[jrockway]

Dogfooding is valuable. If there is a fundamental flaw in their printers, they're going to find out about it firsthand.

It is also nice to know that I can print replacement parts and that my replacement is pretty much identical to the original.

[tpmx]

Sure, but they've kept doing it for many years now. At some point it shifts into some kind of "religion". Or, judging from your response, it's a part of their brand?

[kbob]

Keep in mind that the printers are getting regular upgrade cycles. A 2019 i3 MK3S is nothing like a 2015 i3. Firmware upgrades are released every couple of months. So continuous testing is still valuable.

[jrockway]

I was going to cite this example. I think recent MK3Ss have different fan shrouds than the first batch of MK3Ss, because the community found a better design. All Prusa had to do to upgrade new customers was to send their print farm a different file; total cost $0. Meanwhile, customers with the original design could download a file and print it, getting a factory-quality part for (essentially) free.

Managing a printer farm sounds like a lot work, but they did do that work and can reap the benefit that prototyping and manufacturing are the exact same process. It's like saying "why does Google have all these servers when they could just buy them from AWS?" Well, they got good at building datacenters before you could just rent a datacenter from Amazon. If you're good at something, keep it up!