🧪 Artificial Intelligence Superintelligence

This simulator models the potential development and behavior of highly advanced artificial intelligence systems by treating capability growth as a recursive feedback loop: a system's intelligence at time t determines how effectively it can improve its own intelligence next. Adjust the improvement rate, the returns-to-intelligence exponent and the compute ceiling to move between diminishing-returns "soft" takeoffs, steady exponential growth, and abrupt "hard" takeoffs that consume the available ceiling in a burst — the same decision-making dynamics discussed in superintelligence-scenario research, made explorable rather than abstract. Each unit of capability renders as a node in a growing 3D network, wired to the earlier capability it was recursively built on, so the shape of the graph is a direct picture of the underlying differential equation.