[msandford]

Because for a long time they did precisely that: get stuff into beta really quickly and see how it does. As organizations get bigger there tends to be pressure to stop tinkering with stuff and start Doing Things Right (tm).

I think Google used to have a corporate culture which encouraged this and from what this writer says (so it's just an anecdote, not data) they might be losing that culture. Doesn't mean Google will fail but having that kind of culture is an ongoing job and it's non-trivial. Seems like they might be letting bureaucracy creep into the organization.

[abrahamepton]

That's really not what I'm saying, actually. I think Google continues to be a fabulously innovative place. I loved my time there, and fully expect it to continue being revolutionary and transformative.

But the constraints of keeping a billion people's needs in the back of your mind are real, and they impose limitations on what you consider worthwhile. It just wouldn't make sense for Google to study Chicago Public School utilization data, for instance - it's waaaaay too small a focus - but I find it deeply fun and meaningful, and not having to make something bulletproof for 1/6 the planet to potentially use just means it gets out the door quicker.

[msandford]

"not having to make something bulletproof for 1/6 the planet to potentially use just means it gets out the door quicker" is exactly what I'm referring to. As Google's market share grows, so does the bureaucracy. It's one of the inevitable problems of scale and it takes a very strong conviction on the part of the company's management to actively avoid it. I don't think Google's management is actively resisting the forces to bulletproof things and as a result progress slows down.

But your point about working on fun projects that are too small to matter to the world at large is very well taken.