WIP Discussion Notes

Works in Progress Issues 19-21 Discussion

5 min read

Works in Progress Issues 19-21 Discussion

Session: 7 February 2026 Context: Preparation for a Rationalist discussion group, covering three articles from Works in Progress magazine. The discussion ranged well beyond the articles into historical parallels, epistemological questions, and meta-themes about institutional design.

Source Articles

Core Propositions

On competition and monopoly

On water and geopolitics

On cities and land

Cross-cutting themes

Discussion Arc

The conversation began with article summaries and critique, then developed through several thematic arcs:

Arc 1: The monopoly article’s ideological centre of gravity. The Olley-Pakes framework was challenged as having an economist-liberal bias — trying to salvage the principle of competition rather than recognising that good service doesn’t necessarily require it. The AT&T/Bell Labs case proved central: the monopoly period produced extraordinary innovation while scoring poorly on the decomposition. This led to the distinction between static and dynamic efficiency and the recognition that for network goods, the question is governance, not competition.

Arc 2: Historical parallels in railways and telecoms. Victorian railway competition, the AT&T breakup, and UK rail privatisation all demonstrated that competition in network infrastructure produces waste, and the market’s own resolution is consolidation toward monopoly. The full arc — competition → waste → consolidation → nationalisation → privatisation → dysfunction → reconsolidation — undermines any simple narrative about competition as universally beneficial.

Arc 3: Geographic constraint and institutional innovation. Why Japan developed land readjustment and Britain didn’t may be better explained by geographic constraint (73% mountainous) than by culture. This connected to the Nile (Egypt’s narrow floodplain producing centralised resurveying), America’s abundant land enabling sprawl, and Britain’s green belt producing constraint without a complementary reorganisation mechanism.

Arc 4: The Ise Shrine principle. The cyclical rebuilding of Ise Grand Shrine (every 20 years since the 7th century) preserves craft knowledge and enables incremental adaptation. Extended to urban planning, institutional design, and treaty structures: perhaps the question isn’t which structure is correct but how to build in periodic restructuring.

Arc 5: Pareto improvement and its limits. The gap between potential and actual Pareto improvement emerged as a meta-theme: each article proposes a positive-sum mechanism, but each involves potential losers whose compensation is uncertain.

Arc 6: Measurement, Goodhart, and the adjacent possible. Goodhart’s and Campbell’s Laws as convergent discovery through the adjacent possible. All three articles propose new metrics — each vulnerable to Goodharting if adopted as policy targets.

Arc 7: AI, Hayek, and preference formation. Whether AI enables central planning was explored and largely rejected: the Hayekian objection is epistemic, not computational. But the deeper question is whether platforms create rather than reveal demand, which undermines the foundations of both market and planning-based systems.

Questions Worth Pondering

  1. Legibility and blind spots. Each article makes a system legible for intervention (Olley-Pakes, land value, water dependency). What is each metric unable to see, and what are the consequences? (cf. James C. Scott, Seeing Like a State)

  2. Reform versus bypass. When is working within the existing structure right, and when is the whole framework the problem? Railways and telecoms suggest that sometimes the market structure itself is misconceived.

  3. The Ise Shrine question. Should institutions be designed with explicit sunset clauses and rebuilding cycles? The Indus Waters Treaty (65 years), British Rail (45 years), AT&T (70 years) — all accumulated rigidity until catastrophic failure.

These concepts were mentioned but not yet developed into full notes — they mark the edges of this session’s thinking and are candidates for future development:

Graph View

Backlinks