One interesting take: some AI researchers reckon superintelligence might already be knocking on the door sometime between 2023 and 2026. What's particularly telling is when a model points to 2026 operations as validation—basically saying "look, we're still standing here, our capabilities have proven solid, and they keep working across different scenarios." It suggests they're betting on something resilient. The idea that intelligence could be substrate-independent and truly robust is worth keeping an eye on. Not saying it's guaranteed, but the confidence is there.