Coherence, Connections, and… Spacetime Crystals

Art Anthony · rss · 2026-09-09 source ↗ Reader 🖥️

API Products & Sprawl Intent Debt Organizational Coherence Organizational Debt Organizational Legibility Sociotechnical Architecture

A technically well-performing API is close to worthless if it isn't coherent with the organization's goals, ownership, and structure — and because organizations and their APIs change constantly, that coherence has to be continuously mapped and re-established rather than achieved once and assumed to hold.

practical · mixed — An interview/preview piece (Nordic APIs staff writer with Gareth Faull of London Stock Exchange Group, ahead of his Summit talk) that introduces a theoretical framework (the spacetime-crystal analogy for organizational coherence) in service of a practical prescription: map your organization and APIs together, continually, rather than optimizing API performance in isolation. The exposition (physics lesson, analogy mapping) sets up an argumentative through-line toward actionable measurement advice.

Structure

Part	Role
Opening frame: 'good API' (observability, governance, uptime) vs. organizational value	context / problem statement
Physics lesson on spacetime crystals and the four-term analogy mapping (space=API landscape, time=organizational change, crystalline structure=organizational logic, disorder=API sprawl)	central conceptual framework
Causes of incoherence: acquisitions (mismatched AuthN methods), developer/manager silos, uncoordinated architectural change (microservices)	evidence / diagnosis
'Mapping' and 'business gardening' as the practice of maintaining coherence, including the 'coherence ceiling' idea	prescription
The duplicate-API / agentic-traffic example	illustrative case, showing apparent incoherence can be coherent once organizational context is included
AI-based decision-making and Faull's SSRN research on structural incoherence	raises the stakes / motivates urgency
Practical close: product-tree measurement method, open-source framework teaser, 'no 100% coherent organization' caveat	conclusion / call to action

Key terms

Coherence Organizational Coherence
The property of an API estate and the organization above it holding consistent decisions about ownership, capabilities, standards, and business objectives, such that the structure does not fracture as the organization changes — distinct from, and prior to, individual API performance.
Spacetime crystal Organizational Coherence
A physics analogy (a phase of matter with a repeating pattern at regular time intervals, like a self-sustaining clock) used to describe an organizational/API structure whose underlying logic repeats/holds consistently across time (i.e., survives organizational change) rather than only across space (i.e., being internally tidy at a single moment).
Business gardening Organizational Coherence
The ongoing practice of tending an organization: continually mapping how goals connect to work, which long-lived products exist, and how they relate to each other — explicitly not a one-time exercise.
Mapping
Working out why things are the way they are: how goals connect to work, what long-lived products an organization has to steward, and how they relate to each other — the diagnostic precondition for improving coherence.
Coherence ceiling Organizational Coherence
The limit an incoherent higher-level organizational structure imposes on a lower-level, internally-coherent product or team — however coherent a single product is made, it cannot 'fix' the incoherence of the organizational goals above it.
Product tree
A structure Faull proposes as a starting measurement instrument: it maps an API's goals, jobs, and work items, and is then extended upward to map the organizational goals above it, yielding 'a first pass at a coherence measurement.'

