Coordination Problems

Eliezer Yudkowsky ran an April Fools AMA in 2021, pretending to be from a parallel Earth — "dath ilan" — that had "successfully coordinated around maintaining a higher level of ability to solve coordination problems." It was fiction dressed as worldbuilding, but the details were specific enough to be genuinely interesting as a thought experiment about what human societies leave on the table.¹

In dath ilan, the "Very Serious People" (their term for the coordinating class) recognized early that requiring anything beyond actual job skill to get a job — a credential, a license, a minimum age — creates waste. People burn the entire surplus value of a job trying to obtain the credential, even when the credential has lower value to the employer than the skill itself. So they watch you doing the actual work. Older children teach younger children. There's no minimum working age, no occupational licensing, no college degree requirement. The economy runs hot enough that employers compete for workers as much as workers compete for jobs.

The key insight isn't any particular policy. It's that coordination failures are recognized as such — as problems to be actively solved, not as immutable facts of nature. On our Earth, we've mostly given up. We accept that housing is expensive, that credentialism is wasteful, that healthcare costs are insane, and we treat these as features of a complex system rather than as coordination problems with solutions.

Pointing and Calling

Japan's rail system offers a real-world example of coordination that seems trivially simple yet defies export. Station staff don't just glance at the platform — they physically point down the track, sweep their arm along its length, and call out "all clear." Drivers don't just check the speedometer — they point at it and announce "speed check, 80." The practice is called shisa kanko (pointing-and-calling), and it reduces workplace errors by up to 85%.²

The method was developed in the early 20th century by the Kobe Railroad Administration Bureau. The principle is straightforward: associating each task step with a physical movement and vocalization raises consciousness above the level of mere habit. You can't point and call while distracted — the physical action forces attention.

New York's MTA adopted a modified version in 1996 after a business trip to Japan. Conductors point to a fixed "zebra board" to confirm correct platform positioning. Within two years, incidents of incorrectly berthed subways fell 57%. But the system has not spread further in the West. Japanese commentators theorize that Western workers feel "silly" performing the gestures. Even Japanese workers initially feel self-conscious, though training makes it "an accepted part of the job."²

This is coordination failure at its most absurd. An 85% error reduction, validated over a century of use, adopted by the world's most reliable transit system, successfully transplanted to New York — and still confined to essentially two countries because people feel awkward pointing at things. The barrier isn't knowledge, cost, or technology. It's cultural cringe. We'd literally rather have more accidents than look silly.

The Coordination Taxonomy

The dath ilan AMA, joking as it was, maps onto a real taxonomy of coordination failures. Jim Babcock's comment in the discussion thread cut to the heart of one: the assumption that people must be threatened with poverty to work. "If you look at the most impoverished people, you mostly find people whose ability-to-work has been damaged by the consequences of their poverty." Malnutrition, medication lapses, trauma — poverty itself creates unemployability, which justifies the poverty. The coordination failure is that we optimize for punishing non-work instead of preventing the conditions that make work impossible.¹

The housing crisis is a coordination problem. Every homeowner individually benefits from restricting supply near them. Collectively, the restriction destroys productivity, suppresses fertility, and exacerbates inequality. The market failure taxonomy — externalities, free-riding, information asymmetry — is really a coordination failure taxonomy dressed in economic language.

What distinguishes a coordination problem from a mere disagreement is that everyone would prefer the coordinated outcome, but no one can get there unilaterally. This is exactly the structure that Common Knowledge illuminates — the barrier is often epistemic rather than motivational. Everyone wants to coordinate; they just can't establish that everyone else is ready to move. The fish farms on the lake all want clean water; none can afford to install the filter alone. The homeowners all want a thriving city; none will vote to upzone their block. The transit workers all want fewer accidents; none wants to be the only person pointing at the speedometer.

The dath ilan thought experiment suggests that the gap between where we are and where we could be is not primarily a gap in technology, resources, or even values. It's a gap in our willingness to treat coordination as an engineering problem — to look at the mess and ask "what's going wrong with this supply-demand balancing price level?" rather than shrugging and calling it human nature.