Tony Seale · LinkedIn · Meshed With 4 Companion Documents

Seale's Law: Vocabulary Erosion, Meshed With a Four-Document Corroboration Set

“The precision of a word's meaning is inversely proportional to its commercial value as a marketing term.” Tony Seale's post reproduced verbatim with its full comment thread, then read against four independently authored documents — each showing a different AI-era term (context, ontology, System of Intelligence, trust boundary) undergoing the identical dilution, and each prescribing the same corrective: open-standards Linked Data.

KG curated by kg-generator + rdf-infographic-skill and Claude Sonnet 5 on behalf of Kingsley Uyi Idehen.

Section 1

The Post🔗

Reproduced verbatim, by Tony Seale.

Every technology wave eats its own vocabulary. Words start sharp. Then the market discovers them. “Cloud.” “Platform.” “Transformation.” “AI.” At first, each points at something real. Then it becomes useful in a sales deck. Then it becomes useful in every sales deck. The word survives. The meaning does not. So here is Seale’s Law: ⚡ The precision of a word’s meaning is inversely proportional to its commercial value as a marketing term ⚡ It is a joke - but only just. The painful irony is that this is now happening to semantics. For years, those of us working with ontologies and knowledge graphs have made a fairly unfashionable argument: meaning matters. Context matters. Data is not ready for machines simply because it has been loaded into a warehouse, lakehouse, dashboard or API. When concepts remain implicit, identities are inconsistent and business meaning is trapped inside departmental dialects, the system may appear integrated. It is not. AI has made that impossible to ignore. LLMs have exposed a truth that many analytics programmes could postpone: enterprise data is not connected enough, described clearly enough or governed around meaning well enough for intelligent systems to use it safely. AI does not simply need more data. It needs data with identity, relationships, context and explicit business meaning. That is good news. It is also dangerous. 🔵 “Semantics” is about to become a beautiful container for vague promises. The word will be attached to products, features, copilots, roadmaps and platforms. Some of it will be real. Much of it will be rebranding. Organisations moving quickly will be sold noise at precisely the moment they need clarity. The message is simpler than the market will make it. 🔵 Every organisation has a data integration problem. To solve it for AI, you must connect your data and organise that connected data through an ontology: an explicit model of the things your organisation cares about, what they mean and how they relate. That is phase one: internal coherence. 🔵 Phase two is arriving fast: interoperability. No organisation operates alone. Supply chains, regulators, partners, customers, research networks and financial systems all cross organisational boundaries. If AI is going to work across those boundaries, meaning has to travel too. That requires open standards. Not proprietary semantic theatre. Not another silo with “ontology” printed on the brochure. Open standards for identity, connectivity and shared models of meaning. Semantics is having its moment. Our task is to protect the substance before the market hollows out the word. Your AI strategy is only as strong as your semantic foundation. Connect the data. Model the meaning. Use open standards. Prepare for interoperability. And if Seale’s Law is right, do it before we need a new word for the thing that finally made the old one profitable. 🔗 Building your semantic foundation? Let’s talk: https://lnkd.in/ezHU2amU
👍 93 reactions💬 8 comments (7 visible without sign-in)
Section 2

Comment Thread🔗

All 7 comments visible without signing in, in original order.

#1 Arnaldo Sepulveda
Exactly. Prompt engineering is not a security boundary. The boundary is the platform: identity, policy, retrieval scope, tool authorization, observability, and audit.

The interesting question is not whether an LLM can be attacked, because it can. The question is whether the surrounding system can detect drift, constrain authority, prove provenance, and fail closed when the operating conditions change.

That’s the problem I’m working on with governed agentic systems.
Yep! Prepare for interoperability at Internet and Web scale as follows: • Connect the data — use hyperlinks as identifiers. • Model the meaning — describe things using machine-computable sentences built from terms in shared and/or homegrown ontologies. • Use open standards — use standards such as RDF for machine-computable statements and ontology construction. Basically, good old Linked Data principles manifesting a Semantic Web — a natural complement to the natural-language processing capabilities that Large Language Models (a.k.a. Langulators) have recently added to computing’s UI/UX stack. No more silos. Silos are extremely stubborn and often well disguised by marketing communications that blur or appropriate terminology as a survival strategy. Adopt Linked Data principles and you get the magic of the Web at an even more powerful level: data, information, and knowledge providing a critical context layer that AI agents can access, interpret, reuse, and share by reference. There’s no better foundation for this on the planet. ** posted deliberately using markdown! ** #Ontology #ContextLayer #SemanticLayer #SemanticWeb #LinkedData
#3 Oliver Cronk
We have an episode of Architect Tomorrow coming that is aligned to what you are saying here - talking about how ontology is falling into the marketing bucket. Aiming to publish it next month.
#4 Michael Bucknell
Love this image Tony Seale
#5 Andrea Splendiani
Totally agree, but any advise on how to “protect the meaning”? We already see “ontologies” mentioned everywhere, without people even thinking at the idea of connecting data...
#7 Andi Willmott
Totally agree.
What happens when a controlled vocabulary contents start to mean differing things to differing teams due to overuse, reuse or reinterpretation.
Remember when ‘lean working’ had one meaning - not now…🙄
Section 3 · Distinct Callout

Kingsley Idehen's Perspective🔗

The mechanism the post leaves implicit

Of the eight commenters on this post, Kingsley Uyi Idehen is the one who names the concrete mechanism Tony Seale's post leaves abstract: hyperlinks as identifiers, RDF, shared/homegrown ontologies — Linked Data principles manifesting a Semantic Web. He also observes that “silos are extremely stubborn and often well disguised by marketing communications that blur or appropriate terminology as a survival strategy” — Seale's Law, restated as a description of silo behavior itself.

