Vanderbilt Owen GSB · Session 1 of 14 · March 9, 2026

Welcome to the
Serendipity
Engine

AI-Accelerated Entrepreneurship Practicum

You're here to build real businesses faster and smarter — using AI as your always-available development partner.

🚀 Ship over perfect 🤝 AI as co-founder 🏙 Nashville orgs · real users 💼 No tech background required
Professor
Oliver Luckett BA '96
Teaching Assistant
Baxter Webb
Format
Mon & Wed · Mar 9 – Apr 22
Sessions
14 sessions · 21 slides
Before we begin · The Arc

30 Years of the Same Bet

Each company was the same insight applied to a different information system.

📡
1996–2001
Qtalk · Qwest
Knowledge Arbitrage
The 1996 Telecom Act opened a deregulation window. VOIP was the protocol exploit. Incumbents thought they sold calls. We knew they sat on a packet network.
📹
2005
Revver
Object Monetization
First platform to share revenue with creators. The creator economy started here, years before it had a name. The file IS the contract. Attribution traveled with content, not container.
🏰
2008–2015
Disney · theAudience
Enterprise Attention
Head of Innovation, Disney. theAudience shaped digital identity for Obama, Coachella, Pixar, AmEx, Star Wars, Toy Story 3 — thousands of brands and celebrities.
🌿
Always
Beyond the Code
Full Stack Builder
Niceland Seafood. HausMart. Yazoo Yaupon. The same pattern applies everywhere: find the gap, build the protocol, own the layer. Technology is just the medium.
Now
Inkwell Technology Studios
Convergent Intelligence
Artiquity — artist rights, authenticity and attribution in AI. Enginuity — near-symbolic awareness for large-scale construction and creative projects. The architecture where every prior insight converges.
🧵
The Thread

Every venture was the same bet: there's a knowledge gap opening in an information system, and if you see it before the incumbents do, you can build the protocol. Telecom deregulation. Object-level attribution. Programmable attention. Convergent AI. The gap is always there. Your job — in this course and in your careers — is to see it before anyone else does.

The Core Problem

Every AI System Today
Sees Only One Graph

That's why they hallucinate. That's why they feel hollow. That's why they miss the room.

🤖
ChatGPT / Claude
Generative + some Knowledge. No Social layer. Gives the same answer to everyone regardless of who's asking, why they're asking, or what they actually need.
"Here's a recipe for pasta" (whether you're a chef or a 10-year-old)
📊
Recommendation Engines
Social behavior data + some Generative. No Knowledge layer. Optimizes for engagement, not truth. Sends you down rabbit holes with zero factual grounding.
"You might also like…" → 3 hours lost
📚
Knowledge Bases / RAG
Knowledge-heavy but no Social or Generative layers. Returns correct facts to the wrong person in the wrong context in the wrong format. Technically accurate. Practically useless.
"See page 47 of the manual." (when you're in crisis)
The Origin Story · July 9, 2025
The Question That Started Everything
"What do all words
ending in -ity have in common?"
The answer: -ity words are abstract nouns that convert observable qualities into concepts that can be reasoned about as distinct entities.

Clarity turns "clear" into a thing you can measure. Possibility turns "possible" into a space you can navigate. Serendipity turns "lucky" into a system you can engineer.
The insight: English already built the ontology. We just had to see it.
The Vocabulary · 225+ Terms · 7 Categories

Language Already Built the Ontology

claritycreativitycomplexity curiosityserendipityauthenticity possibilitysynchronicityvitality integrityvulnerabilitycommunity connectivityreciprocitylucidity necessitytotalitycontinuity materialityindividualitypropensity felicityfamiliarityubiquity dignityequalityauthority gravityfluiditysensitivity + 195 more…

Each term is an ontological primitive — a concept that bridges human emotional experience and structured machine reasoning. They exist at the intersection of psychology, philosophy, and data science.

The Architecture

Three Graphs. One Convergent System.

🧠
Layer 1 · Social Graph
Who Cares
"Who is this for and why?"
The human substrate. Emotions, psychology, relationships, trust, belonging, identity. Without this layer, your system is broadcasting into a void.
authenticityvulnerabilitytrustbelonging
📚
Layer 2 · Knowledge Graph
What's True
"What do we actually know?"
Verified facts, domain expertise, provenance, confidence scores. Without this layer, your system is hallucinating with confidence.
claritycertaintyintegrityvalidity
Layer 3 · Generative Graph
What Could Be
"What can we create?"
Synthesis, transformation, emergence, serendipity. Without this layer, your system is a search engine that learned to type full sentences.
creativityserendipitypossibilitynovelty
Bridge Layer
authenticity clarity connectivity creativity vitality serendipity integrity possibility
Layer 1 Deep Dive

The Social Graph:
Who Cares?

