The Cloudcast Time Series for Physical AI
24 snips
Nov 26, 2025 Evan Kaplan, CEO of InfluxData and a leader in open-source time series technologies, dives deep into the world of Physical AI. He contrasts this with generative AI, emphasizing the need for deterministic systems and high-resolution data. Kaplan explains the efficiency of time series databases and how they outperform traditional databases. He also discusses the trade-offs between data resolution and cost, and why real-time processing is crucial for applications in edge computing. Expect fascinating analogies with satellites and autonomous vehicles!
AI Snips
Chapters
Transcript
Episode notes
Evan’s Journey To Time-Series Focus
- Evan Kaplan described his path to InfluxData and belief in time series before AI hype.
- He joined after meeting founder Paul Dix and saw time series breaking into mainstream use.
Time-Series Is Tailored Telemetry
- Time series is telemetry indexed by time for efficient reads and high-volume ingest.
- It enables sub-10ms queries, downsampling, backfill, and efficient long-term storage.
Physical AI Needs Vast, High‑Resolution Data
- Physical AI collects effectively infinite real-world data versus finite digital corpora.
- More resolution yields more deterministic behavior necessary for safety-critical systems.