I would not say that fusors are a waste, but so far they have turned out to not be very efficient. There are several factors working against them, compared to linear accelerator-based devices. Basically what it boils down to is the neutron production cross section.
In a fusor, the ions are in a plasma, and the mean free path is huge. So the effective cross section is small (in other words, ions have a relatively small probability of hitting each other and fusing), and you are spending a lot of energy accelerating ions (and electrons) that fuse very rarely. Perhaps efficiency can be improved, but the results speak for themselves -- the best efficiency I could find was 10^4 neutrons/W [2]. Despite a lot of work over the years, we do not have a good qualitative understanding of what goes on inside a fusor.
Linear neutron generators have a huge advantage because they use solid tritiated metal hydride targets [1]. The ratio of hydride atoms to metal atoms is typically ~2, which presents a very large effective cross section compared to a plasma. In addition, the generated beam(s) of deuterons can be optimized for the target geometry. Furthermore, we have a better understanding of what goes on: it amounts to a fixed-target setup, which has a long history in nuclear physics.
There is the side issue of D-D vs D-T. Since the fusor is fed by gases and typically purges to atmosphere, hobbyists (and most scientists) operating fusors are restricted to the use of deuterium due to safety concerns. For a given energy in the typical range, the microscopic cross section for D-D is lower by an order of magnitude or more. But when the efficiency deficit is 10^4, 10^5, there is still a big gap even if D-T was used in both cases.
[1] See chapter 7 of PhD thesis by J.M. Verbeke, UC Berkeley, 2000
[2] Miley group, UIUC, 2004
