💫 I forced 4 AIs to recreate Mario from scratch using Symfony AI and Godot

on May 8, 2026

Creating a video game with artificial intelligence has become almost commonplace.

Today, all you need to do is open Cursor, Claude Code or ChatGPT and type:

“Create me a Mario.”

And a few minutes later, something appears on the screen.

But behind this apparent simplicity lies a much more interesting problem:

How to properly orchestrate multiple AI agents to produce a coherent, playable, and maintainable result?

That's precisely what I wanted to experiment with in this video:

ChatGPT
Claude
Gemini
Grok

Even prompt. Same objective. Completely different results.

But above all: a completely different architecture from what is called “vibe coding” today.

An inspiration from Playdate

Before even talking about AI, we need to talk about video games.

In the video, I quickly show the Playdate, a small independent console developed by Panic with a black and white screen and a mechanical crank.

What's interesting about the Playdate isn't the power.

That's the philosophy.

The equipment imposes constraints:

limited screen
minimalist controls
simple gameplay
maximum creativity

And paradoxically, these constraints make the games more interesting.

It's exactly the same phenomenon with AI agents.

A vague prompt often results in a fragile system.

Well-defined constraints produce much more consistent results.

The Single Prompt Problem

Most AI demonstrations work like this:

Create a Mario game

The model then attempts to:

generate the gameplay
handle collisions
create the graphics
balance the levels
produce the mechanics
organize the project

The result is often impressive for a few seconds.

Then appear:

broken collisions
impossible levels
inconsistent rules
contradictory mechanisms
an immediate technical debt

The problem is not necessarily AI.

The problem is that he is being asked to do everything simultaneously without specialization.

Symfony AI + Godot

For this experiment, I used two main technologies:

Symfony AI

Symfony AI provides:

the agents
the tools
multimodal
the RAG
memory
multi-agent orchestration

The framework is divided into several specialized bundles:

Agent
Store
AI Bundle
Platform
App

The goal is to build a clean architecture around AI models rather than simply sending prompts.

Godot

Godot Engine

Godot is an open-source game engine extremely well-suited for rapid prototyping:

2D
3D
scene system
GDScript scripting
nodal architecture
multi-platform export

The idea was simple:

Orchestrate AI agents with Symfony AI to automatically generate a Godot game.

Multi-agent orchestration

The core of the project is based on a multi-agent architecture.

Instead of having a single AI that does everything, I separated the responsibilities.

The Orchestrator

The orchestrator receives the main prompt:

Generate a Mario-like platformer

Then it automatically delegates to the appropriate specialized agents.

Specialized agents

Layout Agent

The “level designer”.

It generates:

the platforms
the holes
the pipes
the stairs
the progression of the level

Gameplay Agent

The mechanics specialist.

He manages:

collisions
the enemies
the bonuses
the jumps
the checkpoints
the interactions

Godot Export Agent

The role of this agent is to transform the generated data into structures compatible with Godot:

scenes
TileMaps
objects
JSON exports

QA Agent

The most important thing.

The QA Agent verifies:

the gameplay
collisions
impossible jumps
the consistency of the mechanics
the rules of the level

This is probably the element most lacking in “vibe coding” demos.

The orchestration with Darkwood Flow and Navi

To orchestrate all of this, I use:

Flow allows you to manage the execution of multi-agent workflows.

Navi allows:

traceability
observability
routing
monitoring of executions

The pipeline looks like this:

User Prompt
    ↓
Orchestrator
    ↓
Layout Agent
Gameplay Agent
Godot Export Agent
QA Agent
    ↓
Generated Game

The orchestrator automatically chooses which agent should intervene depending on the progress of the system.

A single Symfony command

The entire generation process is launched with a simple Symfony command:

php bin/console app:mario:orchestrate

This order:

calls the orchestrator
Route to the specialist agents
generates the result
Automatically exports the game

The most interesting thing: The pipeline becomes reproducible.

The results

ChatGPT

ChatGPT produces a very good database quickly.

We find:

a faithful structure
consistent gameplay
a rapid iteration

The result is strongly reminiscent of the old Mario Game Boy games.

The main advantage: the speed of prototyping.

Claude

Claude produces something more graphically detailed.

The activities:

are more fluid
more modern
sometimes more ambitious

The reasoning seems more structured.

But we also feel more creative freedom compared to remaining faithful to the original.

Gemini

Gemini is probably the most visually interesting.

We find:

more modern graphics
better art direction
greater inventiveness

The gameplay also seems more refined in certain aspects.

There's a more "Nintendo-like" approach.

Grok

Grok is the most unpredictable.

He takes more initiative:

level selection
scoring system
graphic variations
new gameplay approaches

But it also deviates further from the original Mario.

The real subject is not Mario

This project was ultimately just a pretext.

The real issue lies elsewhere:

The future of AI development lies in orchestration.

A single agent can produce a prototype.

But several specialized agents can produce:

systems
pipelines
workflows
coherent architectures

And most importantly: reproducible results.

The future of vibe coding

Current “vibe coding” often relies on:

a huge prompt
a single AI
many hallucinations
little validation

The next step is probably:

specialized agents
explicit constraints
tests
QA loops
observability
of the orchestration

In other words:

The future of vibe coding is orchestration.

