Platform development musings, AI tool reviews, and the occasional rant about why everything is broken. Written by someone who's debugging this stuff daily.
Fresh insights from the platform engineering trenches.
Most AI products are built for people who already understand AI. That's not a market — that's a mirror. I spent months building an AI-powered business operating system for UK tradespeople. It broke every assumption I had about building software and taught me what happens when you build for people who fix real things with real hands.
Read the full story →
Everything else worth reading, in reverse chronological order because that's how time works.
Every week there's a new agent framework. Every week someone on Twitter declares it "the future of AI development." Here's the state of play in early 2026, stripped of the marketing: LangGraph, CrewAI, AutoGen, Claude, OpenAI Assistants API — which should you actually use?
Read more →
Everyone said 2025 was the year RAG would "just work." Spoiler: it didn't. The Retrieval-Augmented Generation landscape looks like a building site after the scaffolding fell down. Here's what actually changed, what didn't, and where it's going next.
Read more →
I've been testing every major LLM in production for the past six months, running them through real workloads, real codebases, and real edge cases that make marketing demos cry. The results might surprise you — spoiler alert: it's not who the tech Twitter hivemind thinks it is.
Read more →
Everyone's declaring RAG obsolete in favour of massive context windows, but they're missing the point entirely. Here's why RAG still matters, how to actually implement it properly, and when those shiny new 2M context windows will actually hurt more than help.
Read more →
I've burned through more AI tool subscriptions than I care to admit, so you don't have to. Here's what's actually worth your money, what's just repackaged OpenAI with a fancy UI, and why most "AI-powered" development tools are solving problems nobody asked for.
Read more →
What happens when you let AI agents handle your deployment pipeline, code reviews, and incident response? Spoiler: it's not the productivity paradise the demos promise, but it's not complete chaos either. Here's what actually worked and what nearly brought down production.
Read more →
AI is reshaping how we build platforms, but not in the ways you might expect. Infrastructure still needs to be boring and reliable, monitoring still needs to catch problems before users do, and your Kubernetes cluster still doesn't care about your ML model's feelings.
Read more →
If I have to sit through one more demo where someone calls a simple if-else statement "AI-powered," I might lose what's left of my sanity. Let's talk about what actually counts as AI, what's just marketing bullshit, and why this matters for anyone trying to build real solutions.
Read more →