Since hiring for engineers is so competitive, it is arguable that choosing fancy tech can give an advantage there. It has been done successfully (in a hiring sense)... Twitter pushing Scala, Jet.com on F# are two I can think of.
The downside is it could backfire on who you're able to hire. It could just lure people who will either want to migrate to the next new-and-shiny or just leave quickly.
I think one thing a lot of startups think is that they absolutely have to be able to attract the best and brightest. In 90% of cases, average level programmers can build your product just fine. Unless you are building for enormous scale or your product is using cutting edge technology, you simply don't need a bunch of programmers from the top 1% or even top 10%.
In my company, we hire middle of the pack programmers for a bargain and use commonly known tech stacks.
> you simply don't need a bunch of programmers from the top 1% or even top 10%
Having done some recruiting at a startup last year, even getting the top 50% to reply to an email is a tough proposition. The problem really stems from brand awareness. No one has heard of your startup, so in a sea of "help wanted" postings on Angelist, Stack Overflow or Craigslist or cold emails, it's extremely easy to be ignored.
Using novel technology and getting the word out (as Twitter and Jet.com did), can at least get technical awareness within a motivated subset of engineers who are interested in said new stuff -- and of all skill levels.
That said, Peter Norvig made a pretty good refutation of your core point[1]. Say your initial site was successfully launched by average engineers. If you're not always trying to hire better than the average employee at your company, your team just gets worse over time.
I'm sure he's right when you are building a one percenter company. But my team will likely never grow to a size that it will ever take on a life of its own outside my control. 5 engineers tops when we hit market saturation. Each new hire I make is more skilled than the last because I can pay more as time goes on.
I just don't think research done from the perspective of the biggest tech company in the world has any relevance on 99.9% of small tech startups.
Doesn't this say more about the flaws of the business and/or the management than it indicates this being a good idea? Many of the most successful software companies do everything they can to attract and retain the upper end of software engineers and they don't do this because it's bad for businese. Presumably they do this because it provides more value for the company than having more programmers or trying to save money on salaries/perks.
Perhaps, but I actually think getting the best and brightest is more of a factor of having tons of VC money. In that case, you need to move as fast as possible, and every little bit helps. If you are building a niche business with limited TAM, you just don't need the top tier, assuming of course you aren't building something groundbreaking from a tech perspective.
Interesting perspective. Kind of reminds me of the discussion yesterday about programming contests not being the best indicator of software engineering success. Personally I love programming contests, but reading your comment I can see how it's not the most important thing to be in the top 10% in Algorithms.
Your main focus should always be to ship good user tested product, and I agree you don't have to be the best programmer to build something that users want and like.
The downside is it could backfire on who you're able to hire. It could just lure people who will either want to migrate to the next new-and-shiny or just leave quickly.