GPT 5.6 Builds a Three-Body Problem Simulator
Acceptable-Object390 · reddit · 2026-07-15
The author used GPT 5.6 Sol to have Row-Bot build a "Three-Body Problem" interactive website. The goal wasn't just a demo of three glowing dots, but a technically credible numerical gravity workbench.
Implementation
- Tech stack: React, TypeScript, Vite
- Simulates the planar Newtonian three-body problem
- Employs 4 independent numerical integrators:
- Explicit Euler
- Symplectic Euler
- Velocity Verlet
- Fourth-order Runge–Kutta
- Supports switching integration methods and time steps
Initial Condition Presets
- Fermat figure-eight orbit
- Lagrangian equilateral triangle solution
- Hierarchical three-body system
- Chaotic scattering
Credibility Design
- All presets are first transformed into a center-of-mass frame with zero total momentum
- Continuously calculates physical invariants: total energy, linear momentum, angular momentum, and center-of-mass position
- Tags states as nominal / caution / unreliable based on drift levels
Chaos Experiment
- The system creates a "shadow system" differing by just one initial coordinate (10^-7)
- Measures their separation in phase space over time to observe chaos-induced bifurcation
More from coding & agent
- A Forward Deployed Engineer job really has three stages: audit, evals, deploy — blaizedsouza · 2026-07-22
- 438 sealed tests show coding agents prefer DIY over third-party databases — cramforce · 2026-07-22
- Tenable and AWS launch a Black Hat build event for open-source security agents and MCP servers — Dave_Maynor · 2026-07-22
- Codex helps build Valdiluce, an open-world game with climbing, gliding and gondolas — Dimillian · 2026-07-22
- HeyGen adds a media-sourcing skill for coding agents with 75k images and 10k tracks — HeyGen · 2026-07-22
- Agent search bottlenecks are now about variance, not raw latency — rohanpaul_ai · 2026-07-22