Higher fissile number density = higher enrichment = nonstarter. Fine for submarines, not for civilian power. Re, reflector, you still have to deal with free path issues when determining overall reactor size. The more you're spending on inert mass relative to how much power you're getting, the worse your economics. Plus your reflector is contributing to (n, gamma) and other neutron consuming reactions (although it's possible to use a moderator that you need anyway (say graphite) as a reflector... although there are issues with that as well to deal with)
You'll note that I mentioned and agreed with the mass production argument - if fission power is going to have an actually sustainable renaissance, I would expect modular reactors to be the means. But I nonetheless questioned whether that could be enough to overcome the basic issues on top of the additional challenges that a small modular reactor imposes.