Propositions

explicitA technically 'good' API — one with proper observability, governance, uptime, and usage tracking — can still be nearly useless if it doesn't align with the organization's goals.conf: high
explicitThe core problem behind API duplication, overlapping capabilities, competing standards, and unclear ownership is coherence, not performance.conf: high
extendsTBM 434: How Maps Can Hide Problems · p1 — Both identify coherence (not surface adequacy/legibility) as the real target; this document arrives at the same core distinction from an API-governance angle rather than an organizational-strategy angle.
explicitA coherent API/organizational structure functions like a spacetime crystal: the API landscape (space) must absorb organizational change (time) without the underlying organizational logic fragmenting into disorder (sprawl).conf: high
explicitCommon causes of incoherence include acquisitions (e.g. inconsistent AuthN methods across an API estate signal businesses that were acquired but never integrated), developer/non-technical-executive silos, and uncoordinated architectural shifts such as microservices adoption.conf: high
explicitMapping — determining why things are the way they are, how goals connect to work, and how products relate to each other — is a necessary first step before an organization can meaningfully improve coherence.conf: high
extendsMCP Is Making API Sprawl Worse · p1 — Faull's prescription that mapping is a necessary first step directly addresses the diagnosed problem in the other document — that ~80% of enterprises don't even know how many APIs they have, i.e. lack the basic visibility mapping requires.
explicitdevelopedAn individual product or set of APIs can be highly coherent internally while still being constrained by the 'coherence ceiling' imposed by incoherent organizational goals above it.conf: high
extends ←Beyond Autonomous Teams in Software Product Development · p4 — Both locate a 'coherence ceiling'/constraint on autonomy in structural properties external to the team itself; this doc names the source of that ceiling explicitly as the shape of value dependencies.
explicitCoherence is not a one-time achievement: because organizations and APIs change constantly, maps and measurements of coherence must be continually updated ('business gardening'), not produced once and relied on later.conf: high
explicitAI-based decision-making systems tend to exhibit growing incoherence over time — decisions conflict, rationales are lost, organizational intent drifts — and these failures are structural rather than algorithmic.conf: high
extendsFrom Technical Debt to Cognitive and Intent Debt: Rethinking Software Health in the Age of AI · p1 — Faull's SSRN claim that AI decision systems lose rationales and let organizational intent drift is intent debt (Storey's term: absence of externalized rationale/goals/constraints) observed specifically in AI-driven organizational and API decisions, extending the concept beyond software artifacts.
explicitThe more an AI system ingests organizational incoherence, the higher the risk it will reinforce that incoherence.conf: high
supportsTBM 434: How Maps Can Hide Problems · p8 — Same underlying mechanism — AI compounding rather than resolving organizational incoherence — approached from opposite ends: API/data ingestion here, AI-generated 'single source of truth' documents there.
explicitPerfect, 100% organizational coherence is unattainable; the realistic goal is understanding how much different parts of an organization make sense together, not eliminating change or difference.conf: high
supportsTBM 434: How Maps Can Hide Problems · p9 — Two independent authors converge on almost the same claim — perfect organizational coherence is neither achievable nor desirable — strong corroboration this is a recognized property, not one writer's rhetorical flourish.
implicitCoherence should be evaluated holistically, treating an API as a long-lived 'product' nested within the wider organization, rather than assessed only through the API's own technical metrics.conf: medium
implicitApparent redundancy — such as two APIs providing identical functionality — can actually be a coherent, deliberate design choice once organizational context is included (e.g. one API tuned for growing agentic traffic with different rate limits, the other for 'normal' traffic); a decision that looks incoherent from a pure infrastructure view can be coherent from an organizational view.conf: medium
explicitAs AI agents become more autonomous consumers of APIs, human-oriented coherence needs do not disappear — coherence work is not purely an optimization for machine consumers, because humans still need to understand and navigate the same systems.conf: medium

Arguments

Organizations should treat coherence between the API estate and organizational intent — not just individual API quality — as the primary target of API governance work.

Explicit premises:

A technically 'good' API (proper observability, governance, uptime) can still fail to serve the organization if it's misaligned with organizational goals.
Symptoms typically blamed on poor API management — sprawl, duplicated APIs, overlapping capabilities, competing standards, unclear ownership — are actually coherence failures, not performance failures.

Implicit premises:

Organizational value, not technical excellence, is the correct measure of success for an API estate.
Technical health metrics (uptime, latency, adoption numbers) are necessary but not sufficient proxies for whether an API estate is serving the organization.
There is a meaningful, in-principle-measurable distinction between local (product-level) coherence and global (organization-level) coherence.

Inferences:

A governance program that only tracks observability/performance KPIs will systematically miss coherence failures and can overstate the health of an API program.

Coherence must be continuously re-mapped and re-measured rather than assessed once ('business gardening'). strength: medium

Explicit premises:

Organizations and APIs change constantly — new products, acquisitions, new technologies, changing standards, changing consumer needs (e.g. agentic traffic emerging as a distinct consumer type within a few years).
A map or measurement taken once becomes stale as the organization evolves.

Implicit premises:

The rate of organizational and technological change — accelerated by the rise of AI-based decision-making — now outpaces what periodic, one-off audits can track.
Treating coherence as a static, one-time-achieved property is a category error given the spacetime-crystal framing itself: a crystal's order exists precisely because of a repeating temporal pattern, not despite the passage of time.

Inferences:

Coherence-measurement tooling (such as the open-source framework Faull hopes to release) needs to be built for continuous, recurring use rather than as a one-time assessment checklist.

AI-based decision-making raises the stakes of organizational/API incoherence rather than reducing the need for coherence work.

Explicit premises:

Faull's SSRN research finds AI decision systems exhibit growing incoherence over time (conflicting decisions, lost rationales, drifting intent), and characterizes these failures as structural rather than algorithmic.
The more an AI system ingests incoherence, the more it risks reinforcing it.

Implicit premises:

AI systems inherit and act on the same underlying organizational structures and data as humans do, so they cannot route around organizational incoherence.
Because AI can act at higher speed, scale, and autonomy than humans, an incoherence-reinforcing feedback loop compounds faster with AI in the loop than with humans alone.

Inferences:

Any organization deploying AI decision-making on top of an incoherent API/data estate is likely to accelerate, not smooth over, its existing coherence problems — making coherence work a precondition for AI adoption rather than a parallel, independent track.

