Creative Assembly Loops
Loop v1.2 · 1 shipped · 1 in progress ·
What if we rebuilt Denki's creative assembly lines as agentic loops?
A long-term project to remake how playthings get made, one loop at a time.
Play 001 Pass Master →A one-tap football timing game, on your phone or desktop.
01What
A human-AI co-production loop for tiny, instantly playable browser games. Each loop ships one thin slice, a polished first playable, and leaves the process better than it found it.
02Playthings
Every plaything is also an experiment on the loop, and in service of its iteration.
-
001Pass Master · A one-tap football timing game
Theory into practice: build it, expose the assumptions, codify the loop.
-
002Rondo · Pass Master, evolved
A more nuanced design and iterate loop within preproduction, timed from seed to sign-off.
-
003Amp · A Denki prototype, re-designed and rebuilt from fresh
A Pepsi challenge: the original Denki prototype stays sealed until the rebuild is done, then we compare and critique.
03Why
Everything is a games platform now and making is cheap, so the constraint has moved: choosing what to make, proving it early and polishing it properly. Denki solved that with process, not headcount; agents let a tiny team run that process again, in loops that get better with every pass.
The idea isn't what goes in at the top. It's what flows through the loops.
04Context: lines then, loops now
Nathan Baschez argues that the future of work is shared factories, not personal agents. It made me think of Denki, where we didn't run a traditional games studio: we ran creative assembly lines. In 2006 I joined Denki as its third Entertainment Product Architect, the equivalent of a showrunner. I was responsible for the design, production and delivery of all the company's game titles and digital toys, across digital interactive TV, online, mobile and other formats. Games went from idea to live on Sky in six to ten weeks (TV as a Games Platform), and I ran the same process in short loops to take Quarrel from vertical slice to BAFTA Scotland Best Game.
If you rebuilt Denki today, it would be a tiny team at the heart of a colony of agents. Tom Critchlow's Of Termites & Tokens names the shift: “the opportunity is not to build faster, smaller machines - but to become a colony: a mixed population of humans, agents, bots, scripts, dashboards, workflows, alerts, memories and permissions, all sensing and acting through local signals.”
So this is agentic augmentation, not automation: the Entertainment Product Architect remade as scout, shaper, showrunner, editor and storyteller. (More on lines and loops.)
05How it works
Loop v1.2, correct as of 2 November 2026 (the process updates with each release)
| Stage | Proves |
|---|---|
| 0 · Seed | The intent, and everything risky or unknown about it |
| 1 · Interactive Spec | Risky assumptions proven or killed in toys, before anything is built |
| 2 · Play 1.0 | Core play feels good on the target device, with placeholders |
| 3 · Game 1.0 | The rules and structure work |
| 4 · Thin Slice | It looks, sounds and feels finished |
| 5 · Ship | The public build works live |
| Postscript | What we learned, written up, and fed back into the loop |
- Me
- Vision, direction, shaping, balancing, playtesting, editing and sign-offs.
- Claude
- Questions, specs, toys, builds, tuning, testing and deployment.
Front-load risk, back-load polish.
Nothing moves without a sign-off.
Every post-mortem iterates and evolves the loop.
06Roadmap
The long-term ambition is to build the Denki × Culturematic × Kniberg equivalent of the OODA loop. For now, I'm focused on Version 1, Creative Assembly Loops: one loop, with a layer added for every plaything it ships, and plenty to do inside it. V2 and V3 are where one loop becomes a network of loops.
| Version | Adds | The question |
|---|---|---|
| V1 · Creative Assembly Loops (now) | Make it | How fast can a tiny human-AI team make playthings to a Denki standard? |
| V2 · + Attract Mode | Validate it | How might we put each plaything in front of a real audience and find out whether it pulls them in? |
| V3 · + Attention Arbitrage | Scout it | Which cultural opportunity, audience, mechanic and distribution surface should we combine in the first place? |
Attention Arbitrage identifies what might deserve to exist.
Creative Assembly Loops makes and iterates it.
Attract Mode discovers whether the audience agrees.
07Lessons learned
001Pass Master
In one line: do the hard human work up front, because every shortcut came back as rework.
What went well
- We shared one vision: the high concept held from first build to ship, and the style guidelines gave us a shared language for fast feedback.
- The Denki process worked in theory, then in practice: the more faithfully we followed it, the faster and better it went.
- We iterated in short loops: with real references and a tuning panel of exposed, editable values, we could test an assumption or fix a design snag in a single round.
What could have gone better
- We deliberately cut corners at the Seed and on the iteration path, and paid for it in rework.
- Assumptions and contradictions became direction without being flagged or tested.
- Our tools and records came late: the tuning panel, the making-of log, the repo and naming conventions.
Lessons for next time
- Do the hard human work on the Seed, and flag every risk while shaping, before the Spec is signed off.
- Prove risky assumptions and feel as toys in the Interactive Spec, before building Play 1.0.
- Follow every stage of the loop. Each one is there for a reason.