Notes on AI agents and MCP
Short, plain notes about building AI agents and MCP servers that work in real businesses: what works, what breaks, and why.
June 15, 2026
How I put 10,000 players in one world
Most online games hide their scale — they split players into rooms of 20 or shards of a few hundred. For Helix Empire I set a harder target on purpose: 10,000 players in a single shared world, on one server, live in the browser. This is the full story of how that gets built — the four walls you hit, why the real bottleneck is traffic and not CPU, and the moment a load test proved my pretty number was a lie. It's long, it's technical, and every claim ends in a measurement. The lessons transfer to any high-load system.
- architecture
- methodology
June 15, 2026
The boring AI win is paperwork
The NHS just signed a £120 million deal to give 505,000 staff an AI assistant. Not to diagnose disease — to do paperwork. In trials, the average person saved 43 minutes a day, and one ward cut its backlog of discharge letters by 62% in a month. That's the AI story nobody puts in a keynote: the durable, deployable value is usually the dull, high-volume admin work, not the dazzling demo. Here's why the boring use case is the one that actually pays, and why you should hunt for yours.
- business
- methodology
June 14, 2026
The AI coding speedup is smaller than it feels
In a controlled study, experienced developers using AI on their own codebases were measurably slower on complex tasks — and felt 20% faster the whole time. A 2026 follow-up with better methodology landed near a small positive, not a big one. Meanwhile ~93% of developers use AI tools and headline productivity has barely moved. None of this says AI coding is fake. It says the feeling of speed and the fact of speed have come apart, and if you manage by feel you'll manage wrong. Here's how to tell the difference.
- methodology
- careers
June 13, 2026
A green checkmark can hide a broken middle
Here's the failure mode that eats AI agents in production: an agent runs a multi-step task, makes a wrong turn somewhere in the middle, and still produces a final answer that passes your check. The output looks clean. The reasoning was broken. Researchers found this is exactly how multi-step agents fail — a step-three mistake propagates invisibly into a step-ten summary that reads fine and is wrong. If you only grade the final answer, you're blind to most of how agents actually break. Here's why, and what to check instead.
- methodology
- agents
June 13, 2026
The biggest context window doesn't win
Every model launch brags about a bigger context window — a million tokens, two million, the whole codebase at once. But an analysis of enterprise deployments found that nearly 65% of agent failures came from context drift or memory loss during multi-step work, not from a window that was too small. The teams shipping reliable agents in 2026 aren't the ones with the biggest window. They're the ones who curate hardest what the model actually sees. Here's the difference, and why more is often worse.
- agents
- methodology
June 13, 2026
Your agent works 57% of the time
A March 2026 report looked at 6,259 AI agents running in real production and found an aggregate success rate of 56.6% — barely better than a coin flip. The same studies show a 37% gap between how agents score on benchmarks and how they do in the real world. That gap is the whole story. The demo always works; the job is making the agent work the other 43% of the time. Here's why the number is so low, and what the teams above it actually do differently.
- agents
- methodology