Charity Majors (Honeycomb) draws a clear line: > "DevOps is about automation and managing infrastructure. Platform is about not having infra to run." DevOps assumes infrastructure exists and optimizes its management. Platform engineering asks whether the team needs to run that infrastructure at all — outsourcing to managed services or cloud providers until scale forces insourcing. Platform teams work higher up the stack than operations, DevOps, or SRE teams. Their tectonic-plate metaphor: sitting between infra code (slow-moving) and business code (fast-moving), enabling abstraction at the boundary. ## Cross-Domain Applications **Startup scaling decisions**: When to bring infra in-house vs pay variable SaaS costs — platform thinking frames this as "total cost of ownership" vs "developer cycle cost." **Organizational design**: Choosing between an SRE model (you own infra excellence) vs platform model (you eliminate infra exposure) shapes team topology, hiring profiles, and vendor strategy differently. **Product management**: DevOps maturity metrics (DORA) measure delivery speed within existing infra. Platform maturity metrics measure cognitive load reduction and self-service adoption — different success criteria. ## References - [[Platform Engineering, DevOps, and Cognitive Load]] (InfoQ, Oct 2022) - Charity Majors, CTO Honeycomb ## Related Concepts - [[Premature Abstraction Trap]] — Manual validation needed before committing to infra abstractions