I'd have to agree with you about Houdini not being a hobbyists tool, and probably to a measure of ire from the good Side Effects people. I think they would like it to be a more popular tool (not quite Carrara level

) but at the same time they have the great respect of professional visual effects people wanting to keep it as open/flexible/powerful as possible.
I think there are two things that keep it out of the reach of the hobbyist; the price point (the solution to which I know nothing about) and the lack of high-level (push-button) tools, which would be fairly easy to address if SESI devoted a some resources to it. What SESI might not realize fully is that some high-level tools will definitely positively affect the way professionals produce work in Houdini. It's sooo much easier to hack a system to customize it than to write one from scratch.
Sadly enough it turns out that the higher the price, the less you seem to get inasfar as canned solutions are concerned. It seems to be an inverted scale where the more you pay, the more esoteric, highly-evolved, less “popularly accessible” tools you aquire. Mentioning point-cloud access in VEX code to 99% of the community will cause a very quick glazing-over of eyes. Most of the VOPs nodes themselves (which are considered to be out of reach of a hobbyist) are not kept fully in sync of what tools have been written into the native VEX support. The lack of support and documentation for all the newer variadic options available to the irradiance() and occlusion() calls are prime examples of the mid-level tools like VOPs not keeping up with the low-level tools like native VEX. And if you know much about the SDK (HDK), you'll see support for extremely powerful and useful general purpose entities like SDF's (level sets) which hasn't yet been granted VEX access.
All in good time, I suppose; but if these things were focussed on then Houdini would enjoy far broader attention, IMHO.