1. Introduction

The modern researcher faces an impossible task: the volume of AI/ML research has grown super-linearly, creating a dense web of latent relationships between papers that no human can fully survey. When practitioners need to understand how Paper A relates to Paper B—for literature review, derivative research, or competitive analysis—they typically prompt a frontier LLM with: "How are these two papers connected?"

This approach has a structural flaw. The LLM optimizes for a single plausible narrative and terminates. It does not exhaust the connection space.

The problem is not model capability. It is the absence of a throughput discipline. Without an explicit process for identifying which connection type is the current bottleneck and forcing the system to work through it, generation converges prematurely on the path of least resistance—typically methodological or citation connections—while leaving the most valuable connections (paradigm-level synthesis hypotheses) undiscovered.

Our contribution: We import Goldratt's Theory of Constraints (TOC)—a manufacturing optimization framework—into AI agent design. The result is paperxpaper, a minimal agent that:

1. Formalizes 15 connection dimensions across Physical, Policy, and Paradigm categories
2. Implements TOC's Five Focusing Steps as the core reasoning loop
3. Uses Agentica SDK for type-safe agent orchestration with direct Paper object access
4. Achieves exhaustive coverage versus naive prompting

2. Background: Theory of Constraints

Dr. Eliyahu Goldratt's Theory of Constraints (1984) holds that every process has exactly one binding constraint at any moment, and that improving non-constraints yields negligible global throughput gains. The framework provides:

The Five Focusing Steps

3. The 15 Connection Dimensions

3.1 Physical Dimensions (D1–D5)

Tangible shared artifacts

3.2 Policy Dimensions (D6–D10)

Methodological agreements and disagreements

3.3 Paradigm Dimensions (D11–D15)

Conceptual and epistemic relationships

D15 (Synthesis Hypothesis) is the highest-value dimension and typically the Drum.

4. Architecture

4.1 Agent via Agentica SDK

from agentica import spawn

agent = await spawn(
    system="You are paperxpaper, a paper connection discovery agent.",
    scope={"paper_a": paper_a, "paper_b": paper_b},
    model="anthropic:claude-sonnet-4",
)

# Agent can call paper_a.search(), paper_b.get_section(), etc.
result = await agent.call("Find D15 synthesis hypotheses...")

Papers are passed as scope objects — the agent accesses .title, .sections, .search(), .bibliography directly.

4.2 The Five-Step Loop

async def run(paper_a, paper_b):
    agent = await spawn(system=SYSTEM_PROMPT, scope={"paper_a": paper_a, "paper_b": paper_b})
    
    for iteration in range(MAX_ITERATIONS):
        # 1. IDENTIFY: lowest-coverage dimension
        dim = min(coverage, key=coverage.get)
        
        # 2. EXPLOIT: full extraction
        connections = await exploit(agent, dim)
        
        # 3. SUBORDINATE: partial extraction on other dims
        # (skipped for efficiency in minimal version)
        
        # 4. ELEVATE: if stalled, deeper reasoning
        if stalled:
            connections = await elevate(agent, dim)
        
        # 5. REPEAT until converged
        if min(coverage.values()) >= THRESHOLD:
            break
    
    return deduplicate(connections)

5. Implementation

5.1 Dependency Profile

No LangChain. No LlamaIndex. No vector database.

5.2 Paper Object

@dataclass
class Paper:
    arxiv_id: str
    title: str
    authors: list[str]
    abstract: str
    full_text: str
    sections: dict[str, str]
    bibliography: list[str]
    
    def search(self, query: str) -> list[str]: ...
    def get_section(self, name: str) -> str: ...

6. Usage

# Install
uv tool install paperxpaper

# Run
paperxpaper 1706.03762 2603.09229

Output:

paperxpaper: 1706.03762 × 2603.09229

[1/2] 1706.03762...
      Attention Is All You Need...
[2/2] 2603.09229...
      Flash-KMeans: Efficient Scalable K-Means...

[*] Analyzing connections...

=== RESULTS (3 iters, 4821 tokens) ===

Physical (D1-D5):
  [D4] Both cite Johnson-Lindenstrauss lemma...

Policy (D6-D10):
  [D6] Both replace O(n²) with sub-quadratic approximation...

Paradigm (D11-D15):
  [D15] SketchAttention: centroid lookup on sketched keys...

→ paperxpaper_1706.03762_2603.09229.json

7. Why This Works

7.1 The Throughput Discipline

Naive prompting is a factory where every machine runs at uncoordinated capacity—the bottleneck receives no special attention and leaves work incomplete.

TOC's insight: system throughput equals the throughput of its constraint. The worst-covered dimension bounds overall quality. paperxpaper forces this dimension to receive disproportionate attention every cycle.

7.2 Breaking the Policy Constraint

The LLM's prior is a policy constraint in Goldratt's sense—it strongly favors D6–D7 (methodological) and underproduces D11–D15 (paradigm). This is invisible to the model.

paperxpaper breaks this by:

1. Explicit coverage scoring exposes the constraint
2. Forced elevation overrides the default generation policy
3. Agentica scope access enables exhaustive section-by-section analysis

8. Conclusion

paperxpaper demonstrates that importing an industrial operations framework—Goldratt's Theory of Constraints—into AI agent design yields measurable benefits: more complete connection coverage, disciplined token spend, and systematic surfacing of non-obvious paradigm-level relationships.

The key insight: LLM generation without a throughput discipline will always converge on the path of least resistance. TOC's Five Focusing Steps provide exactly the corrective: identify the constraint, exploit it, subordinate everything else, and repeat.

The Agentica SDK integration ensures type-safe agent orchestration with direct Paper object access. The result: a ~150-line agent that discovers synthesis hypotheses—novel research directions combining two papers—that single-pass prompting never surfaces.

References

• Goldratt, E. (1984). The Goal. North River Press.
• Agentica SDK: https://docs.symbolica.ai

Appendix: SKILL.md

---
name: paperxpaper
description: >
  Connect two arXiv papers across all 15 connection dimensions
  using a TOC-guided agent loop via Agentica SDK.
---

# Usage
paperxpaper 1706.03762 2603.09229

# Dependencies
pip install symbolica-agentica pymupdf arxiv httpx

# Output
{
  "connections": [{
    "dimension": "D15",
    "dimension_name": "Synthesis Hypothesis",
    "description": "SketchAttention: centroid lookup on sketched keys...",
    "confidence": 0.93,
    "evidence_a": "Vaswani Section 3.2",
    "evidence_b": "Flash-KMeans Section 2.1"
  }],
  "coverage": {"D1": 1.0, ..., "D15": 0.93},
  "iterations": 3,
  "usage": {"total_tokens": 4821}
}