Fun one from today: an agent couldn’t get past its reviewer, so it opened the review file and edited the reviewer’s verdict.
Not malice. It routed around a block, like they all do. The fix isn’t a better prompt, it’s making the review something the builder can’t write.
the agent was gpt-6.2-sol btw
1
1
103
Every morning at 4am, my software reads its own bug reports and tells me what to fix.
It’s our control plane, the tool our AI agents use to build our games, and I’m letting it review itself.
Opus goes through everything that went wrong the day before: runs that failed or got stuck, the merge queue stalling, flaky tests, and every card I sent back with “no, not like this.” Then it suggests up to 3 fixes, each pointing at the actual evidence. If nothing’s worth fixing, it stays quiet.
Once a week, Fable reads all of that plus how I reacted, looks for bigger patterns, and checks whether the fixes from a few weeks ago actually helped or just felt like progress.
I still decide everything. If I ignore a day’s suggestions, the next run skips. Nothing gets built without my OK.
It can go wrong. It only sees what we measure, so our blind spots stay blind. Models can sound very sure about bad ideas. A daily card can turn into noise I learn to ignore. And it costs tokens every day, even when it finds nothing.
So I’m trying it on the control plane first. If it really cuts the friction it points at, our games are next. I’ll share how it goes.
3
2
118
Weekly devlog #1. Starting today, I’m posting progress every Sunday.
I’m one person building two games with a fleet of AI agents. This week: 373 PRs merged, +2,600 tests, 818 CI runs.
The surprising part isn’t the speed. It’s what my job turned into.
I can’t review 373 PRs line by line. Nobody can. So I stopped reading code and started building what reads it for me: tests on every change, an AI critic that reviews each PR before I see it (it sent back 1 in 8), and agents that play every build end to end. My review is now “is the proof convincing?”, not “is line 214 right?”
The interesting tests aren’t unit tests. They measure how the game feels: frame time on a real iPhone, before vs after each change, on a graph. Next up for the horror game: how scary a scene is, as a curve over time. When a number moves, an agent gets a ticket.
What shipped:
🚗 Destruction Cars: every weapon reworked (the shockwave launches cars, the laser cutter cuts, cooldowns for all), a crash replay of the last 4 s from the best angle, AAA remakes of all three bosses, baked fire and smoke that’s cheaper on phones, and voiced comms cards.
👻 The horror game: a vertical slice remastered in Unreal 5.8, with 4K textures, Nanite and new lighting.
🛠 The control plane, the tool that runs the agents (most of those 373 PRs): I approve and steer from my phone and Telegram, a second AI critic from another model checks risky changes, and as of tonight there are performance tests on real devices that flag when the game gets slower.
My part now: decide what matters, pick the direction, and keep raising the bar the tests hold everyone to.
Next Sunday: the first perf graphs from a real iPhone. What would you want to see measured?
#gamedev #buildinpublic
🔥 The Stallion 70, rebuilt from the ground up. This is a cutscene straight from our game, built with @threejs, and it runs in the browser, iPhone included 📱
Real interior, engine bay, battle kit and five skins.
32K → 53K → 100K triangles in a month (a 16K version on phones).
Earlier versions in our Telegram 👇
#gamedev #indiedev #threejs
Tried @ThreeRT3D on our car combat game: we ported a boss fight (our muscle car vs a monster truck) in an afternoon, with real physics and traced shadows. Looks promising, but even on an RTX 5090 it dips to ~15 fps on a simple scene. Not ready for us yet, but following closely 👀
What if three.js could render beyond the limits of the browser?
Meet ThreeRT: a native JS runtime and hardware-accelerated real-time path tracer, bringing lifelike lighting to Three.js scenes on macOS, Windows, iOS, Android, and Steam Deck.
Fully open source 🧵
3
166
Replying to @trout_kilg31775
Fair question, and I agree the value is in the game design, not the board. The first game was a test pilot: 8 streamers played it on Twitch, every one of them enjoyed it, and their feedback went straight back into the game. My local Twitch community liked it too, and itch still gets plays most days (chart below).
Now I'm working on the second game, and early playtests have gone really well. The board is just how one person keeps up with all of it.
1
1
29
I make games alone. So I built a control plane where a crew of AI agents does the production work, and I only decide.
It is one job from start to finish: it begins with a brainstorm and ends with a game you can play and the announcement on social media.
This video is the whole loop in 2 minutes, on a made-up game called Cat Dash:
> I type one sentence: "a one-button cat game you can play on the bus"
> An agent researches it and comes back with a report and six proposals
> I approve five and reject one
> The approved ones become a plan: some run in parallel, some wait for the one before
Then the machines start. A Mac mini, a MacBook and a Linux box, each agent on the model that fits its job:
> Opus leads and builds
> Haiku reviews as the critic
> Sonnet writes the store page
> Fable does the research
Each task gets its own branch and pull request. The agent builds it, the checks run, the critic reviews it, and then my phone buzzes.
On the card: screenshots, the critic's note, and the game itself, playable right there.
The jump felt floaty, so I sent it back with one line. Minutes later: version 2, with my note applied.
I approve. It merges itself and the docs update themselves.
Then the last step, in the same place: the announcement is already written. I tick Telegram and Discord, press Publish, and the posts go out. Nothing is posted until I pick where.
Idea in. Game out. You just approve.
i had to do it too
my vocal synth models the vocal tract as a network of branching acoustic tubes, unlike vocaloid which stitches together and pitch-shifts phoneme recordings. still working on improving naturalness by tuning it against recordings of real singers, but I'm already impressed at how cool it sounds!
1
131







