R&D is supposed to produce something of value. If you agree that iterating rapidly using something like Tailwind can get you ~80% of the way towards some ideal design, do you think that last 20% is worth the upfront cost? I think there are very, very few cases where this will be true, and even for those cases, you can get first to market using the rapid iterated design and then refine it more towards the "ideal" before the upfront design approach even gets out the door.

Nue's approach to styling sounds nice in theory, but it seems like it's only a good fit for a domain you already know well, where the structure of the solution is already well understood and so the upfront design cost is actually minimal. For instance, the example project is a blog, a thing that's been around for like 20 years and whose structure, components and features are already well understood. I just don't think that's very common, but if I'm wrong I would certainly like to someone tackle a project that they don't understand using this approach.