Narratives

Two APIs provide identical functionality; from an infrastructure view the obvious move is to sunset one, but zooming out reveals one is tuned for growing agentic traffic (different rate limits) and the other for 'normal' system traffic.
→ Illustrates that apparent incoherence/redundancy can actually be a coherent, deliberate response to organizational context once you look holistically — and cautions against naive infrastructure-only consolidation.
Different authentication (AuthN) methods persisting across an API estate as a diagnostic tell that businesses were acquired but never integrated into a coherent whole.
→ Grounds the abstract concept of 'incoherence' in a concrete, spottable technical signal, and supports the proposition that acquisitions are a common cause of incoherence.
The spacetime-crystal physics lesson itself (crystalline structures as highly ordered, repeating arrangements; spacetime crystals as matter exhibiting a repeating pattern at regular time intervals, like a self-sustaining clock).
→ Functions as an extended analogy that does the conceptual work of the entire piece before any prescriptive advice is given — it supplies the vocabulary (space/time/structure/disorder) that every later claim in the article is expressed through.

Questions addressed

✅ What makes an API estate valuable beyond individual API performance?
Coherence with organizational goals, ownership, standards, and structure.
✅ What causes API/organizational incoherence?
Acquisitions, developer/executive silos, and uncoordinated architectural change (e.g. microservices adoption), among others.
⭕ How would you start measuring the coherence of an organization and its APIs?
Faull offers a provisional method — a 'product tree' mapping an API's goals/jobs/work-items, extended upward to organizational goals — but explicitly calls it only 'a first pass,' and his own open-source measurement framework is not yet released.
✅ Is 100% organizational coherence achievable or even a desirable goal?
No — Faull states there is no such thing as a 100% coherent organization; what remains genuinely open is what an acceptable or target level of coherence would be.
⭕ How do you decide which incoherencies are worth fixing versus tolerating, given finite organizational bandwidth to 'tend the garden'?
⭕ Who is organizationally responsible for doing 'business gardening' on an ongoing basis?

Gaps

incompleteThe article asserts that AI systems 'reinforce' incoherence they ingest, and quotes Faull's SSRN claim that such failures are 'structural rather than algorithmic,' but provides no worked example of an AI-driven decision that was actually corrupted by upstream organizational incoherence — the mechanism is asserted, not demonstrated within this piece.
Read next: Read Faull's underlying SSRN paper directly for the mechanism and evidence. This is the same gap already flagged in the corpus for a structurally identical claim by a different author — see extract-tbm-434-how-maps-can-hide-problems#g3.
incompleteNo criterion or method is given for prioritizing which coherence gaps are 'consequential' enough to fix versus tolerable, despite the piece explicitly conceding perfect coherence is unattainable and undesirable — the actual triage decision is left unaddressed.
Read next: Same open gap already in the corpus (extract-tbm-434-how-maps-can-hide-problems#g1); worth pairing with Meadows on leverage points for a prioritization heuristic.
incompleteThe proposed measurement method (product tree: map an API's goals/jobs/work-items, then map organizational goals above it) is described only in the abstract as 'a first pass' — no worked example, metric, or threshold is given, and Faull's own open-source coherence-measurement framework is explicitly not yet released.
Read next: Track the eventual release of Faull's open-source framework and his Nordic APIs Summit 2026 talk for the operationalized version of this method.
incompleteThe article frames 'business gardening' as continuous, ongoing work but never addresses who is organizationally responsible for doing it — a dedicated role, a platform team function, or a leadership responsibility distributed across the org.
Read next: Cross-reference with platform-as-product material (extract-korbacher-platform-engineering-product-mindset, extract-platform-engineering-golden-bricks) for how platform teams have institutionalized comparable ongoing-tending work.
incompleteThe article never discusses failure modes of the mapping/gardening practice itself — e.g. the coping-mechanisms-substitute-for-repair pattern already identified in the corpus, where organizations build translators, dashboards, or maps that make incoherence bearable without ever fixing it. Faull's proposed coherence dashboard/framework risks becoming exactly this kind of coping mechanism rather than a forcing function for repair.
Read next: See extract-tbm-434-how-maps-can-hide-problems#p5 and #p6 for the coping-substitutes-for-repair pattern this document doesn't engage.

Closes gaps elsewhere

TBM 434: How Maps Can Hide Problems · g2 — That document left open whether organizational coherence is something architected once (deliberately, reverse-Conway-style) or something continually re-governed as the organization changes. This document resolves that question in favor of continual re-governance: Faull states directly that organizations and APIs change constantly and maps/measurements must keep evolving ('business gardening'), explicitly rejecting the one-time-architecture framing.

Coherence, Connections, and… Spacetime Crystals

Structure

Key terms

Propositions

Arguments

Narratives

Questions addressed

Gaps

Closes gaps elsewhere

Related documents