Experiments ·
He Asked Three AI Models to Simulate a Galaxy Collision
Someone ran the exact same simple prompt, simulate the milky way andromeda collision, through three different AI models and compared the results side by side. Each one spent about the same amount in tokens, yet the outputs looked completely different.
One simulation rendered a dynamic, physically accurate looking particle merger, while another came out far more static, showing how much models can diverge on identical instructions.
Made by u/Fleischkluetensuppe. This write-up follows the original post; it is not an independent product test.
Keep up with what gets built.
Read the daily AI briefing, browse the tools, or sign up for briefing email updates.