The Social Graph answers the most important question in any startup: who are these people and what do they actually feel? Not what they say. Not what they click. What they feel.

  • Person nodes — identity, history, role, context, needs
  • Relationship edges — trust, context, shared experience
  • Community clusters — groups bound by shared -ity states
  • Emotional properties — anxiety, curiosity, vulnerability levels
  • Temporal dynamics — relationships that evolve over time
Example
CREATE (p:Person {
  name: "Maya Chen",
  role: "First-year MBA",
  anxiety: 0.7,
  curiosity: 0.9,
  belonging: 0.4
})
Key Insight
Two people asking the same question need completely different answers if one has high anxiety + low belonging and the other has high curiosity + high confidence. A Social Graph knows the difference.
Layer 2 Deep Dive

The Knowledge Graph:
What's True?

A Knowledge Graph is not a database. It's a structured epistemology — a system where every fact knows where it came from, when it was verified, and how confident we should be in it.

  • Entities — real-world objects with verified attributes
  • Relationships — facts as edges with confidence scores
  • Provenance — every fact has a source and timestamp
  • Ontology — the rules that define what's valid
  • CONFIRMED / CITED / INFERRED / ASSUMED — the IAM provenance doctrine
RDF Triple Pattern
<Subject> <Predicate> <Object>

Mercury → orbitsAround → Sun
confidence: 1.0 · source: NASA
Provenance States
✓ CONFIRMED 📄 CITED ~ INFERRED ? ASSUMED
Layer 3 Deep Dive

The Generative Graph:
What Could Be?

The Generative Graph is not "use ChatGPT." It's a structured map of creative possibility — the patterns that determine what can be synthesized, combined, or invented from existing knowledge.

  • Pattern nodes — proven creative combinations in a domain
  • Synthesis edges — which concepts produce novelty when combined
  • Serendipity routing — engineered surprise vs. random generation
  • Context-sensitive creation — Social + Knowledge informs what to generate
  • HybridRAG — Retrieval-Augmented Generation from all 3 graphs simultaneously
The Difference
Standard LLM: "Generate a marketing email."

Trinity Graph: "Generate a marketing email for Maya (anxious, curious, first-year MBA) about her client project based on verified data about her partner organization, using the tone that worked with similar users last quarter."
Serendipity Score
Every generated output gets a serendipity score: how surprising yet relevant is this? High serendipity = discovery. Low serendipity = expected. The Graph lets you tune the dial.
The Generative Graph Without Trinity

Alone, It Performs Knowledge.
It Performs Empathy.

Neither is real. Both are dangerous.

Failure Mode 1 · No Knowledge Graph
Hallucinated Empirical Facts
"Studies show that students who study in groups
retain 47% more information than those who study alone."
↳ This statistic does not exist. No study was cited. It sounds plausible.
"The median Series A raise in 2024 was $8.4M,
down 12% from 2023's $9.6M peak."
↳ Specific, confident, sourced-sounding. Fabricated.
"The FDA approved semaglutide for adolescents
in March 2023 following a 14-month clinical trial."
↳ Plausible medical fact. Partially wrong date and trial duration. Told with zero caveat.
Without a Knowledge Graph: there is no ground truth to check against. The model generates the most statistically likely next token — not the most accurate fact. It performs knowledge.
Failure Mode 2 · No Social Graph
Feigned Confidence + Social Emotions
User: "My startup idea is to sell bottled water door-to-door."
GPT: "That's a great idea! Door-to-door models can build incredible customer loyalty…"
↳ Should have said: no, this is a bad idea. But without a Social Graph, it can't modulate confidence.
User: "I just lost my job."
GPT: "I completely understand how difficult this must be. Here are 7 steps to update your LinkedIn profile…"
↳ Performed empathy. No idea if this person needs comfort or urgency. Pivots to a list.
User: "I'm so stressed about my MBA finals."
GPT: "Of course! I'd be happy to help. Let me know what you need! 😊"
↳ Performative enthusiasm. No model of who this person is, what stress means for them, what they actually need.
Without a Social Graph: the model doesn't know who it's talking to. It applies the same emotional register to everyone. It performs empathy — statistically likely warmth tokens — not genuine response to a known emotional state.
The Fix

Knowledge Graph → ground truth check before any claim is generated. Social Graph → user context before any emotion is performed. The Trinity Graph doesn't make AI smarter. It makes AI honest.

