The compute continuum
Compute is a continuum. Not every service needs to be a server.
Compute does not begin and end in a regional cloud AZ. End-user laptops, phones, SBCs, factory devices, and purpose-built hardware are part of the same continuum.
Designing every workload as a server is a habit from a thinner world. Some work belongs at the edge for latency, cost, privacy, or resilience. Some belongs centrally for aggregation and control. The skill is placing work on the continuum deliberately.
Value augmentation on latent compute—devices already paid for and already powered—will keep expanding. Architectures that assume “always a cluster” will overpay and underperform.
