Training Data Physics Gets a Vote: Nominal Cofounders on Hardware Development in an AI World
90 snips
Mar 10, 2026 
Jason Hoch, cofounder of Nominal, builds test and validation tooling for complex hardware. Cameron McCord, cofounder of Nominal, designs data and AI platforms for hardware engineering. They discuss bringing software practices to hardware testing. They cover the simulation versus real-world testing gap, using AI to automate repetitive test work, and how telemetry and validation infrastructure enable physical AI.
AI Snips
Chapters
Transcript
Episode notes
Hardware Needs Its GitHub Moment
- Hardware is entering a software-like era where centralized platforms (like GitHub did for code) will replace fragmented local workflows.
- Cameron McCord argues testing and DevOps for hardware are unsolved problems today and ripe for productized infrastructure to accelerate development.
SpaceX Built Its Own Test Platform
- SpaceX built proprietary internal tooling early because existing test analysis software wasn't good enough and empowered their engineers to move faster.
- Jason Hoch cites SpaceX's in-house platform as a major reason for their acceleration and Nominal's motivation to offer that externally.
Testing Still Runs On PDFs
- Most hardware test data still lives locally and is shared via PDFs and PowerPoints, preventing centralized analysis and traceability.
- Cameron McCord describes engineers downloading data to laptops and sharing screenshots rather than structured, cloud-backed datasets.