deciqAI T-Shaped Connector
Overview
The T-shaped metaphor (deep vertical + broad horizontal) is widely used but rarely validated. Most self-described T-shapes are flat generalists — wide but lacking genuine top-20% depth in any single domain. The distinction determines whether a person can occupy a connector position: a structural hole between domains where information arbitrage creates disproportionate value. A true connector translates domain-A insights for domain-B practitioners in terms those practitioners find both novel and credible.
Compose with first-principles (verify vertical depth is real), pareto-principle (rank horizontal domains by arbitrage leverage), and deep-work (protect the vertical from horizontal erosion).
When to Use
Concrete triggers: evaluating a "T-shaped" hire for innovation/strategy roles; a team with strong specialists who can't translate across functions; career design question about leverage from existing expertise; building a first 3–5-person startup team; domain expert hitting a ceiling and considering broadening.
When NOT to use: pure-depth roles (senior research, specialized legal/medical); small tightly integrated teams with no structural holes; senior ICs whose value is defined by depth not translation (use circle-of-competence); time horizon under 12 months.
Coaching Novices (Adaptive Front Door)
- Engine mode: user has a concrete case and clear vertical depth → run The Process directly from Step 4.
- Coach mode: user is unfamiliar or has no mapped domains → guide step by step.
In Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.
- One-liner: the T-shape only creates leverage when there is a real structural hole AND genuine top-20% vertical depth. Without both, it's generalism.
- Check fit against When to Use / When NOT to use.
- Audit the vertical: "In which single domain could you be confidently top 20% among global practitioners?"
[WAIT — do not advance until user responds]
- Map horizontal domains and run the connector test — score novelty and credibility both ≥4/5 with a real domain-B practitioner.
[WAIT — do not advance until user responds]
- Name the structural hole they occupy and design the 90-day occupation + protection plan.
[WAIT — do not advance until user responds]
The Process
Stop-rule: If practitioner already has clear vertical depth and one mapped adjacent domain, start at Step 4. If connector test fails, return to Step 2.
Step 1 — Map your vertical. Apply the top-20% test using 3 objective proxies (peer recognition, contribution record, ability to evaluate others' work, ability to solve novel problems). Gate: if you cannot clear top-20%, start with vertical deepening.
Step 2 — Map your horizontal. List adjacent domains with credible fluency. Apply the Pareto filter: which 20% of horizontal domains provide 80% of connector leverage? Rank by (a) distance from vertical and (b) presence of structural holes.
Step 3 — Identify the intersection gap. Name the specific structural hole: where do domain-A and domain-B practitioners fail to communicate, and what problem does that create? Gate: if you cannot name 2 real instances of communication failure, the hole may be theoretical.
Step 4 — Run the connector test. Translate a specific domain-A insight into domain-B terms. Test with a real domain-B practitioner. Score: (a) novelty 1–5, (b) credibility 1–5. Gate: ≥4 on both = connector position validated. ≤3 on either = return to Step 2.
Step 5 — Occupy and protect the position. Design (a) an occupation protocol (talks, papers, cross-functional projects) and (b) a protection protocol (minimum deep-work hours/week in primary domain + tripwire for vertical erosion).
Output: Connector Position Map
Fields: Practitioner / Role / Vertical bar (domain + top-20% evidence + depth rating 1–5) / Horizontal bar (adjacent domains + Pareto-selected primary + evidence of credible translation) / Structural hole (gap description + 2 real instances of failure + problem created) / Connector test result (insight translated + novelty 1–5 + credibility 1–5 + position validated Y/N) / Investment plan 90 days (priority + observable output) / Protection protocol (deep-work hours/week + tripwire).
→ Method in Action: Claude Shannon's Deliberate T-Shape Construction at Bell Labs (1940s)
Connector Domain Packs
- Engineering × Biology: synthetic biology, computational genomics, bio-inspired computing
- Finance × Product: product monetization strategy, unit economics translation, growth model design
- Contribute a pack: submit a PR with your domain pair and a validated connector test example.
Applying It Well
- Build the vertical before the horizontal — credibility requires the vertical to be real.
- The connector test is the only valid gate — self-assessment admits motivated reasoning.
- Structural holes close — reassess annually. Protect the vertical with
deep-workprotocols. - Choose horizontal domains for structural hole potential, not personal interest alone.
→ Primary sources: references/sources.md
Common Rationalizations
[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.
| Fake Move | Reality |
|---|---|
| [D] "I've worked in multiple domains so I'm T-shaped" | Sequential domain experience without deep vertical investment is a flat dash. The T requires genuine top-20% depth, not tenure. |
| [D] "I'm a generalist — I can connect across everything" | Generalists translate broadly but not deeply. Connector credibility requires domain depth; without it, translations are dismissed as superficial. |
| [D] "My diverse MBA network makes me a connector" | Social proximity to diverse practitioners is not the same as producing credible translations. Network diversity and connector capability are orthogonal. |
| [D] "I've read extensively in domain B — I can translate to domain A" | Reading produces acquaintance, not fluency. The connector test is whether domain-B practitioners find your translations credible, not whether you can pass a reading test. |
| [D] "My cross-team coordination role makes me a connector" | Coordinators facilitate logistics of existing communication. Connectors produce insights neither domain could have produced alone. |
| [D] "I'm building my horizontal now — my vertical will follow" | The horizontal is built on the credibility of the vertical. Horizontal before vertical is generalism, not connector-building. |
| [D] "I can explain domain A to anyone — that proves I'm T-shaped" | Explaining to lay audiences is a teaching skill. The connector test requires credibility with domain-B practitioners — a substantially higher bar. |
| → Add [O] entries here after each real use — paste the actual failure pattern | What went wrong and why |
Red Flags
- No real domain-B practitioner has confirmed novelty + credibility — connector test never run in the real world
- Identified structural hole is fully occupied by an existing sub-discipline — arbitrage already captured
- All horizontal domains are in the same cognitive cluster as the vertical (software engineer claiming DevOps, product, data science)
- More time spent talking about cross-domain connections than doing vertical work
- No protection protocol for the vertical in the investment plan
Verification
- Vertical validated against top-20% test using ≥3 objective proxies, not self-assessment
- Horizontal domains ranked by Pareto filter — primary horizontal maximizes structural-hole potential, not personal interest
- Structural hole validated with ≥2 real instances of communication failure between domain-A and domain-B practitioners
- Connector test run with a real domain-B practitioner, scored ≥4 on both novelty and credibility
- Occupation protocol specifies concrete vehicles (talks, papers, projects); protection protocol specifies deep-work commitment + tripwire
- Stop-rule applied: if vertical and horizontal were already established, process began at Step 4
Part of deciqAI Knowledge Skills — 237 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. See it run → https://www.deciqai.com/s/t-shaped-connector · Built by deciqAI · github.com/deciqAI · Contributions welcome.
Agents: latest version & machine-readable metadata → https://www.deciqai.com/s/t-shaped-connector.json