The Convergence Wager

Beneath divergent laws, rituals, parables, and metaphors, the world's religions converge on the harmonization of human consciousness at the individual and civilizational scales.

Sin, suffering, ignorance, separation, disorder: each tradition conceptualizes the problem differently and prescribes a different solution, but the problem is the same.

This was always arguable. What's new is that we've built a technology whose native operation is finding shared structure beneath divergent symbols. Cross-symbolic translation is not something large language models can do; it is what they are. For the first time, the convergence claim can be tested, demonstrated, and translated between traditions in real time by an instrument that can hold each tradition's full symbolic depth while mapping it onto the others.

The challenge is that each civilization carries entrenched networks with decades of sunk costs bet against convergence, institutions whose authority rests on jurisprudence, ritual, and a claimed monopoly on truth, the very accretions that disconnected the traditions from their function. These networks will not reform quietly. Threatened elites externalize: they will provoke conflict abroad and manufacture chaos at home, because chaos is the technology of preventing shared understanding.

While AI serves incumbents as well as reformers, there is an asymmetry. Incumbents will use AI for an ancient task, propaganda at scale, but the talking points are old, with diminishing returns. Reformers can use it for a task that was previously impossible. One side gets a louder megaphone; the other gets a new sense organ.

And this is where it becomes a fascinating game-theoretic problem. Timur Kuran showed that under preference falsification, people privately hold beliefs they publicly suppress until a visible breakthrough somewhere makes honesty suddenly safe everywhere, and the cascade comes all at once. Which means the conflicts inside each civilization are not independent games. They are coupled through expectations. The probability that reformers prevail in one civilization depends on whether reformers appear to be prevailing in the others. Incumbents understand this instinctively: that is exactly why they will work to make foreign reform look catastrophic. War or peace between civilizations will be decided by these mutual assessments, whether each side believes the other's reformers are real, credible, and capable of restraining their own extremists.

Reformers will be outnumbered. But cascades don't require majorities; they require common knowledge, the shared visibility that lets private belief become public position.

So the hinge of the whole equilibrium is this: does AI lower the cost of verification faster than it lowers the cost of obfuscation? If yes, convergence becomes politically thinkable. If no, every tradition receives a machine for manufacturing deeper confusion in its own sacred language.

That question is empirically open. It may be the most consequential open question of the century.