Category: IT Trends & Reviews

One review per quarter, going back to the start of 2017: what shipped, what turned out to matter, and what was noise. Each entry covers the same ground — frameworks and language releases, cloud and Kubernetes, AI research and products, hardware, security incidents, and regulation — so the quarters can be read against each other. Read in sequence they show how quickly the centre of gravity moved. In 2017 the argument was TensorFlow against PyTorch and whether Kubernetes had won. By 2020 it was GPT-3 and Apple Silicon. By 2023 open-weight models had arrived and the EU AI Act was being drafted; by 2026 the same act is being enforced and inference infrastructure is the story. These are written after the fact, with the benefit of knowing which announcements went anywhere — which is the point. A quarterly review written in the moment records hype; written later it records outcomes. Each review links to the previous and next quarter, so the whole series can be read as a timeline. Where a quarter introduced something covered in depth elsewhere on the blog — a model runtime, a vector database, a Python library — the review links across to it.