Description
What if AI could explore hundreds of product paths in parallel—and never lose track of what changed? In this TiDB SCaiLE 2025 keynote, PingCAP Co-Founder and CEO Max Liu rethinks the future of building with AI by shifting from code-only versioning to state versioning: memory, data, config, environment, and history.
Max contrasts today’s single-path, human-steered workflows with a future where AI explores 100× paths simultaneously with full visibility. This keynote sketches a multi-tenant multiverse of users, each spinning up countless branches on an agentic AI platform, backed by data infrastructure that can keep up.
An ideal blueprint for a solution includes data branching, data versioning, flexible schema management, compute–storage separation, and distributed transactions—all aimed at unlimited scalability. Max reveals how TiDB is the only distributed database that meets these requirements for parallel exploration.
You’ll learn:
- Why state versioning unlocks parallel exploration
- The essential primitives for universal branching
- How to design data systems for an agentic, multi-tenant future
- Where TiDB fits when exploration meets scale
Speaker
Max Liu
Co-Founder & CEO, PingCAP
Related Resources
Agentic AI Architecture for Scale: A Unified Database on TiDB
Agents at Scale: The Manus Story and What Comes Next for AI
Customer Story
4 GenAI Use Cases That Power Intelligent, Scalable, and Context-Aware Systems
Agentic AI Architecture for Scale: A Unified Database on TiDB
Agents at Scale: The Manus Story and What Comes Next for AI