Another lesson that may apply: fad driven development isn't necessarily a good idea. Sometimes "old" (i.e. established, tested, widely used) technologies are better choices than whatever the latest fad offers.
If today I were putting together an MVP with an eye toward getting funding, I'd only use things that somebody on the team knew very well. Learning new stuff is great, but it means you will be unpleasantly surprised a lot more often. Those WTF delays really add up, and they often come at the worst time.
For example, I've worked in a number of environments that hired for node.js, but where the devs were actually recent RoR converts who were moderately proficient in node.js and its ecosystem. In the case of getting an MVP out quick, it might still be better to use RoR.