The article discusses the author's experience with infrastructure decisions at a startup over the past four years, including choices such as using AWS over Google Cloud, adopting EKS, and utilizing various AWS services. The author endorses some decisions, such as using AWS and EKS, while regretting others, like using Datadog due to its high cost. The author also shares lessons learned from implementing various processes and tools, including GitOps and incident management. The goal is to provide insights for other startups making similar infrastructure decisions.