A specific category of AI expert commentary deserves more scrutiny than it typically receives, precisely because it's easy to source and appears frequently: commentators whose primary public visibility comes from commenting on AI generally, without a specific, ongoing role building, deploying, or researching it directly.
Why this category is easy to over-rely on
Professional commentators are, by design, more available, more quotable, and more consistently on-message than practitioners with a demanding primary job, which makes them a convenient and frequent source for time-pressured coverage. Convenience isn't the same as expertise specific to the question being asked.
What gets lost by over-relying on this category
A commentator whose expertise is in commenting on AI broadly, rather than in a specific technical or operational domain, is often speaking from general pattern-matching rather than direct, current, hands-on knowledge of the specific question at hand - a distinction that gets lost when a quote reads as equally authoritative regardless of its actual grounding.
What we try to do instead
We prioritize sources with current, direct, first-hand exposure to the specific question a given piece is asking, even when that means a source who's harder to reach or less naturally quotable than a professional commentator readily available on short notice.
Why this distinction matters for readers
Two quotes can read with identical confidence and fluency while representing very different levels of actual grounding in direct experience. We think readers are better served by sourcing that prioritizes direct grounding over polished availability, even when the latter is easier to acquire on deadline.