00:00:01
[Music]
00:00:04
[Applause]
00:00:08
[Music]
00:00:11
so I think I will add to the
00:00:12
introduction and something I've never
00:00:13
told you um which is when Sati and I
00:00:16
started talking about uh combining
00:00:18
Microsoft and Linkedin I knew there was
00:00:20
a bunch of different very good ways in
00:00:23
which the missions were friends the
00:00:24
cultures cared about how do you elevate
00:00:27
human potential and how do you make them
00:00:29
productive and I knew that that was
00:00:31
going to be part of the magic of
00:00:32
combining the companies what I didn't
00:00:35
fully know was that part of the magic
00:00:36
would be also working with you um
00:00:39
because not only is SAA a extremely
00:00:42
smart and accomplished skilled executive
00:00:45
but also genuinely looks at this kind of
00:00:47
large scale and what are the things that
00:00:51
uh the world is coming and how do you
00:00:52
take your responsibilities for that
00:00:54
world very seriously in shaping the
00:00:56
world you want to get so it's been a
00:00:58
honor and a pleasure being on this thank
00:01:00
you so much re and the next time the
00:01:01
board does my evaluation I'll ask you to
00:01:03
put that in yes I I would be delighted
00:01:06
to do that um so let's start from the
00:01:10
the from the very scope of the world
00:01:11
which is you know obviously we have um
00:01:14
Wars uh pandemics and markets om my we
00:01:17
have you know climate change world on
00:01:19
fire we'll be talking to Bill about that
00:01:21
at the you know closing uh of the last
00:01:24
session of the conference but one of the
00:01:26
things that you're doing and very much
00:01:28
focused on your mission is how do you uh
00:01:32
essentially uh use technology to help
00:01:34
solve these large problems what are the
00:01:36
what's the shape of the kind of the the
00:01:38
ways that people should think about you
00:01:40
know kind of technology and they solve
00:01:43
solving scale problems in the world
00:01:44
ahead in this world that has so much
00:01:47
challenge yeah no it's it's something
00:01:50
that I've been thinking a lot about in
00:01:53
fact uh you talk since you mentioned
00:01:56
Bill the month when that popular
00:02:00
Electronics magazine came out you know
00:02:03
which I guess Bill picked up along with
00:02:05
Paul Allen and started
00:02:08
Microsoft the same week there was a news
00:02:12
week cover which talked about the energy
00:02:16
crisis
00:02:18
inflation uh and so that juxtoposition
00:02:20
of those two magazine covers um sort of
00:02:24
I think are back with us uh in some
00:02:28
sense which is you have all the things
00:02:30
that you talked about uh as the
00:02:32
challenges or constraints and so so this
00:02:36
idea of doing more with less of course I
00:02:39
come at it with the bias of software as
00:02:42
the most malleable resource that we have
00:02:45
that we now got to use to tame uh all of
00:02:48
what's happening but the way I think
00:02:50
about let's talk about more what is what
00:02:52
is more I would say economic growth I
00:02:54
mean if I look at human history the last
00:02:57
250 years have been pretty unique right
00:03:00
it is the Scientific Revolution combined
00:03:02
with a whole bunch of other things that
00:03:04
happened uh which created Modern Life uh
00:03:07
We've enjoyed it except we do have real
00:03:10
challenges let's take energy right then
00:03:13
whatever economic growth that comes next
00:03:16
cannot break the planet uh we definitely
00:03:19
also need more social cohesion inside
00:03:22
the United States in the world so that
00:03:24
means we have to deal with
00:03:26
inequities um right so so to me thinking
00:03:30
about technology that drives economic
00:03:33
growth while we in some sense we need a
00:03:36
new constraint solver where you need
00:03:38
economic growth but don't break the
00:03:40
planet create more trust create more
00:03:42
Equity uh and that's what I think I I'm
00:03:46
I'm most excited about or at least
00:03:48
that's what Microsoft's mission is and
00:03:50
in fact the way I describe it is what if
00:03:52
we now maybe during the founding of
00:03:55
Microsoft you picked one magazine and
00:03:57
