Motivated reasoning
Mark Seemann · article · 2026-05-04 source ↗ Reader 🖥️
The author's recent criticisms of AI in software development are likely shaped by motivated reasoning — love of craft and economic self-interest — yet even if ninety percent of his concerns prove irrelevant, the remaining ones may still matter, and raising counterpoints to techno-optimism has value regardless.
theoretical · reflective essay — A metacognitive essay that applies the concept of motivated reasoning to the author's own recent series of AI-critical articles. More confessional than argumentative — the primary move is epistemic self-audit rather than persuasion.
Structure
| Part | Role |
|---|---|
| Definition of motivated reasoning | Establishes the cognitive-bias framework the essay will turn on itself |
| Application to AI writings | Confesses that his recent AI-critical series is likely influenced by this bias |
| Love of the craft | Names the first motivational root — intrinsic love of programming-as-puzzle-solving, which AI threatens to eliminate |
| Incentives / economics | Names and partially defuses the second motivational root — economic self-interest — by noting he is near retirement and not primarily motivated by money |
| Conclusion: counterpoints still valuable | Argues that even biased concerns may contain one or two real problems, and dissenting voices serve an epistemic function |
Key terms
- motivated reasoning
The tendency to arrive at desired conclusions through reasoning processes heavily influenced by personal motivations, often subconsciously; smarter people are more susceptible because they can construct more convincing rationalizations - love of the craft Intrinsic Motivation & Making
Intrinsic attraction to programming as puzzle-solving that made decades of otherwise uninteresting business domains engaging - counterpoints to techno-optimism AI Backlash
Dissenting perspectives that question AI's net benefits, serving an epistemic balancing function even if most specific concerns prove wrong
Propositions
- definitionalMotivated reasoning is a cognitive bias in which personal motivations unconsciously steer reasoning toward desired conclusions.
- empiricalMore intelligent people are more susceptible to motivated reasoning because their superior mental resources allow them to construct more convincing rationalizations.
- self-assessmentThe author's recent AI-critical articles are likely influenced by motivated reasoning.
- self-assessmentThe primary non-economic motivation biasing the author is love of programming as craft — solving puzzles, not managing entities that solve puzzles.
- evaluativeBeing told to become a 'manager of agentic LLMs' is functionally equivalent to becoming a manager, which the author explicitly rejected decades ago.
- probabilisticEven if ninety percent of the author's AI concerns are irrelevant, the remaining ten percent could contain real problems.
- normativePosting dissenting opinions has epistemic value as counterpoints to techno-optimism, even when some concerns are wrong.
- evaluativeThe last few decades of the software development job market — where enjoyable work was also well-paid — are most likely abnormal.
Arguments
The author's AI criticisms should not be dismissed simply because they may be motivated by personal interest strength: medium
The prospect of becoming an 'LLM manager' is not an adequate replacement for programming as craft strength: high
Narratives
→
Questions addressed
- ⭕ Is the author's criticism of AI in software development just motivated reasoning?
- ⭕ What motivates the author's AI skepticism?
- ⭕ Should the author stop writing AI-critical articles given the acknowledged bias?
Gaps
- analyticalThe essay never identifies which specific AI concerns the author believes are most likely to survive the motivated-reasoning filter. The '10% that might be real' is left entirely unspecified.
- evidentialThe claim that smarter people are more susceptible to motivated reasoning is stated without citation or nuance. The underlying research (Kahan et al.) shows this primarily for identity-protective cognition on politically charged topics, not necessarily for professional self-interest.
- omissionThe essay treats 'love of craft' and 'economic incentive' as the only two motivational roots, but does not consider identity/status — being a recognized expert in a field that AI might devalue — as a distinct and potentially stronger bias source.