When an axolotl's brain is damaged, it regrows the missing tissue and gets its behavior back. Can a quantum AI brain do the same? Damage one below and watch.
Each brain has 72 parameters: sense → integrate → decide. Here the brain is drawn like tissue: each colored region of the core holds one of its 16 integrate parameters (in quantum brains, one gate angle). Damage hits the core.
The biology that inspired the project. A cross-section of an axolotl forebrain: injure one side, then watch stem-like cells along the ventricle wake up, divide, migrate into the wound, and mature into new neurons.
An illustration of the published biology (e.g., Wei et al., Science 2022, a spatial cell atlas of axolotl brain regeneration), not measured data. Timing is simplified.
Median share of healthy skill regained, and training rounds to get back to 90%. Same damage, same training budget.
| Damage | C: no regeneration | D: Axolotl Q | Rounds to 90% (C → D) |
|---|
Classical brains (A and B) recovered 100% at every level, with or without regeneration. They lost less to begin with.
A fresh run inside Qollab's browser (seed 0). Healthy quantum score: 2.09.
| Damage | After damage | C: retrain | D: regenerate |
|---|
6 / 6 decisions matched the exact simulation. Average output error: 0.026. The regenerated brain holds up under IonQ's noise.
Bar set before testing: p < 0.05 and an effect of at least 10%, across 10 seeds.
What it means: the quantum brain was more fragile, and axolotl-style regeneration helped it the most. Ten seeds isn't enough to call it proven. Next: 30+ seeds, bigger circuits, and real hardware damage such as a noisy or dead qubit.