[thwarted]

"Our biggest mistake in our startup was scrapping a prototype that was built in a certain framework for first PHP and then eventually Python/Django. ... This applies not just to the programming language or framework, but to hardware and software applications — keep costs low, use what you already know, and move on"

This is one that requires a large grain of salt. Switching technology may very well be compatible with "use what you already know" and "keep costs low". You can't switch key players in the middle and expect consistency in either quality or cost, which a lot of people think is possible. "You can’t afford to have a religion" works both ways: you can't have a religion based on what you perceive to the value of what you've already done either. I've come into a few projects after a not insignificant money was spent on "cheap" freelancers and contractors to "get the prototype up". Now all I'm supposed to do is add a few more features and polish it to take it to the next level or deploy it. This is often an impossible task, or one that will take longer or cost more than a rewrite because it merely does the prototypical features and wasn't built for expansion or even deployment. But people fear the "rewrite" because it sounds like you're starting over. Get the right people in early, use #8 when necessary, and #6, #7 will come easier (at least in terms of the technology).