Any question answered...

printing 3d plastic shapes, how twee.

How about a laser sintered stainless steel rocket engine?

Or what about the idea of 3d-printing houses to incorporate termite colony style air conditioning into the walls?

Or maybe the (so far very speculative) idea of printing continuous flow chemical processing plants? (just think of the drug market / terrorist implications)

Even twee plastic shapes have their uses. I heard one story that rather than trying to keep stock of all the breakable plastic bits of their equipment, the US army print some of their replacement components.

There's a prof at Bath uni who has made a 3d printer for £400. Has open sourced it too so that everyone can hack and improve it. reprap it's called.

They will be everywhere in a few years.

Certainly they will be everywhere that that there are geeks who think they are cool. They'll probably put some useful pressure on commercial manufacturers to improve, but how long it will take before they can compete economically in anything but the most contrived situations remains to be seen.

I was initially quite sceptical of 3d printing as a general purpose way of making things. I still think people are underestimating the awesomeness* of a lot of the manufacturing techniques they are hoping to replace, but i think my initial guess that it was a stupid approach (entropy) doesn't stack up, and i'm becoming more optimistic that 3d printing will develop its own awesomeness that will stand up to (but be different from) traditional manufacturing.

Mostly i'm hoping the open source 3d printing culture is a beach-head for rolling back the magicisation of the manufactured world.

• all the banging and scraping in traditional manufacturing does important things to material micro-structure. Whether aligning the grain boundaries of metals in a drop-forged component or the chains chains of polymers and their crystallinity in an inflated PET plastic bottle, or polishing something smooth to optical precision, traditional manufacturing can work simultaneously on the macro-structure and micro-structure, while 3d printing works only at the scale of its print head's resolution.
  
  

3d printing's major points are:
Access to the insides allows construction of more or less arbitrary shapes, from a small pallette of input materials, so supporting a simple CAD model -> 3d object process.
Build costs are a function of object volume and precision, not complexity.
Any more?