When clouds take over, all the weather patterns change.
That’s the clear subtext in the trends mapped out in this year’s version of the Internet Trends 2018 presentation from Kleiner Perkins Caufield Byers’s venture capitalist Mary Meeker, presented at Code Conference Wednesday.
She points to the adoption of cloud computing — notably, the launch of Amazon AWS S3 in 2006 — as one of computing’s recent Big Bangs, with the other being consumer mobile (exemplified by the iPhone’s launch in 2007).
Meeker quotes the Amazon S3 launch announcement: “Until now, a sophisticated calable data storage infrastructure has been beyond the reach of small developers.”
But the impact of cloud services goes far beyond just offering big tools and storage to small-scale programmers. She noted that one AWS service in 2006 has now become 140+ services this year, even while cloud computing costs declined and cloud service revenue (for Google and Microsoft, as well as Amazon) accelerated, up 58 percent year over year.
From her presentation:
It’s this revenue boost, optimized because of cloud infrastructure, that has allowed Google, Amazon and to some extent Microsoft create these new services, and to dominate in their respective fields. In other words, it’s because they’ve become major cloud services that they are now occupying a heavenly position.
The acceleration, however, is not just about costs and revenue. Meeker points out that data gathering, sharing, optimization and “sensor pervasiveness” since 2006 are ramping up at a “torrid pace.”
Such torridity doesn’t just result in a massive increase in information. It creates entirely new services, like visual navigation through Google Maps, shared transportation in Mobike, predictive maintenance by Samsara or fitness tracking through Motiv.
And it leads to models based on large aggregates of data carved out for individual use cases, creating data-driven personalization that results in higher engagement. Often, personalized data is married with collective data to create new merged experiences, such as preferential newsfeeds, content discovery services, Netflix, Waze and NextDoor community recommendations. Here’s another slide from her presentation:
More data, Meeker notes, leads to better analytics, which generates better products, bought by more customers, which leads to more data…
And then came AI
That “data flywheel”, she noted, supports new artificial intelligence services that have become affordable for a wide range of developers. Amazon’s AI platform, one of her slides points out, has emerged from the AWS infrastructure and now offers — for hire — image recognition, language processing, machine learning and scalable compute clusters.
Similarly, an AI platform has emerged from the Google Cloud, offering a Cloud Vision API, Tensor Processing Units to accelerate neural net processing, a conversational platform and custom machine learning models.
AI, she quotes Google CEO Sundar Pichai, “is more profound than electricity or fire.”
That pervasive intelligence service is now beginning to impact new fronts, as the data and intelligence flywheels gain speed.
One new arena ripe for new services built on clouds — utilizing massive data collection and processing, personalization based on aggregated models, and AI — returns us to where this all started: physical stores.
Meeker points to a comment this month from Google Chief Economist Hal Varian: “We’ve seen how technology can make online shopping more efficient, with lower prices, more selection & increased convenience. We are about to see the same thing happen to offline shopping.”