SIML Worm Agent Release

This release includes cross-platform binaries and a Godot visualizer for the SIML artificial worm simulation from the following paper:

Artificial Intelligence is for Amateurs: SIML, Free Energy Principle and Cognogenics as the Foundations Toward the Birth of True Artificial Life

---

🐛 Contents

You can download all the files in the latest release.

• siml-linux-x64 – Linux binary
• siml-windows-x64.exe – Windows binary
• siml-apple-silicon – macOS (M1+) binary
• siml_visualizer.pck – Godot 4.4.1 visualizer (platform-independent)
• README.md – You're reading it

---

🚀 Quick Start (Visualization Required)

--

🧠 Visualizer (Godot-based)

macOS Only

• Download and double click on apple-silicon.zip
• Right-click .app → "Open" → bypass Apple Gatekeeper if needed
• Extract both files
• In the command line, we need to set up sandbox to run the files with this command since we didn't pay $99 for an Apple Developer Account:

xattr -dr com.apple.quarantine siml-visualizer-macos.dmg
xattr -dr com.apple.quarantine siml-apple-silicon

• Double click the siml-visualizer-macos.dmg file, and click on the godot-command file that opens
• Requires macOS 12+ with Apple Silicon

Linux and Windows

Install Godot 4.4.1

Make sure it's the same version used to export the .pck (e.g., Godot 4.4.1+).

OS	File	Instructions
Linux	siml_visualizer.pck	./path-to-Godot-executable --main-pack siml_visualizer.pck
Windows	siml_visualizer.pck	./path-to-Godot-executable.exe --main-pack siml_visualizer.pck

---

⚙️ SIML Binaries

Choose your platform and run the binary from the terminal:

OS	File	Instructions
macOS	siml-apple-silicon	chmod +x siml-apple-silicon && ./siml-apple-silicon
Linux	siml-linux-x64	chmod +x siml-linux-x64 && ./siml-linux-x64
Windows	siml-windows-x64.exe	Double-click to run or use PowerShell

---

Available Flags

Flag	Description
--record-mode	Output mode for recording simulation (bit or rgb). Default: bit.
--runs	Number of worms to simulate sequentially. Default: 1.
--use-memory	Load memory from previous agent run and persist schema learning.

Example: Run 10 worms with existing memory

Be sure to place the agent_memory.bak and agent_memory.bin file in the SAME folder as the siml executable if you want to use the existing trained agent.

./siml-[platform] --runs 10 --use-memory

The simulator will save a .bin file to record your runs, and start over with the saved .bin every time. If you want to start from scratch again, simply rename or remove the .bin.

Logs

You’ll see logs like:

[SIML] Starting Capsule…
[SIML] Run 0 spawned at (x, y, z)
...

📜 License

This is a binary-only research release for reviewers and collaborators. Can not be used currently for any commercial purposes. There are headless formats and other flags that can be added for research.

All source code is withheld under internal license.

🔗 More Info

For citations, papers, or future inquiries, contact:

Author: Richard Everts

Deep SIML Labs