As a PI, I have quite a few projects that have gone nowhere. For some of these projects, I have submitted a manuscript that has been rejected. In others, I have written portions of a manuscript, but feel like the results are rather inconclusive. I could submit these manuscripts as they are, but I find it difficult to write papers with inconclusive results (there's no story to tell). It feels unethical to resubmit papers where I have not addressed the flaws pointed out by the reviewers. I suppose I can just be as upfront as possible about the flaws, but this also takes some time and finesse. It seems like the amount of time I'd have to put into these manuscripts to address the flaws would be better spent on projects that are working and have promising results.
My grad school advisor seemed to be of the mindset that you should just keep revising and resubmitting a manuscript until it gets accepted. I sometimes felt like he was succumbing to a sunk cost fallacy and that the time we put into rescuing these projects wasn't worth it. There's also the argument that you should publish any work you do, even dead ends, because it will avoid duplication and save other people time. But putting anything in a publishable form takes time and effort, which takes away from other projects. What is the opinion of the community on these issues?
Another complication can be the involvement of others in the project (students, technicians), who would like to have a better publication record. However, in many cases, it might be faster to make them part of a better, more promising project than to dump time into a project that isn't working.
I'm trying to be more choosy about the projects I take on in the future to avoid this dilemma, but it's hard to know where a project will go. I'd like some advice on how to handle such projects/unpublished manuscripts. Just submit them to a preprint server and forget about them? Submit them to journals that have a low enough bar? How do you decide whether work is worth salvaging?