95% of technology startups are really just service organizations with the false belief that they need to invent knew technology to be successful. Sadly, most of these firms will high low level "engineers" to build essentially a website with application like features. Those engineers, working with a focus on either 1) recoding something they did previously but so horribly they were fired for it 2) select technologies and solutions which will improve their marketability at the expense or producing a usable site 3) solve scaling/performance technical issues their employer may never see, 4) ignore massive quantity of quality third party open source projects / solutions / toolkits / services because they only see the coding effort and wholly ignore operations/testing/code maintenance.
Although my education was in C on Unix, I find developing marketable, scalable, significant products on Microsoft .NET, when accounting for the cost of engineering, operations, licensing, maintenance (on shore and off) favors .NET when the organization / staff is primarily Microsoft centric. (Duh).
We used to say ... it's faster to go from 0 to 60 with Microsoft, but if you want to get to 100, you need to be on *nix. I still believe it's true. But that said, 95% of the shops out there won't ever need to go 60 mph ...
GM