Eugene Yan's essay on choosing problems in data science and machine learning frames project selection as a set of tradeoffs rather than a search for the "best" problem. The first axis contrasts short-term gain with long-term gain: some problems deliver quick wins, while others build compounding value over time. The second axis contrasts incremental innovation with disruptive innovation, asking whether teams should optimize existing systems or attempt bigger leaps.
Yan also warns against resume-driven development, where practitioners pick problems mainly because they look impressive on a CV. Such choices can favor fashionable topics over problems that matter to the business or users. The essay suggests that awareness of these incentives is part of making sound decisions.
Because this is a single source, there is no independent viewpoint to compare. The value here is in the framing: it gives data scientists a vocabulary for discussing why they choose certain problems and what they are implicitly giving up.