>> The challenge for us is it'd likely take an engineer a week or two to optimize our test suites. That's enough time for them to ship meaningful features.

That does seem like a possibly worthwhile choice though? A "stop chopping the tree to sharpen the axe" moment?