feedback agent and tileset POI API for faster loading. Tripadvisor re-styles Mapbox Standard with interactive pins, a collision algorithm and Nearby, driving 70% lift in user engagement and enabling broader gains like 88% fleet efficiency mentions.
AI-written summary · about
chapters (11)
- 0:00Welcome and introduction
- 2:16Keynote kickoff
- 3:26Evolution of maps today
- 10:05Challenges in map building
- 17:13New AI tools for developers
- 22:53Mapbox Standard platform
- 25:17Tripadvisor collaboration
- 28:19Building best travel map
- 34:13Key features and impact
- 44:40AI assistant for travelers
- 56:19Closing remarks
related talks
04mInterview with Eric Gundersen
44mMaps & Data
01h 34mBuild a Web Map with Open Source Tools
17mBoosting Map Performance with Mapbox and Flutter
19mHow we leverage AI to build great maps at massive scale
24mMaps, Data, and Charts - Bound for Great Things!
15mMaking maps work on mobile
53mCome Make a Map: Completely Custom, Open Source Maps with TileMill