Hacker News Reader

Comment by iainmerrick | Hacker News Reader

What always irritated me about Google X was exactly that rhetoric around "moonshots", as if Google had discovered this completely new technique for getting stuff done.

Not only have quite a few companies and organisations successfully completed huge, ambitious, forward-looking projects; but Google haven't, at least nothing remotely comparable to the actual moonshot they happily co-opt. Google has done lots of amazing stuff, but usually in a remorselessly incremental way.

Gmail and Google Maps (especially Street View) are the closest I can think of to "moonshots" in that they were big and bold improvements over what had come before. But neither one would be called a "moonshot" by Google X standards. In each case a back of the envelope estimate would be enough to show that it would work out (in technical terms, if not business terms).

So why does Google act like they're uniquely qualified to make moonshots work?

View on HN

GMail and Google Maps were also both ruthlessly incremental. They seemed like big improvements over what had come before because they had incubated inside the Googleplex for a couple years before they were released, and because the competition hadn't improved in nearly a decade.

It's more that Search was a non-incremental leap forwards, and that Google was basically the only company of the dot-com era that had attempted an ambitious technological goal and succeeded. Google Search really was fundamentally different from previous technologies; its main rival Alta Vista, for example, ran on a single box and used very standard text-ranking techniques, while Google thought very much out-of-the-box in terms of ranking, scale, and architecture.

Google X is very much a Larry & Sergey project, and Larry (while Eric was CEO) was often very critical of how incremental employee projects were and how they weren't taking enough risks on fundamentally different approaches.

Google search used a couple of clever innovations -- the PageRank algorithm, and scaling via lots of small computers. But apart from that, isn't it almost the definition of incremental? You don't need a massive investment to make it work, just a couple of grad students with 10 PCs to make a proof of concept, then 100 PCs to make it better, then 500...

Thinking about it, Google's key business trick has never been moonshots, it's been "throwing things against the wall to see what sticks". It's a wide approach (invest speculatively in lots of different ideas) rather than a deep one (bet heavily on a really long shot).

They practically operate their own business incubator.

Amazon is a much better example of a moonshot than Google search. The key idea is easy to state: an easy-to-use online store for almost anything, with fast and cheap delivery. But how on earth do you make that work in practice? You need a ton of business deals with various manufacturers and brands, you need warehouses everywhere, you need a massive delivery network... the investment required is just staggering. And indeed, for many years they were criticised for not making a profit, but the investment is clearly paying off.

At the beginning, Jeff have only thought of an online book company, instead becoming the 'Everything Store'. This legacy still exists as ASIN (Amazon Stock Indentification Number) is extended on ISBN to work on legacy systems.

"In each case a back of the envelope estimate would be enough to show that it would work out (in technical terms, if not business terms)."

I would think that applies to the Apollo project, too. An envelope can hold a quick calculation of how large your rocket needs to be, but it isn't large enough to hold any of the minor details that may cause problems. One also could argue it was done "in a remorselessly incremental way".

This bugs me especially when Silicon Valley companies get involved in life sciences, closer to my field. So much of the buzz is around "innovation" and daring to "think outside of the box," while so much of the actual work comes in implementing and tinkering with basic ideas. It's very weird to see a company say "We're not bound by the slow pace of the establishment, we're going to dream big, and look for a cure for cancer itself!" People have been looking for a cure for cancer for thousands of years--the problem in healthcare has rarely been too-small dreams.

>Not only have quite a few companies and organisations successfully completed huge, ambitious, forward-looking projects

Hmm, and what would those be?

The first example that came to mind (and made me add "organisations") is the creation of the National Health Service in the UK. A massive undertaking, especially as the country was still recovering from WWII, but massively successful.

Apollo, of course. The Manhattan project. Lots of NASA space probes.

Hmm, to be honest I'm less sure about companies... The phone network? Or laying railtrack right across the US?

Oh, and a topical one: the ARM chip! It was totally ridiculous for a tiny company like Acorn to jump straight from the 6502 to their own custom, cutting-edge microprocessor.

Okay, to me some of your examples seem high risk, but still incremental. There was plenty of prior research/innovation/development in those areas that they built upon. But to be fair, I also think that its almost impossible for someone to independently do something 'earth shattering' without building upon other people's work.

You're right, I suppose it's a continuum -- everything is incremental to some degree or other.

The Manhattan project is probably the best example, because there was nothing to build on except "science strongly suggests we could make a bomb out of this; we have no idea yet how to solve the practical issues, or even what all the practical issues are." It was partly incremental in that a much smaller nuclear bomb would still have been useful, but there was definitely a big risk that it just wouldn't work at all.

Reply on news.ycombinator.com