I'm lucky to have lived through several eras of the ongoing revolution in computing.
I think you could argue that revolution is a bigger force shaping our world now than
it ever has been. Each era of the computing revolution has been defined by boundaries
and constraints. We overcome the constraints and push past the boundaries and we
then forget the boundaries existed as we accelerate breakthroughs on the path to ubiquity.
The great author and futurist Arthur C. Clarke's Profiles of the Future contains a quote
that I think embodies the spirit of exploration and discovery perfectly: "The only way of
finding the limits of the possible is by going beyond them into the impossible".
There are a number of big-picture developments that are going to change how we all
approach technology in the future. The end of Moore's law, the explosion of data,
the evolution of the edge among them.
We've been able to take for granted that there's a PC on every desk and in every home
for so long that we forget how impossible of a goal that must have seemed in the late
70s and early 80s. Bill Gates wrote his first commercial program on an Intel 8008
microprocessor. The most powerful GPU used for training AI models right now is billions
of times more powerful than this device. If you combined every 8008 ever made,
that would still be several orders of magnitude less computing power than you carry
around in your pocket in 2020.And that wasn't all. The scarcity of compute at the
beginning of the PC era was just one of many obstacles that had to be overcome before
we could have personal computing so ubiquitous that we all have the luxury of
taking it for granted. And here's the interesting thing: these impossible
-sounding constraints that Bill Gates and a bunch of other personal computing pioneers
confronted in the early days of that era/ seemed more like challenges and opportunities
to them /than obstacles that should deter them. None of those pioneers could have
accurately predicted what personal computing would become, but they had a pretty
good idea of the trends that would enable what they were attempting, as well as
the technical challenges that they had to overcome. They invested into those trends
and focused their ingenuity and creativity on overcoming those challenges.
So today I want to focus specifically on the trends and constraints that are influencing
one of the most significant developments in the history of computing:
the explosion of large-scale machine learning models and rapid
advancements in AI. Microsoft has been able to do some amazing things to drive
this trend forward, but none of the major shifts in our computing revolution have
happened just because of the efforts of one company. PCs became a phenomenon not
because of the PC itself but because of the breadth of things people did with them.
PCs are a platform for others to build on top of.
The internet changed the world not because of TCP/IP and HTTP but because it was
a platform for individuals and entrepreneurs and big businesses to create and innovate.
Even what you do every day on your mobile device isn't about the technology in the device
but more about the breadth of applications on those devices and the people, content, and
services those applications connect you to. At every one of these inflection points, the
leaps forward we made were because of the contributions of developers. Any platform is
only as good as the developers who use it, so for AI to achieve its full potential it must
be a platform, and this platform will need to be powered and delivered at truly
unprecedented scale and democratized so everyone can innovate and build on top of it.
That's where all of you come in. The future of AI will be in the hands of the developers
who can harness its powers. And you won't need to be a data scientist or machine learning
specialist to do so. For any developer, the wildest dreams of what you can achieve and
create with technology will become accessible as we build AI at scale. The reason that
I wake up every morning excited about my job is not because we get to work on these
things, but because we get to work on them in service of others. That's why I feel so
fortunate to be speaking today with all of you, the explorers who are going to help
us expand the boundaries of what is possible.