I agree; why not do the work to show that it scales better? It violates scientific integrity to have a hypothesis, run the experiment, find that your hypothesis

I agree; why not do the work to show that it scales better? It violates scientific integrity to have a hypothesis, run the experiment, find that your hypothesis was wrong but then say it is probably right regardless