A LinkedIn Post by Tony Seale
From Chaos to Order
The universe tends towards disorder. That’s the second law of thermodynamics in a nutshell. And yet, life exists. Stars form. Galaxies coalesce. How does breathtaking order emerge from chaos?
Tony Seale
Semantic AI & Knowledge Graph Strategist, author of the post.
The LinkedIn Post
Published June 26, 2024, exploring the implications of emergent order for organizational structure.
The Science of Emergent Order
The post builds on fascinating concepts from physics to explain how structure can arise spontaneously.
Second Law of Thermodynamics
In closed systems, entropy (disorder) always increases. But reality is a web of open systems.
Dissipation-Driven Adaptation
Physicist Jeremy England's theory: systems pushed by external energy reorganize to better dissipate that energy, leading to structure.
Dissipative Structures
Mentioned by commenters, Ilya Prigogine's work on how open systems exchange energy with their environment to create order (negentropy).
The Analogy: Organizational Entropy
Just as physical systems can create pockets of order by exporting chaos, our organizations have done the same, with disastrous results.
Local Optimization
We drew boundaries around applications and databases, incentivizing teams to perfect their own small systems.
Exporting Chaos
Each optimized application, in minimizing its own internal disorder, exported chaos to the wider organization.
The Result: Data Silos
This created immense 'organizational entropy': brittle integrations and isolated data that cripple our ability to adapt and innovate.
The Solution: The Knowledge Graph
To become AI-native, we must dissolve the artificial walls between systems. The Knowledge Graph is the technological pattern for this shift.
Knowledge Graph
A shared connective tissue that creates a coherent, intelligent whole from warring, entropic parts.
How to Weaken Application Boundaries
1. Ontological Core
Establish a shared model of key concepts and relationships.
2. Map Data Sources
Connect existing, siloed data using ontologies.
3. Implement Progressively
Implement for a relevant domain and expand the Knowledge Graph over time.
4. Build Semantic APIs
Expose unified data, allowing new apps to be built on a coherent foundation.
Key Questions Answered
How did 'organizational entropy' arise?
It arose from creating boundaries around applications for local optimisation. This practice caused each system to export chaos—like data silos and brittle integrations—to the rest of the organisation.
What is the function of a Knowledge Graph in this context?
A Knowledge Graph acts as a shared connective tissue of links and ontologies. It dissolves the artificial walls between systems to build a coherent, intelligent whole from disparate parts.
Why is this shift more than a 'data project'?
It is considered a profound technological, political, and philosophical shift that redefines the boundaries of an organisation to prepare it for a future where adaptability is the ultimate competitive advantage.
Under the Hood: Exploring the Knowledge Graph
Key Facts
- Uses terms from the Schema.org Ontology.
- Generated using the OpenLink AI Layer (OPAL).
- Deployed to a public Virtuoso Data Space instance.
Explore Directly
How This Document Was Created
- Read the document of interest (the original LinkedIn post in this case).
- Use the OpenLink Structured Data Sniffer (OSDS) browser extension to trigger the document's upload to the OpenLink AI Layer (OPAL) for processing via one of its loosely-coupled Large Language Model (LLM) bindings.
- Export the generated Knowledge Graph from OPAL to a Virtuoso instance (or any other SPARQL 1.1 compliant endpoint that supports secure INSERTs via the Web).
- Copy the Knowledge Graph from OPAL and provide it to a Google Gemini instance, requesting the generation of this HTML-based infographic based on the information in the graph.
This workflow demonstrates the original vision of the Semantic Web project: connecting documents and data using the fundamental web standards of HTTP, HTML, and RDF. Here, RDF is used to represent both the descriptive metadata and the rich relationships between entities, with Linked Data principles emphasizing the critical role of hyperlinks and persistent entity names (IRIs).
Community Insights & Discussion
"...As we enter the Age of AI, a similar shift is unfolding—this time focused on entities and their relationships..."
- Kingsley Uyi Idehen
"Good silos can should spread output usable order. Bad silos export chaos."
- John Evans
"...My early 70s physics education (e.g., Ilya Prigogine and his 'dissipative structures'...) helped me to understand that this could also be described in terms of entropy/disorder/decay and negentropy/order/organization."
- Roy Roebuck
Foundational Reading
The author shared these links for additional context.