Essay · 2026-10-06

Ask the field first

Research that starts from a technology goes looking for a problem. It usually finds one that fits the technology.

Mynd Labs Research starts the other way. Each field we work in has one plain guiding question, such as how a clinician can keep their judgement while a system does the paperwork, or how an automated financial decision can carry its own audit trail.

Those questions are our current best guesses. The first job of each program is to take its question to people who live with the problem and find out whether it is the right one.

What we learn goes into notes and open questions, including the questions that turn out to be wrong. We would rather publish a corrected question than defend an uncorrected one.

The AI Lab builds shared instruments for all of this: replay, memory, evaluation, safety testing and efficient systems. Every field shapes what those instruments need to do.

Related project: Fieldwork