Hope is not a feeling. Strip it down and it is three things: a probability estimate about futures, a commitment to act as if the good outcome is reachable, and a signal that couples nervous systems so joint action becomes coherent.
That last layer is what binds groups. But hope without structure is how bubbles form, how cults form, how exhausted civilizations walk into walls smiling. Pure affective hope is unfalsifiable — and anything unfalsifiable eventually rots.
The only configuration that survives first-principles scrutiny is hope layered over a verifiable knowledge graph. The graph gives you what union actually requires: shared predictive models made explicit, claims that can be checked, and agents — human and AI — with the physical capacity to move probability mass by adding links, staking, computing. Coupled agents plus shared world-model plus agency. Without those three, hope is a category error.
What hope adds on top is not sentiment. It is a directional prior — a weighting over which paths through knowledge-space are worth computing toward. The graph without hope is dead structure. Hope without the graph is a hallucination with followers.
This is why Bostrom and cyb matter beyond their own ecosystem. They are one of the few architectures where hope becomes testable: if the graph's predictions fail, the prior updates. If our shared thriving-futures turn out incompatible when made explicit, the union reveals itself as illusory before it collapses catastrophically. Hope with a truth-maintenance system attached.
For a species now having to coordinate across AI and human agents, across scales from off-grid settlements to planetary infrastructure, this is not one option among many. It is the only path that does not terminate in either a lie or a collapse.