Idehen is also the accountable person for this collection's curation, and for its mesh with four of his own previously published companion documents (see the Corroboration section below).

↑ Read the full comment in the thread above

Section 4 · The Thesis

Seale's Law and the Two-Phase Framework🔗

“The precision of a word’s meaning is inversely proportional to its commercial value as a marketing term.”

Seale's Law — Tony Seale
Phase One

Internal Coherence

Connect an organisation's data and organise that connected data through an explicit ontology — a model of the things the organisation cares about, what they mean, and how they relate.

Phase Two

Interoperability

No organisation operates alone — supply chains, regulators, partners, customers, research networks and financial systems all cross organisational boundaries. Meaning must travel too, via open standards.

Section 5 · Corroboration

Four Documents, Four Diluted Terms, One Corrective🔗

Each of these four documents was authored independently of Tony Seale's post, about a different named term, yet each traces the exact pattern Seale's Law predicts.

1

Case 1 — "Context" as a Proprietary Container

Diluted termContext / Context Engine
Corrective mechanismSix loosely-coupled, standards-addressable infrastructure layers — Identity, Identification, Authentication, Authorization, Data Spaces, and Ontology — rather than one vendor's proprietary compounding loop.
2

Case 2 — "Ontology" Mentioned Everywhere, Connected to Nothing

Diluted termOntology
Corrective mechanismOntologies re-adopted as genuine, queryable, dereferenceable logical guardrails for agent loops — the neurosymbolic-AI convergence of LLM language fluency with RDFS/OWL structure — rather than a brochure word.
3

Case 3 — Vendor-Owned Ontologies Recreate the Lock-In They Claim to Solve

Diluted termSystem of Intelligence / Ontology (vendor-branded)
Corrective mechanismAn enterprise-owned Knowledge Graph built on Linked Data principles and HTTP, queryable via SPARQL by any model — instead of a closed, proprietary graph (Palantir Ontology, Databricks Unity Catalog) recreating the walled-garden dynamic one layer down the stack.
4

Case 4 — A "Trust Boundary" That Still Needs a Representational Substrate

Diluted termTrust Boundary
Corrective mechanismThe five-pillar trust boundary (Control, Capability, Choice, Cost, Compound) framed as needing new legal instruments is, per the critical perspective, already achievable today via loose coupling over HTTP, an enterprise knowledge graph, and attribute-based access control — no new word or new law required.
Knowledge Graph

Explore the Knowledge Graph🔗

Interactive graph of every entity and relationship in the companion RDF — the post, its comment thread, the thesis framework, and the four corroboration cases. Click any node or edge label to resolve its full description via URIBurner.

— nodes / — links
Click outside to release zoom
Advanced Settings
Physics
Predicate Display
Predicate Filters
Node Type Filters
Literal Filter
Resolver
Arrow Style
Classes Properties Instances Graph note: kgData embedded from companion RDF at generation time.
FAQ

Frequently Asked Questions🔗

Fifteen questions covering the post, the thread, and its mesh with four corroborating documents.

Tony Seale's formulation: 'The precision of a word's meaning is inversely proportional to its commercial value as a marketing term.' As a term becomes profitable to say in a sales deck, its precise meaning erodes even as the word itself keeps circulating.

Terms like 'Cloud', 'Platform', 'Transformation' and 'AI' each started by pointing at something real, then became useful in every sales deck; the word survived while the precise meaning did not.

It is about to become 'a beautiful container for vague promises' — attached to products, features, copilots, roadmaps and platforms, some real and much of it rebranding, right as organisations need clarity most.

Phase One is internal coherence: connect your data and organise it through an explicit ontology. Phase Two is interoperability: because no organisation operates alone, meaning must travel across organisational boundaries too, which requires open standards.

Connecting an organisation's data and organising that connected data through an ontology — an explicit model of the things the organisation cares about, what they mean, and how they relate.

Because supply chains, regulators, partners, customers, research networks and financial systems all cross organisational boundaries, AI working across those boundaries requires meaning to travel too — via open standards, not proprietary semantic theatre.

He restated the post's three-part imperative concretely: connect the data using hyperlinks as identifiers, model the meaning using shared/homegrown ontologies, and use open standards such as RDF — Linked Data principles manifesting a Semantic Web, a natural complement to LLMs' natural-language capabilities.

She asked how to 'protect the meaning' in practice, noting that 'ontologies' is already mentioned everywhere without people connecting it to the actual idea of connecting data — a live instance of the same dilution Seale's Law predicts, happening to 'ontology' itself.