sort of created history this time around
00:03:59
pick both the covers uh and then go to
00:04:02
work on it and and speaking of creating
00:04:05
history we are in this huge shift in
00:04:08
artificial intelligence and one of the
00:04:10
things that I think that uh Microsoft is
00:04:13
is doing very well is thinking about how
00:04:15
to amplify human beings and one of the
00:04:17
products we already have in the market
00:04:19
is co-pilot yeah so say a little bit
00:04:21
about co-pilot and say how that isn't
00:04:22
just like why it's really interesting on
00:04:24
for engineering and developers and how
00:04:26
technology the creation of Technology
00:04:28
but also go to how this is a way that
00:04:33
all professionals all creatives should
00:04:35
be thinking about what kinds of
00:04:36
Transformations are coming down the pike
00:04:38
yeah I mean this if you step back for a
00:04:42
second uh I think uh Greg Brockman is
00:04:45
also coming later uh what's happening
00:04:47
with these large uh multimodal models um
00:04:54
is pretty phenomenal um and essentially
00:04:57
the curve on intelligence
00:04:59
basically by throwing more compute
00:05:01
you're creating more
00:05:03
intelligence um so let's take that um
00:05:06
you know if you don't you know Greg I'm
00:05:08
sure we'll talk a lot more about it but
00:05:09
let's take that as a
00:05:11
given the I love this idea or this
00:05:14
metaphor of a co- pilot uh because it's
00:05:18
it goes back to I think to the theme of
00:05:19
Masters of scale it's such an empowering
00:05:22
word even right so I'm the pilot and I
00:05:24
have a co-pilot um the GitHub co-pilot
00:05:27
it's essentially it's an AI pair
00:05:29
programmer right uh in fact it's it's
00:05:32
the thing that makes me feel just great
00:05:35
every weekend whenever I feel like I
00:05:37
bring up vs code you know it's kind of
00:05:39
believe me I mean I wish I could code
00:05:40
better uh but you know but nowadays with
00:05:43
thanks to get up co-pilot I can get into
00:05:45
anything and feel like a super person
00:05:48
like super programmer um the idea that
00:05:52
you can really tame the learning curve
00:05:56
on the most sort of intense knowledge
00:06:00
work right software programming is just
00:06:02
pretty phenomenal I mean one of the the
00:06:05
you can prompt for whatever it is that
00:06:08
you want to do um and it'll generate the
00:06:12
code for you yeah but it's just not that
00:06:16
right you can just highlight a bunch of
00:06:18
code this is stuff that we have in the
00:06:19
labs now which is you can highlight a
00:06:21
bunch of code it'll explain it for you
00:06:23
in English right which is sort of a
00:06:25
great way to learn it let's say you know
00:06:27
one of the things that's happening is
00:06:28
everybody's saying hey C++ to rust right
00:06:31
which is sort of how do I move to
00:06:33
perhaps a more security first language
00:06:36
from C++ and it'll translate it for you
00:06:39
yes so if you have something of that
00:06:43
nature and it's like inside of the data
00:06:45
we have it's around 30% 40% of the code
00:06:48
is being generated by the AI programmer
00:06:49
but you're still in the loop right the
00:06:51
way the design even works is you don't
00:06:54
feel the loss of control if anything you
00:06:56
feel more in control more productive
00:06:59
more EMP powered and I love that human
00:07:01
feeling and somebody sort of described
00:07:03
this to me which is when the first
00:07:04
computers came how people felt using
00:07:07
them is how I think the co-pilot
00:07:10
recaptures it and to your point it'll
00:07:13
change everything like we're working
00:07:14
with Autodesk on Maya uh so you can
00:07:17
imagine architecture industrial design
00:07:20
uh pretty much any field you can have a
00:07:22
co-pilot for everything and so that I
00:07:25
think is a great metaphor uh there will
00:07:28
be Automation and full automation
00:07:29
somebody gave me this nice sort of AI
00:07:32
metaphor of you are um in the loop on
00:07:37
the loop or off the loop uh it's sort of
00:07:40
a good way to think about the three
00:07:42
designs uh this one is you're in the
00:07:44
loop uh and that's a place to focus on
00:07:47
