FFC NL 2025 TALKS

Cognitive load has become a central theme in platform engineering, often framed purely as a developer experience problem. Golden paths and self-service platforms aim to simplify delivery, but cognitive load cannot be eliminated - only shifted. Much like the conservation of energy, this is what I call the conservation of cognitive load.

In this talk, I’ll explore how platform engineering is fuelled by the desire to manage cognitive load but focusing only on developers gives an incomplete picture. The load doesn’t disappear; it shifts. In practice, much of it lands on the platform engineers who design, run, and evolve the abstractions. This raises new questions: how do we reduce their cognitive load, how do we scale architecture so they can manage it, and how do we scale platforms themselves - often through federated or layered groups of platforms - to spread the weight? Cognitive load comes in different forms; intrinsic, extraneous, and germane, and each requires different strategies, not just for developers, but for the platform teams carrying the complexity.

I’ll share an example from Fidelity International’s migration from Cloud Foundry to Kubernetes to show how the conservation of cognitive load plays out in practice. From the “hotel model” of our Kubernetes platform to the challenges of hosting infrastructure workloads, I’ll explore where complexity settled, how abstractions failed to absolve teams, and why invisible complexity still shapes architecture. The aim is to give a practical lens for deciding where cognitive load belongs — and how platform teams can scale to carry it.