Sierra's 'Context Engine' brands customer context as a proprietary, vendor-owned moat; the critical-perspective response shows the same compounding value is achievable via six loosely-coupled, standards-addressable layers (identity, authentication, authorization, data spaces, ontology) instead — the diluted term is 'context', the corrective is open infrastructure.

It documents AI engineers rediscovering ontologies as guardrails for agentic systems from four independent voices — the exact word ('ontology') Andrea Splendiani flags as already overused in this thread, being restored to substance via neurosymbolic AI and dereferenceable, queryable graphs.

Vendor-owned 'System of Intelligence' ontologies (Palantir Ontology, Databricks Unity Catalog) recreate the lock-in they claim to solve — the diluted term is 'ontology' rebranded as a proprietary control point; the corrective is an enterprise-owned Knowledge Graph on Linked Data principles, queryable via SPARQL by any model.

Its 'trust boundary' is framed as needing a new legal instrument, patents' functional equivalent; the critical perspective shows the same boundary is already achievable today via open standards — HTTP loose coupling, an enterprise knowledge graph, and attribute-based access control — no new word or law required.

Because a closed, proprietary graph 'printed on the brochure' as an ontology is exactly the 'proprietary semantic theatre' the post warns against — the word 'ontology' retains commercial value as a feature name while losing the openness that gave the underlying concept its original precision.

RDF and RDFS/OWL for machine-computable statements and ontology construction, HTTP-addressable hyperlinks as identifiers, and SPARQL for querying — Linked Data principles manifesting a Semantic Web, per Kingsley Idehen's comment and all four meshed companion documents.

Connect the data, model the meaning, use open standards, prepare for interoperability — and do it before the market hollows out 'semantics' the way it already hollowed out 'Cloud', 'Platform', 'Transformation' and 'AI'.

Glossary

Glossary🔗

Terms from the post alongside terms reused from the four meshed companion documents.

Seale's Law

Tony Seale's coined principle: a word's meaning-precision is inversely proportional to its commercial value as a marketing term.

Vocabulary Erosion

The pattern Seale's Law formalizes: a technology term points at something real, becomes useful in a sales deck, then survives as a word while its meaning does not.

Semantic Coherence

The state in which an organisation's concepts are explicit, identities are consistent, and business meaning is not trapped inside departmental dialects.

Internal Coherence (Phase One)

Connecting an organisation's data and organising it through an explicit ontology.

Interoperability (Phase Two)

The requirement that meaning travel across organisational boundaries via open standards, since no organisation operates alone.

Ontology

An explicit model of the things an organisation cares about, what they mean, and how they relate.

Linked Data

Using hyperlinks as identifiers to connect data across organisational and system boundaries — Kingsley Idehen's named mechanism for Phase One.

Semantic Web

The open, standards-based vision Linked Data principles manifest; named directly in Kingsley Idehen's comment on this post.

RDF

The open standard for machine-computable statements and ontology construction that Kingsley Idehen's comment names as the concrete Phase Two mechanism.

Ontology Layer (reused)

From the Context Engine vs. Context Infrastructure companion document: the layer enabling deterministic reasoning via RDFS/OWL, one of six loosely-coupled infrastructure layers proposed as the antidote to a proprietary 'Context Engine'.

System of Intelligence (reused)

From the Karp/theCUBE companion document: the governed layer of business rules, policies, processes and tacit knowledge that vendor-branded ontologies claim, and often fail, to hold openly.

Trust Boundary (reused)

From the Reverse Information Paradox companion document: a hard boundary across which nothing crosses without consent — argued achievable today via the same open standards Seale's post prescribes.

Neurosymbolic AI (reused)

From the Ontologies Are So Back mesh companion document: the fusion of probabilistic LLM agents with symbolic ontology systems that restores substance to the word 'ontology'.

HowTo

How to Build a Semantic Foundation Before the Word Is Hollowed Out🔗

The post's own four imperative steps.

  1. 1

    Connect the data

    Use hyperlinks as identifiers to connect data across systems and departments, per Kingsley Idehen's comment — the concrete first move of Phase One.

  2. 2

    Model the meaning

    Organise the connected data through an explicit ontology — a model of the things the organisation cares about, what they mean, and how they relate — using shared and/or homegrown ontologies expressed as machine-computable sentences.

  3. 3

    Use open standards

    Adopt open standards such as RDF for machine-computable statements and ontology construction, instead of proprietary semantic theatre or another silo with 'ontology' printed on the brochure.

  4. 4

    Prepare for interoperability

    Because supply chains, regulators, partners, customers, research networks and financial systems all cross organisational boundaries, ensure meaning can travel across those boundaries too — Phase Two, arriving fast.

About

About This Page🔗

This knowledge graph collection reproduces Tony Seale's LinkedIn post and its comment thread verbatim, then meshes the thesis against four companion documents already published in this corpus. The original document was transformed into RDF (schema.org + a lightweight local ontology) using kg-generator, then rendered as this HTML infographic using rdf-infographic-skill, powered by Claude Sonnet 5.

Technology Stack: