[bliving]

Specialization would be my first guess. But then the problem is, how do you identify a niche that stays under Google's radar long enough to exploit any advantage? So, you just have to find a knowledge domain to specialize in that Google doesn't cover. And how hard is that?!?

You're right... how do you compete against this?

[mr_luc]

Great Teams Can Do Anything versus Great Surfers Need Great Waves.

I always felt as though the discussion was missing something -- and I think that this is it. Competition is different from most startups' vision of success. Growing to compete with the titans of your industry is always harder, so maybe you do need to ride a wave to do it.

I'll be the methods are the same, at least at the start:

    make something people want

That's what facebook did, and they've now got some pretty cool infrastructure running in-house.

I wouldn't say -- particularly after reading that article, yikes -- that anyone is really in a position to compete directly with google via frontal assault.

But if you really want to take on Google instead of be bought by them, the starting point is probably the same as it would be for an upstart in manufacturing back in the day: make something people want, a killer product; become the leader in that product, that thing, and you'll probably be more competitive than when you started, even if you don't have the economies of scale of your competition.

Google does move really fast, and their moat is massive.

I personally have neither the desire nor the ability to pole-vault that moat. Still, the trend against them is the iron-clad rule that small teams are always more productive than large teams (they manage it incredibly well, still -- somewhere, somehow, this is a drag on them), small organizations can move faster than big ones, etc.

Small organizations will be able to make things people want, the first step towards competitiveness. And as for infrastructure advantage the trend is in their favor:

    cloud-on-demand; or, DIY big storage (backblaze guys)
    crawling-on-demand (8legs et al)
    pbx-on-demand (twilio etc)
    ... lots more ...

That ecosystem empowers tiny companies to do a lot more before they have to think about in-housing difficult infrastructure -- possibly, enough to fund it!

( This is all just a thought exercise; I don't know why someone would want to wade the moat and take down google as opposed to sell to them. There are more pleasant things to do with your life. )