Strategy Science · 2026 · Featured publication
Mean Articulation Machines
Where do LLMs provide genuine strategic leverage—and where does their association-based architecture remain fundamentally mismatched to the task?
Plain-language summary
A continuum of tasks, not a verdict on “AI.”
Large language models are powerful because they can detect and redeploy linguistic patterns across enormous bodies of documented knowledge. The same architecture helps explain their limitations.
The paper decomposes strategic decision-making into increasingly difficult cognitive activities. LLMs provide genuine leverage in search, aggregation, paraphrase, comparison, professional drafting, and synthesis of established knowledge. Their reliability declines as tasks require causal explanation, sustained deduction, long-range planning, counter-consensus judgment, undocumented knowledge, or genuinely novel scientific and strategic breakthroughs.
The practical conclusion is neither to dismiss LLMs nor surrender judgment to them. Organizations should match the tool to the task and preserve human responsibility where strategy depends on rare, contrarian, causal, and tacit insight.
The capability continuum
From strong fit to human-critical judgment
Strong LLM fit
- Search and aggregation
- Consensus synthesis
- Paraphrase and translation
- Professional drafting
- Known framework assessment
Human-critical
- Causal theory construction
- Counter-consensus reasoning
- Novel strategic breakthroughs
- Tacit contextual judgment
- Responsible commitment
Formal abstract
Abstract
The performance of LLMs, both good and bad, derives from their core architecture as text pattern detection and generation machines that are sensitive to the frequency of the data upon which they are trained. They are amazing “mean articulation machines” in this sense. Using conceptual analysis and recent benchmark data, the paper identifies those strategic tasks that fall within the reliable competence of LLMs and those which remain fundamentally misaligned with LLMs’ associationistic architecture.
The result is a practical continuum identifying where LLMs offer genuine leverage and where human cognition remains indispensable. The most challenging tasks—novel scientific and strategic breakthroughs—are currently out of reach for LLMs due to inherent limitations in their architecture. Because breakthroughs are described with text does not imply that we can simply mine text for the next novel breakthrough.
In clarifying the boundary of current LLM capabilities, the paper aims to help strategic decision makers deploy these tools more effectively as powerful assistants for the majority of tasks that lie on the tractable side of the continuum.
Implications
Use the system relative to the task.
The further a decision moves toward novelty, causality, tacit knowledge, and contrarian commitment, the more it requires human theorizing and judgment.
Decompose strategy
Do not treat strategic decision-making as a single benchmarkable skill.
Exploit documented knowledge
Use LLMs aggressively for synthesis, drafting, comparison, and research support.
Protect the frontier
Do not confuse high-frequency articulation with rare insight or causal explanation.
Retain responsibility
People remain accountable for judgment, commitment, and action under uncertainty.
Read and cite
Mean Articulation Machines
McBride, R. (2026). Mean Articulation Machines. Strategy Science, 11(1), 31–54. https://doi.org/10.1287/stsc.2025.0439