The Myth of Neutral Technology

There is a comfortable assumption underneath most hiring: that a powerful technology, and a person credentialed in it, will work more or less the same wherever you put them. Plug in the model. Plug in the expert. Get the outcome.

That assumption has a long history of being wrong.

What an old book gets right

The idea that shaped this firm came from an unlikely artifact: Teknologi Kampungan: A Collection of Indigenous Indonesian Technologies, compiled by Craig Thorburn and published by the Appropriate Technology Project at Volunteers in Asia in 1982. It is a modest field manual. Its argument is not modest at all.

“Appropriate technology,” the introduction insists, is not a class of gadgets. It is a stance: that a technology’s value cannot be judged apart from the social, economic, regulatory, and cultural setting it enters. The same tool can be indispensable in one context and useless, even harmful, in another. Technology is not neutral. Its worth is decided by fit.

We are an American firm. We do not claim that tradition as our heritage, and our brand takes nothing from the communities the book documents. What we take is the discipline the author articulated, because it turned out to describe our work exactly.

The same logic, applied to leadership

If a technology is not context-free, the person leading it is not interchangeable either. This is the failure mode the market keeps pretending does not exist.

The numbers are not subtle. In the past year, 59% of organizations made a bad AI hire: a candidate fluent in the interview who could not apply it on the job. 46% of new executives fail within 18 months, and roughly 89% of those failures are attitudinal and contextual, not a deficit of raw skill. The all-in cost of a failed executive runs 10 to 15 times annual salary.

Look closely at why, and the pattern is always a fit problem dressed up as a skill problem:

• The model-builder hired to set enterprise AI strategy. Both are real skills. They are not the same skill.
• The brilliant operator from an unregulated startup, placed where every model decision now meets the FDA, the EMA, and an audit trail.
• The leader whose public fluency was mistaken for demonstrable depth.

None of these is a story about a weak candidate. Each is a story about a strong candidate evaluated as if capability were context-free.

Verification is the refusal of that myth

Vetting an AI leader is not a coding test and it is not a reference call. It is the discipline of asking the question appropriate-technology thinking demands of any powerful tool: not “is this impressive,” but “does this hold here, in this organization, under these constraints, in this person’s actual hands.”

That is the whole of what we do. We separate what a person genuinely owns from what they observed. We test judgment against the specific failure modes of the role. We weigh readiness for the regulatory and organizational reality the hire will actually inhabit. Then we hand you a judgment you can defend to a board.

The myth of neutral technology is comfortable because it makes hiring feel like procurement. It is not. The most consequential variable in your AI program is not the model. It is the person you trust to lead it, and whether anyone checked.