I think the exactly opposite, if someone was able to build a framework that doesn't overly constrain the problem, and doesn't require weeks of screwing around with the build, integration of half baked components and insane amounts of boilerplate, that would be a fantastic contribution in and of itself even it didn't advance the state of tensor compilation in any other way.