I don't think one's principles should be determined by what skeptics say or do. "Just build" a device sounds good in theory, but most startups are there to prove a thesis and get backers - not to commercialize at industrial scale. That is what happens in all venture industries. M&A is more important for this area than IPO.
I admire Mills for his intense work ethic and showing that you can publish mainstream science whilst doing commercial R&D. I am sure he never in a million years expected that even with that, few physicists would not look at his data seriously. Still, you don't do what's right because of the whims of others.
I also think publishing science helps bring other important backers on board - who may otherwise not show up - be it granting agencies, angel investors, and other scientists.
Changing the world isn't easy, and we should always respect those willing to engage in the "Impossible" - it literally goes against all of our cultural / survival programming.
Publishing is all well and good, but the process is sloooooooow. It's slow even when you don't have an entire community biased against your claims from the get-go. When you have all the major journals barring at the door anything smacking of "cold fusion," it could take an eternity to push something through. Is it a good use of SAFIRE's time to pursue that avenue in hopes of convincing some scientific minded angel investor, or advisor to a potential investor, to hop on board? Probably not.
I admire Mills' attention to detail too, but at Brilliant Light's current pace, they are unlikely to beat SAFIRE to commercialization--and Mills has a couple decades' head start. I argue that most investors don't care much about why a device works, just that it works. There is a relative handful of people who even understand Mills' model, and they are not likely to be the ones investing.