yeah I mean it's one of the things that
00:07:49
I think is gets so often lost in the
00:07:50
discussion of artificial intelligence is
00:07:52
that there because it's like well okay
00:07:54
you know look at Hollywood Video most of
00:07:56
it's like Terminator exmachina etc etc
00:07:59
we the revolution we're in is enabling
00:08:01
human beings right we will get to other
00:08:03
questions that will be important
00:08:04
questions around Equity important
00:08:05
questions are handling it but the
00:08:07
amplifying giving all human beings
00:08:09
superpowers like that was one of the
00:08:10
things that Dolly which are also doing
00:08:12
with designer right which is you know
00:08:14
how do you make everyone and all of
00:08:17
their skills more able to do amazing
00:08:19
things 100% And then then the one other
00:08:21
thing Reed you mentioned let's take this
00:08:25
entire energy transition I had not
00:08:28
understood this and there are many more
00:08:29
experts you're talking to bill later but
00:08:32
the the thing that I I didn't realize is
00:08:35
in order to move the world from one
00:08:38
Paradigm to the other we kind of have to
00:08:41
compress maybe 150 years of chemistry
00:08:44
into like 5 years 10 years 20 years
00:08:47
there's no way we can do that without a
00:08:49
co-pilot for it uh like in other words
00:08:52
just finding the new molecules that
00:08:54
completely change uh what is our
00:08:56
petrochemical base that's the type of
00:08:59
stuff that I think is doing more with
00:09:00
less to me yep and and what do you think
00:09:03
when when when people think about uh
00:09:06
what are the ways that they should kind
00:09:08
of approach their own skills and careers
00:09:11
like you also mentioned already with
00:09:12
co-pilot your own coding but when
00:09:14
they're thinking about these tools how
00:09:15
should they think about how do I adjust
00:09:18
how do I help how do I lead my
00:09:19
organization what's the ways to think
00:09:21
about this yeah
00:09:23
it's first of all one of the things that
00:09:25
are most when when we think about
00:09:27
digital capability build right I go back
00:09:31
you know when the PC was born um I I I
00:09:36
distinctly remember the first time I
00:09:38
used a spreadsheet uh I don't know how
00:09:40
you all felt but you know it's just well
00:09:42
I mean Excel is my most
00:09:47
favorite you know what I might as well
00:09:50
admit it right or uh the re but think
00:09:53
about it right when you first were able
00:09:56
to make sense or numbers right your
00:09:58
relationship with numbers changed yes um
00:10:02
and so knowledge work in particular in
00:10:04
the early ' 9s as it started spreading
00:10:06
uh you could even say late 80s but early
00:10:08
90s is when with sort of the graphical
00:10:11
interface completely changed um
00:10:14
interestingly enough the question is
00:10:16
what is that what is the moral equalent
00:10:18
of that um I do believe a domain expert
00:10:22
who has I'll call it Excel class skills
00:10:26
can they create automation can they
00:10:29
create workflows that's sort of I think
00:10:31
the real uh way to think about even
00:10:34
Skilling so one of the things is there
00:10:35
will always be elite skills no no you
00:10:37
know like the people doing these large
00:10:39
scale AI models is one side but the most
00:10:42
exciting thing to me is the Excel like
00:10:45
skills that are going to spread around
00:10:47
the world in every domain in fact one of
00:10:49
the things that I've seen even in the
00:10:51
pandemic where there wasn't enough time
00:10:53
to go to the digital team to have them
00:10:56
build stuff for you uh do M experts were
00:11:00
turning to robotic process automation or
00:11:03
these low code no code tools to create
00:11:06
applications and guess what they were
00:11:08
being assisted by AI in fact we have in
00:11:10
Power Platform the idea that you can
00:11:12
just take pictures of say some design
00:11:14
you drew on a piece of napkin and turn
00:11:17
it into an app right so there's AI that
00:11:20
does the forms recognition generates the
00:11:22
code and then you can probably add a
00:11:24
formula too that type of skill is is in
00:11:28
some sense takes the domain skill you
