This is a subject which has been discussed recently on the G3 forum and we are probably a mini-representation of the views I've read here – with one or two members making very extensive (and impressive) use of CNC to produce their models and others seeing it akin to learning Black Magic.
I've been a late convert to 3D Print but certainly don't see it as a 'do-all'. I see people printing things that would be more easily & quickly made in other materials for instance. I've looked at the gantry 'router' type machines but a G3 friend has advised they are simply "not rigid enough" – which I think also raises the question of "for what"
To my mind CNC is particularly useful to a Hobbyist in two areas. The first is to machine parts which have complex geometries. My current approach to this problem would often involve fabrication. The second is the area of 'mass production'
As someone has already commented – most hobbyists don't make multiple parts – although they do in fact often use them, in the form of commercially produced parts and fixtures. Using purchased components is usually the most sensible method to pursue, assuming that the parts you require are actually available and are affordable in the quantity you require. Has anyone tried to purchase 16BA nuts or bolts recently?
There are also some operations where it's all too easy to make mistakes if your mind wanders for a second (my mind does this frequently). Drilling multiple holes in small parts has been already mentioned but there are other things where it's very easy to make mistakes when doing repetitive actions .
I don't really need CNC for largish 'flat' parts in steel – laser cutting is a viable option for this and even smallish flat parts can be assembled (and brazed) to produce useful fabrications.
So within my specialist area (Gauge 3 Railway Modelling) I see several uses for (inexpensive) automation.
1/ A 'Micro' CNC machine – a kind of automated mini capstan lathe & mill – not removing huge amounts of material but capable of producing multiple small parts. It wouldn't need to be ultra quick (like its commercial counterparts) – simply fast enough to do a few dozen parts over a day or so. Not waiting for a commercial solution but hoping for a public domain one to pop up.
2/ An automated drilling table (probably already available in the form of the 3020/6040 type routers) – the issue being mainly cost for something that might only have limited uses.
So I think that 'capable of machining most things' CNC may be overkill in terms of price and technology for my uses – but smaller, cheaper, probably more specialist machines might be a better solution (for me). I certainly do use my 3D Printer but for things (I consider) best suited for it – e.g. not everything.
Just my thinking on CNC currently – but as things evolve so quickly these days, who knows what we will have available in another year or two.
Regards,
IanT