Anyway, I don't get what the big deal is about duplication of effort -- if it makes people happier to reinvent the wheel than to copy someone else's wheel -- let 'em; it doesn't hurt you.
New distros and package formats hurt everyone:
1. The "Linux" community only has so many knowledgeable volunteers and developers at any one time. Maintaining a general-purpose distribution takes a whole fleet of people, each of whom understands the intricacies of one or more subsystems. When you create another distro, you are implicitly hoping that you can get a whole bunch of people to stop contributing to some other project and instead contribute to yours; and/or you are hoping to divert new volunteers from other projects. New distributions spread us more thinly.
If you can make a newer distro that is significantly better than anything else, you might be able to kill off an older distro and/or grow the community enough to compensate for the above. But if you create a distro that is only marginally better than its predecessors, you will needlessly consume a section of the volunteer base. It would be better to take your ideas to an existing distro and improve that, instead.
2. If a programmer wants to write and test software for "Linux", the number of different distributions to target and test on keeps getting higher.
3. COTS vendors who are tempted to support "Linux" sometimes look at the mess of distributions and give up. When they do provide support for some Linux distros, their Linux customers sometimes whine that they aren't supporting $distro_of_choice, which makes COTS vendors hate us. Unnecessary distros make this worse.
4. New packaging formats, in particular, create additional burdens on cross-distro tools for package management and file browsing.