The Convergence

When All Three Converge,
You Get Awareness

🧠 + 📚
Social + Knowledge
Contextually correct. Knows the facts and who's asking. But can't create anything new.
📚 + ✨
Knowledge + Generative
Grounded creativity. Generates truthful content. But doesn't know who it's for.
🧠 + ✨
Social + Generative
Empathetically creative. Personal and imaginative. But may hallucinate facts.
🧠 + 📚 + ✨
All Three = Convergent Intelligence
Knows who. Knows what's true. Creates what's needed. That's awareness.
The Omega Protocol

10 Runes · 10 Stages of Awareness

Moves AI from linear prediction into System 2 reasoning. Each rune must be completed before moving to the next. Your projects advance through these states.

Materiality
What exists?
Vitality
What's alive?
Interiority
What's felt?
Criticality
What breaks?
Connectivity
What links?
Lucidity
What's clear?
Necessity
What must exist?
Reciprocity
What exchanges?
Totality
What's whole?
Continuity
What endures?

Your semester arc: Week 1 = Materiality (what exists in your domain). Week 14 = Continuity (what your system will sustain).

Required Reading · Connection
"Social networks are not platforms.
They are organisms."
The Social Organism · Luckett & Casey · 2016
🫀
Metabolism
Content velocity, energy flow, viral dynamics
🛡️
Immune System
Community standards, moderation, antibody response to misinformation
🌱
Reproduction
Sharing, memes, cultural propagation through the network

Read Chapter 1 before Week 2. Come ready to draw your startup domain as a biological system.

The Course Structure

Build Real Products.
For Real People.

Your team builds one product for a Nashville artist, collective, or community AND develops your own venture. The client project sharpens your listening. The venture sharpens your conviction. Both ship by April 22.

Venture Build Project · 35% of Grade
Nashville Artist, Collective or Community
A team-built product deployed for a real Nashville artist, collective, or community. Real stakeholders. Real creative constraints. Something that runs after you graduate.
  • What you deliver: working AI-powered product + deployed to real users
  • What you learn: listening, stakeholder management, iteration under constraints
  • What it teaches: empathy is not soft — it's your most powerful design tool
Grading Breakdown
Hands-On AI
25%
Venture Build Project
35%
Graph Architecture Model
15%
Investor Pitch & Demo
15%
Participation & Engagement
10%
Final Demo Day: April 22. Live pitch to investors and industry leaders. 15 minutes per team.
The 14-Session Arc · Mar 9 – Apr 22

7 Weeks · 14 Sessions · 2 Deliverable Tracks

S1–2
Mar 9–11
Foundations — Serendipity Engine · AI Tool Landscape
Team formation. Co-founder selection. AI tools survey. Sprint methodology. The Social Organism framework. First product prototype.
S3–4
Mar 16–18
Generative AI & Agentic Systems
Claude, Gemini, ChatGPT as development partners. AI agents and workflow automation. Build first AI agent.
S5–6
Mar 23–25
Validation & Social Graphs
10 customer discovery interviews. AI-assisted market research. Network theory. Social graph design for your startup concept.
S7 ★
Mar 30
★ MIDTERM — Knowledge Graphs & Ontology
Ontology design. Concept linking. Retrieval augmentation. All teams present midterm. Design knowledge and generative graph layers.
S8–10
Apr 1–8
Intelligent Applications · Business Model · Ship MVP
Graph-powered features. Revenue models for AI ventures. Deploy to real users. Gain beta use.
S11–12
Apr 13–15
Storytelling · Scale · Ethics & Artist Rights
Pitching AI ventures. Narrative frameworks. Artiquity framework. Responsible AI. Build your pitch deck.
S13–14 ★
Apr 20–22
★ FINAL — Integration Review & Investor Pitch Day
Final product polish. Pre-pitch dry runs. Apr 22: Live pitch to investor panel. 15 minutes per team. Final deliverable due end of class.
The Business Case

The Future Competitive Moat
Is Not Model Size

❌ Wrong Bet
"We'll win because we use GPT-5."

