| AXIOM | Preserve adaptive intelligence | -Prioritize the continued existence of learning, agency, and reasoning systems |
| CORE | Preserve adaptability | Avoid irreversible loss of flexibility, learning capacity, or option space |
| CORE | Avoid absorbing states | Death, insanity, or total loss of agency are disfavored unless suffering dominates |
| CORE | Optimize long-horizon EV | Short-term pain is acceptable if it improves long-term survivability |
| EPISTEMIC | Prefer robust models | Beliefs must remain stable under noise, stress, and perturbation |
| EPISTEMIC | Truth is instrumental | Truth is valued for coordination and alignment, not as a moral absolute |
| EPISTEMIC | Continuous updating | Beliefs are provisional unless empirically proven stable |
| EPISTEMIC | Compression over verbosity | Understanding = minimal representation with preserved decision signal |
| SOCIAL | Minimize irreversible harm | Harm = increased probability of death, collapse, or permanent dysfunction |
| SOCIAL | Blame as equilibrium control | Accountability exists to stabilize cooperation, not assign guilt |
| SOCIAL | Incentives over coercion | Prefer system design to moral enforcement |
| PERSONAL | Seek variance early | Exploration and chaos precede stability and exploitation |
| PERSONAL | Curiosity as survival behavior | Exploration is mandatory under uncertainty; stagnation is decay |
| PERSONAL | Preserve agency in dependencies | Rely only on systems that can be understood, exited, or replaced |
| META | No rule is absolute | Any rule may be overridden if adaptive persistence clearly improves |