There is something fascinating about 3D printing, but I think some of that fascination and resulting hyperbole is misplaced.
The fascination seems to stem from the conjunction of a couple of ideas. One is that 3D printing is a quick and easy additive process. The second is the possibility that a machine could "print" the piece parts for its own assembly therefore suggesting the idea of self-replication, something that was previously only associated with biological systems.
But let's look a little closer.
Would 3D printing seem quite so marvelous if we simply called it 3D additive manufacturing? Also, in the general case a machine that could print itself could only make smaller copies of itself -- just as a mother makes a baby. I am excluding special topological cases, such as a donut-shaped machine that could extrude a sausage-shaped daughter-machine designed to then curl up to form a full-sized repeat of the mother-machine.
But in the conventional and general case we can see that self-replicating machines are not really enabled by 3D printing. For sure, a 3D printer could make the piece parts for a copy of itself, but it would still require other machines -- or humans -- to assemble the piece parts together to make the full-sized replica.
And even this neglects one important consideration: that the material that is feedstock for 3D printing generally has to be different from the material that is used to manipulate that feedstock, that is, the material from which the 3D printer is made. So, typically 3D printers made from metal (for strength) and plastic make parts made of plastic. A final point that rarely seems to be addressed directly is that the convenience of 3D printing usually comes with a tradeoff in terms of the strength and durability of the material.
3D printing can produce fantastic models and 3D visualizations and even functioning boxes, cogs, levers, and piece parts. But for most industrial applications expensive full-metal tooling and specialized high-strength materials fabricated using tried-and-trusted methods are still required.
3D printing is a useful additional technique for creating models, widgets, lightweight machines, and prototypes, and it makes these complex shapes affordable in low numbers in the academic and hobbyist environments. But engineering remains a complex discipline, and that means that 3D printing it is not about to completely transform industries based on an understanding of materials science built up over hundreds of years.