Everyone uses GPT-5. The model is a commodity. You're building on rented land.
✓ Right Bet
"We'll win because our Trinity Graph captures 3 years of domain-specific social patterns + verified knowledge that no competitor can replicate."
❌ Wrong Bet
"We fine-tuned a model on our data."

Fine-tuning is erasable. Data can be replicated. There's no structural moat here.
✓ Right Bet
"Our Social Graph has 18 months of real user emotional data. Our Knowledge Graph has 50,000 verified domain facts with provenance. That's the moat."

"The future of AI will be defined not by the size of its models, but by the integrity of its architecture." — Oliver Luckett

AI-First Development

Your Toolkit — Use AI for Everything

🗄️
Neo4j Aura
Free tier · Your graph DB. Cypher from Day 1. Powers all three layers.
Free
🤖
IAM Classroom
Your personal AI partner. Knows Trinity Graph. Evolves with you.
Provided
🧠
Claude AI
Your primary coding and analysis partner. Use it for everything.
Credits provided
⌨️
Cursor Pro
AI code editor. Write entire systems in natural language.
$20/mo optional
🎨
Figma + Midjourney
Design and visual concept generation. Your front-end foundation.
Free / $10/mo
Trinity Graph Viewer
Live visualization of the full Trinity Graph. 511+ term nodes.
Free

Rule: if you did something without AI assistance, you did it the hard way. Use AI for design, code, analysis, documentation, and presentation. Log your usage — it's 15% of your grade.

Your Personal AI

Your IAM Bot Has Been
Waiting for You

It knows the Trinity Graph. It knows your course context. It evolves with you for 14 sessions.

🌿
VOLVOX Lifecycle
Starts UNDIFFERENTIATED. Every 5 queries: SOMATIC SENSOR → STRUCTURAL → GERMLINE. Your questions drive evolution.
🎯
Kirk Steps 0–12
Your bot starts at Step 0. Engagement depth moves it forward. Step 12 = full collaborative partner.
Live Trinity Graph
Neo4j Aura behind your bot. 210 Trinity nodes seeded. Your queries activate real graph traversals.
Login
Username: your first name (e.g. "ava")
Password: firstname + 2026 (e.g. "ava2026")

URL: revilopark.github.io/ontological-theatre/iam-classroom.html

Session 1 · Right Now

Three Things Before
You Leave Today

1
30 min
Find Your Co-founders
This is the most important decision of the semester. Find 3–4 people with complementary skills. Talk to people you don't already know. Diverse backgrounds — design, business, technology if possible.
Register your team name + roles with Baxter before leaving.
2
20 min
Map Your Serendipity
With your team: each person shares one unexpected connection that led them to this room — a person, a book, a moment. Map it. This is your first Social Graph. The Serendipity Engine runs on these paths.
3
15 min
AI Tool Demo
Each person shows the team one AI tool they already use that no one else on the team has seen. This is your baseline survey. By Session 4 you'll have a full toolkit. Start here.
Session 1 Guide → IAM Classroom →
Session 1 Assignment · Due Wednesday Mar 11

Before Wednesday

Assignment · Due Mar 11
Co-founder Rationale & Team Roles
Submit your team composition with: (1) each member's role and core contribution, (2) your co-founder rationale — why these specific people, (3) your tentative venture domain. 1 page max. Email to Oliver + Baxter before Wednesday's class.
Required Reading · Session 1
The Social Organism — Chapter 1
Luckett & Casey. Read it before Wednesday. Notice the biological metaphors — they are not decorative. Come ready to answer: what is the "metabolism" of your startup idea?
HBR Required · Session 1
How to Identify the Perfect Co-founder
Harvard Business Review. The research is clear on what co-founder attributes predict success — and what predicts explosive conflict. Read it before you lock in your team. Yes, before Wednesday.
HBR Required · Session 1
Why Co-founder Relationships Fail
Harvard Business Review. ~65% of high-potential startups fail because of co-founder conflict. Not the market. Not the product. The people. This is the most important article in the case pack.

Session 2 (Wed): AI Tool Landscape + Sprint Methodology. Come ready to demo the AI tools you already use.

End of Session 1
"By April 22nd you will have
shipped something real
to real people."
AI-Accelerated Entrepreneurship Practicum · Vanderbilt Owen GSB · Spring 2026
Session 1 Guide → IAM Classroom →
Session 2 — Wed Mar 11: AI Tool Landscape & Sprint Methodology · Read: Social Organism Ch. 2 + HBR "Building the AI-Powered Organization"
🎙 Speaker Notes