AWS re:Invent post 3 of 3
A guiding principle for Amazon is its focus on “stuff that won’t change over time:” for example, it’s extremely unlikely that people will ever want stuff at higher prices or slower delivery.
As an entrepreneur one should equally focus on the ‘basics’ of the respective problem domain: ask yourself, “what are the key user requirements, that won’t change in 3-5 years?”
If you want to innovate:
AWS is experiencing great growth rate in enterprise, gov’t and education
Thanks to the growth of online commerce and information sharing, consumers now (rightly) feel ‘entitled’ to near-perfect information (eg comparison shopping).
Hence, the power is shifting from marketing to product development: businesses can no longer rely on ‘less-informed’ customers being bamboozled by clever advertising and marketing ploys – also, in the age of Facebook, and social media in general, word spreads quickly: this cuts both ways.
This also applies to AWS where greater transparency around resources consumption can be accessed by developers so they can optimize their code, but just as equally exposes incompetence and laziness.
At Amazon, he enables everyone on the team to ‘pull the cord’ – fix defects nearer the source: this is a reference to the Kaizen manufacturing principles pioneered by Toyota, where, on the assembly line, anyone is allowed (in fact, encouraged) to pull a cord, stopping the assembly line and effectively stopping production, if they see a defect that needs fixing.
This is counter-intuitive, but incredibly effective, as it fixing defects ‘far from the source’ (in production) is way more expensive (and may well be nigh to impossible) than during development and testing.
Operating a low-margin business is hard: there’s nowhere to hide; high margins cover a lot of sins and waste (and I guess everyone of us in the audience could hear the subtitle: Oracle, HP, IBM wouldn’t stand a chance, were they not able to hoodwind customers into overspending for wasteful services).
Good entrepreneurs eliminate risks, gradually, until they have a viable business: so you need a systematic way of identifying risks in your product development and distribution chain.
A question he hears often is “Can AWS go after Enterprises and startups at the same time?
Apparently the needs are very similar around low-cost, high availability of computing and storage resources.
Other areas, like certification and security, developed mostly to meet the needs of large organizations, will benefit start-ups too, at a fraction of the cost, will make them ready for the day they will no longer be “start-ups” any more, and remove the need to migrate to some more expensive, high-margin, wasteful service.
10,000 year clock
Bezos is one of the investors behind this project of having a clock that will work for 10,000 years – he made a joke about it (“yes, I know, you’re all thinking ‘and he seemed sane until now'”), but he undertook it as a symbol for long-term planning.
When the focus is on such a long timescale horizon, we can accomplish things that would not be otherwise feasible.
He sees it as a means to change our way of thinking to foresee and avoid dangerous activities.
Another of his projects, this is the design and production of a vertical take-off reusable space vehicle: the goal is to make space travel (or, at least, sub-orbital trips) affordable; the only way to do so is to have a re-usable vehicle.
They are currently at the 3rd development iteration, and on course for sub-orbital trips.
Of course, they leveraged AWS for lots of simulations, requiring massive computing power.