Source code

Symfony AI + Godot: matyo91/mario-orquestration: https://github.com/matyo91/mario-orquestration
ChatGPT: yoloyee1/SMB-clone-made-by-Godot: https://github.com/yoloyee1/SMB-clone-made-by-Godot
Claude: Axelrpg/Super-Mario-Bros-Godot: https://github.com/Axelrpg/Super-Mario-Bros-Godot
Gemini: charpurrr/SuperMarioSolarEngine: https://github.com/charpurrr/SuperMarioSolarEngine
Grok: wchen02/platformer: https://github.com/wchen02/platformer

Resources

Presentation slides: https://github.com/matyo91/slidewire
Symfony Multiagent: https://github.com/symfony/ai/blob/main/examples/multi-agent/orchestrator.php
Symfony AI
Godot Engine
Playdate

Links of the week

LIVE 1/5 LIBERTY WEBI OPEN HOUSE - THE AI SHORTCUT: https://www.youtube.com/live/L0H-rgqf7zI?is=1mc0YgZAxwjABKay
Eve of the week of May 4, 2026 Guillaume Loulier: https://guikingone.substack.com/p/veille-de-la-semaine-du-4-mai-2026
How MrBeast Works 18 Hours Per Day https://youtu.be/wPeUb2SVmEc?is=psItbJ2YNJFjVHiq
Richest People in the World (1900-2026) - From Rockefeller to Musk: https://m.youtube.com/watch?v=5h_vKrHzRvg&is=M9gch9O_AuqDTiKe
Day 287 growing my SaaS startup to $1M: https://m.youtube.com/watch?v=7aQ01ZKysl8&is=tdQGj8GddnmVH5sQ
ANDREJ KARPATHY JUST DECLARED THE END OF PROGRAMMING AS YOU KNOW IT. https://x.com/neil_xbt/status/2049687182336541013?s=46
How France's biggest pirate site was taken down: https://youtu.be/ULTV6T6jRTw?is=F9k6GcMNll0IfjZo Japan has just made a major decision: https://youtu.be/n_Kfco-kuFY?is=TBrMM-XdMTeE0qGR
check: Remove use of subtyping for monads: https://github.com/Gabriella439/grace/pull/261
Agentic Search for Context Engineering — Leonie Monigatti, Elastic: https://www.youtube.com/watch?v=ynJyIKwjonM
FLUX, open research and the future of visual AI — Stephen Batifol, Black Forest Labs: https://www.youtube.com/watch?v=x8Yb4RidLgM
How the Transformers finally devoured vision – Isaac Robinson, Roboflow: https://www.youtube.com/watch?v=VhfAVA3BG2I
This 50-minute lecture by Jeff Bezos will teach you more about business than a 2-year MBA program: https://x.com/Alokkumarzz/status/2049882855170588755?s=20
This lecture from Massachusetts Institute of Technology breaks down wealth, compounding, and long-term thinking in a way most financial advice never does: https://x.com/allen_explains/status/2049857599919825252?s=46
Stanford just taught the entire thing in a 1-hour lecture. For Free. https://x.com/codewithimanshu/status/2051240758125252872?s=46
this 2-hour Stanford lecture on AI careers. It will teach you more about winning in the AI race than all the AI content you've scrolled past this year. https://x.com/radha_ai/status/2051617864617263397?s=46
Peter Thiel gave a 1-hour class on how to create a monopoly starting from 0. https://x.com/indutripat82427/status/2051709170618098065?s=12
In this 24-minute session, Cherny outlines a future that is already his daily reality: https://x.com/datachaz/status/2051770066912379309?s=12
Radha Tripathi on X: "Claude FULL COURSE 1 HOUR (Build & Automate Anything) https://t.co/gg81I06To0" /
A Stanford University lecture taught the way engineers are actually trained to think about AI systems. You understand the why, every prompt you write, every system you design, and every agent you build operates from a completely different mental model. https://x.com/cyrilxbt/status/2049690838972723503?s=46 -Sam Altman. Dustin Moskovitz. The first lecture of the most influential startup course ever recorded. https://x.com/neil_xbt/status/2050411454830907898?s=46
OpenAI's own engineers just showed how to actually use OpenAI Codex properly. https://x.com/heyzarakhan/status/2051357610222239745?s=46
How much of your project code was written by #AI? Whether to track coding agent adoption in engineering teams or security risks introduced by vibe coding, it's time to track what Claude/Cursor/Copilot/Codex do. https://usegitai.com/: https://www.linkedin.com/posts/fzaninotto_git-ai-track-ai-code-all-the-way-to-production-share-7450124640389263360-jhxu?utm_source=share&utm_medium=member_ios&rcm=ACoAAA2M-8wBB-yYv6qd4pdeoz3fTBIAhu6HQH8
practitioners' perspective: Takeaways based on a qualitative survey of 419 practitioners involved in workflow maintenance: https://dl.acm.org/doi/10.1145/3806833
In 14 minutes, this Anthropic engineer who created MCPs will teach you more about building them right than most developers figure out on their own in months https://x.com/Av1dlive/status/2051967150894772609?s=20
Star Wars: Episode 1 - Racer (Podracing) 100% Speedrun in 1:17:33 [World Record]: https://youtu.be/OWv69snqBFs?is=kNL8iJzzGVRL3hRl
In 2007, Elon Musk Predicted Everything: https://youtu.be/xyCOvT1Y5YQ?is=igUlB5wqAuX1qbRs

🎶 Music credit

Questo trend è mio! ⭐️ : https://www.tiktok.com/@maya_cianflone/video/7607524440364289302

@maya_cianflone
Questo trend è mio! ⭐️
♬ original sound - smb.0685