“We are entering an era when the compute
infrastructure, amount of data, and the algorithms are all
coalescing so we can get to knowledge and wisdom at scale.
I think that justifies saying that we’re entering
a new era—the data era.”
control and optimization. Then there are
public clouds, which give you aggregation at
an industry level.
Edge is the newest layer, and it offers a
place where you can push some of your processing
and analytics capabilities out to the
physical location of the people and devices
you interact with in real time. There are two
advantages of doing this.
bandwidth—it takes a lot of time, energy, and
money to move huge quantities of data back
and forth across the internet. The edge gives
you the ability to run the analytics locally so
that you don’t have an urgency to move all the
data into the cloud or into a private data center.
You might eventually move the data, but
you don’t have to move it in a priority manner
that costs a lot of money.
29
What are those advantages?
First, because the compute and analytics are
close to where the users are, the speed-oflight
issue that we have when we move data
over distance is no longer an issue. You can
operate and make decisions in real time. This is
incredibly important for things like autonomous
vehicles and their ability to react to their environment.
But there are several use cases—
in factories, healthcare, gaming—where the
goal is, first and foremost, to do things quickly
and not have the latency of crossing the internet
involved in the real-time service.
The other advantage to the edge is—as
much as we think the internet has infinite
There’s a lot of hype around AI, both
positive and negative. Where do you see
AI having the greatest impact?
I’m very bullish on machine intelligence. I
believe we can’t progress without sharing the
burden of thinking tasks with machines, and AI
is a set of tools that allows us to do that.
Largely today, we don’t use machines to
do thinking tasks. We store data. We process
data. But the actual empirical decision almost
always happens at a human level. As we look
at the world today, there are so many things
that, quite frankly, are incredibly inefficient
and would improve if machines were to take
over more of that responsibility.