Oh boy. Who thought it was a good idea to run their own P&P in house?
Shaking my head. Even with the low volume turnkey I've hired out I know better than that.
If the best case scenario is just gaining parity with overseas assemblers, you gain nothing. You need to factor in the immense learning curve, wasted time futzing with the machine, etc. A buddy of mine babysits these things for a living. 6 months in and they are just barely getting a handle on things.
In-house PnP capability depends upon a lot of factors. Good for some, terrible idea for others.
e.g. Sparkfun, Adafruit, and 3D robotics are 3 examples of an ideal case for in-house PnP. They spin a lot of different products (hundreds per year), and small-ish runs (100's to 1000's). The logistics of getting a subcontractor to do hundreds of different boards per year is not easy, so makes sense to bring it in house.
Now I'm not entirely sure about Saleae, but from what I gather they don't spin that many new boards every year, they just churn out existing designs. In that case an in-house PnP isn't a great option because you are just trying to maintain an expensive and complicated machine at a labour price to compete with a contract manufacturer who usually knows how to do it better, likely pays their staff less, and has better access to capital to buy better gear (read faster + more capability) than you can afford on your own.