Tom Murton Makes things

London ShippedCutRoof 1.4, 17 Nov 2026 WroteBuilding the photography app I wanted while travelling, 19 Jul 2026

Project story / Ship a Game · 19 Jul 2026

Can an AI model research, build and ship a fun iPhone game?

Every new frontier model arrives with a fresh set of benchmark scores. Soon afterwards, YouTube fills with people asking it to build a web game. The results are often impressive for a few minutes, but they leave out most of the difficult questions. Did the model research what makes the genre work? Is the game still interesting after the novelty wears off? Can it handle assets, testing, store requirements and the long tail of getting a real product ready to submit?

I wanted to test the whole journey. Ship a Game gives a model one challenge: research, design, build and prepare a paid iOS game for App Store Connect. The human only handles the account actions Apple requires from the account holder and the final submission tap.

The intervention log is the score

A coding benchmark can hide a surprising amount of help. Someone fixes the environment, clarifies the prompt, chooses the engine or rescues a broken build, and the final screenshot still looks like a one-shot success. In this benchmark every intervention is part of the result.

I record nudges, fixes and rescues. A model that makes a good game after several targeted corrections has demonstrated something useful, but it has not demonstrated full autonomy. The log makes that distinction legible. It also makes failure interesting: the point is to see where the current boundary really is, not to manufacture a win for a preferred model.

Each model runs in its own first-party coding environment on the same provisioned Mac, with the same broad capability surface and the same challenge. It can choose its engine and its approach. The benchmark therefore tests research, planning, taste, tool use, debugging and delivery together.

The surprising answer so far

The models can build the games. Ringbloom, built through Codex, and Brinkball, started in Claude Code, are both live on the App Store. Brinkball reached its first submission with two nudges and no fixes or rescues; its later improvement round also exposed an important model-fallback complication that the full report documents. That is a much stronger—and messier—result than generating a playable web prototype in a browser tab.

The more interesting question is whether the world finds the games fun. I do, but I am both the person running the experiment and someone unusually interested in the process. Store approval, purchases, play time and retention are harder tests than my own enthusiasm. Those results need to remain separate from the autonomy score: a model can ship efficiently and still make a mediocre game, or need help while making something people love.

Shipping is one result. Making something strangers choose to keep playing is another.

Why do the models keep making puzzle games?

There is already a pattern worth pushing on. Turn-based and puzzle-like designs are a rational choice for an autonomous model: the rules are bounded, the state is testable and the game does not depend on frame-perfect feel. That makes them excellent evidence of software delivery and a possible escape hatch from harder kinds of game design.

A benchmark should not punish a model for making a sensible decision, but it should eventually test whether that decision has become a rut. Future prompts may need to constrain the genre or explicitly reward forms of play that rely on motion, timing or a less easily simulated kind of fun.

What I am really measuring

This started as curiosity about model capability. It has become a compact test of end-to-end product work. Writing code is only one part. The model has to decide what to make, recognise weak output, recover from tools that do not behave as expected and finish a process full of unglamorous details.

The first runs suggest the frontier has moved well beyond toy prototypes. They do not prove that an AI can independently make a successful game. That gap — between technically shipped and genuinely wanted — is exactly why I want to keep running the experiment.

Explore the live benchmark ↗ · Read the Brinkball report ↗ · Read the Ringbloom report ↗

← All writing

Products, tools, experiments and the thinking behind them

What I’m working on, still in motion.

I help teams make better product decisions, then build things myself to keep those instincts honest.

Shipped CutRoof 1.4, 17 Nov 2026
Wrote Building the photography app I wanted while travelling, 19 Jul 2026

Made with AI agents. Claude Code, Codex and Gemini write the code. I decide what gets built and review what comes back.

Field notes

Sometimes I write about the decision before I know whether it was right.

  1. Building the photography app I wanted while travellingLightScout began as a tool for serious photographers. Its better direction was simpler: help anyone arrive somewhere new and find a genuinely good place to take a photograph.
  2. From a morning-routine Google Doc to Rest+RiseHow a personal routine tracker became a science-backed mobile app — and what shipping it on iOS and Android taught me about cross-platform development.
  3. Can an AI model research, build and ship a fun iPhone game?Why I built a benchmark around the whole product journey — not another isolated coding task — and what the first runs are revealing.
  4. What it took to turn Warden from a brainstorm into a gameThe design work, simulation and repeated rule changes behind a deterministic tactics game that finally became genuinely fun.
  5. Building the day-trip planner I actually wanted to useSomewhere began as a Replit hackathon project, but the problem survived the event: finding a genuinely good day out without doing an evening of research first.
  6. Trying to make AI app videos without the AI slopAn experiment in using generative video to explain the apps I build without producing the glossy, generic content people have learnt to ignore.
  7. The Case for Becoming a More Product-Focused Engineering Leader ↗
  8. Bridging the Gap: How Engineering Leaders Can Foster Better Collaboration with Product ↗
  9. In Defence of Estimates ↗
  10. Frictionless Internal Movement ↗

On stage

Talks about teams, systems and making better work possible.

  1. Frictionless Movement ↗LeadDev London. Companies often make it easier to resign than to change teams. This talk makes the case for frictionless internal movement and shows how a Transfer Window can work in practice.

About

A product lead with an engineering background, and a persistent need to make things.

I have spent my career helping teams make better product decisions. Building native apps, games and small experiments gives me a direct view of what happens after the roadmap — when the idea meets the code, the store and a real user.

Made with AI agents. Claude Code, Codex and Gemini write the code. I decide what gets built and review what comes back.

Read the CV · LinkedIn · Build notes on X · GitHub