[roymurdock]

I like to scan these articles for numbers before reading. If I don't see any, I exit out.

If you're going to talk about financial/economic phenomena please at least conduct a modicum of quantitative research and give us some concrete data to discuss, not just opinions in a void.

[diminish]

There's maybe one small number observation, I m not sure if it's correct as there maybe high growth YC companies which will soon pass those 3/4.

"In 2011, Y Combinator’s poster-child alumni were — already — AirBNB, Dropbox, and Stripe. Can you think of any Y Combinator companies from the last five years as well-positioned today as those Big Three were then? Maybe Instacart, if their unit economics work. That’s it."

[nostrademons]

That observation has a huge amount of hindsight bias. AirBnB was founded in 2006, DropBox was founded in 2007, and Stripe was founded in 2010. What did the ideal startup look like in 2007? It was one of:

"A social network for X, like Facebook"

"A social news site for X, like Digg"

"A way to massively improve your e-mail experience, like GMail"

In other words, people always chase the massive startup that got popular about 3 years previously. This article is pretty much on-time if it's holding up AirBNB, DropBox, and Stripe - those are the massive startups that got popular about 3-4 years ago.

But in 2007, these were most decidedly not the hot industries to go into. Hotels were a done deal: Hilton, Marriott, Holiday Inn, and others chains owned it, and who would think a tiny team could take them on? Filesharing was an incredibly crowded market with 20-30 players, and in any case, if it got popular Google was going to crush them with Google Drive. PayPal owned payments; everybody knew it was a regulated industry with strong network effects, so why bother to compete?

[monkmartinez]

You don't need hard numbers to philosophize about the tech industry, nor economic phenomena in general. Read blogs like ibankcoin/flyblog or zerohedge and you will find financial/economic articles written with data that is ancillary. That is, the hard data is not immediately required to make the point.

[Periodic]

It could have had better supporting arguments and evidence, but I couldn't find enough of it and just skimmed to the end to make sure I wasn't missing it.