00:11:31
have and makes it a digital skill and
00:11:34
that's just fundamental amplification
00:11:36
yeah this blurring and crossover of the
00:11:39
world between atoms and bits is actually
00:11:41
one of the things I think is really
00:11:42
leading to the to the amplification so
00:11:45
let's let's let's uh add to this uh
00:11:49
metaverse and um start with digital
00:11:51
twinning right so like because part of
00:11:54
the whole thing about being able to
00:11:55
amplify the world we live in the world
00:11:57
we navigate is to use these tools within
00:12:00
simulation environments within uh and be
00:12:03
able to apply them in a much stronger
00:12:04
way so what's the what's what's the way
00:12:06
that you're thinking about this yeah I
00:12:09
mean the way at least I think about it
00:12:11
is all of um the software categories
00:12:15
we've always had are all about
00:12:18
essentially
00:12:19
creating digital twins uh of people
00:12:22
places and things right so there's you
00:12:24
know Bill would always say to me there's
00:12:26
only one software category it's called
00:12:27
information management and I never
00:12:29
understood what the heck does he mean by
00:12:30
that but it fundamentally is that's all
00:12:33
it is you schematize the world uh and
00:12:36
then that helps you re you know reason
00:12:38
about the world but the point about the
00:12:41
meta wore to me is we've done it by uh
00:12:45
reference right so you schematize
00:12:47
something you know about a person a
00:12:50
place a thing whereas for the first time
00:12:53
you now can do it by value right so you
00:12:56
can schematize the actual place
00:13:00
um or or digitize the actual place the
00:13:03
actual thing the actual person and then
00:13:05
you put the two things together right
00:13:07
embodied presence in a place without
00:13:10
being there that's a killer app right I
00:13:12
mean teams meetings are fantastic but
00:13:14
what if we can have this type of setting
00:13:17
where embodied presence uh can actually
00:13:20
happen that to me is the killer app uh
00:13:23
you know in a world where we move
00:13:25
towards the metowers but to your point
00:13:27
about digital twin
00:13:30
I must say this a even in the pandemic
00:13:34
one place where we saw a lot of
00:13:35
acceleration is every Warehouse every
00:13:38
Factory uh effectively got digitized um
00:13:41
and the benefits of digitization mean
00:13:44
that you can run in simulation uh that
00:13:47
is less less waste less energy more
00:13:50
accuracy the ability to sort of remotely
00:13:53
debug things so I think that to some
00:13:56
degree that thinking about this in as
00:13:59
just the next phase where we are
00:14:01
digitizing by value not just by
00:14:03
reference is probably a good way to sort
00:14:06
of think about the meters yep and how
00:14:07
does the the metaverse work which
00:14:10
obviously you know people it's very easy
00:14:12
for people to imagine things like snow
00:14:14
crash and entertainment and all the rest
00:14:16
but part of what I wanted to highlight
00:14:18
because this is a bunch of the work that
00:14:19
that Microsoft and you guys are doing is
00:14:20
it's not just that it's how you work
00:14:22
it's how you solve difficult problems
00:14:24
it's how you uh you know kind of uh game
00:14:27
out things
00:14:29
now how is that uh a core part of the
00:14:33
cloud Revolution so talk a little bit
00:14:35
about how metaverse and Cloud go
00:14:37
together yeah I mean the
00:14:41
the thing that I find um that cloud has
00:14:46
just done is for any category whether it
00:14:50
is AI or for metowers is created
00:14:54
abundance of compute that at that at the
00:14:57
foundational level I mean if you think
00:14:58
about it right you know you had the main
00:14:59
frames you had the minis you had the
00:15:01
client server revolutions they were all
00:15:03
about sort of really creating more
00:15:05
elasticity in some sense of compute but
00:15:08
the cloud is truly the first time where
00:15:11
you can uh just throw lots more computed
00:15:15
any problem you want the other thing is
00:15:18
especially at a time when Mo's law seems
00:15:20
to be running out of gas uh you now need
00:15:24
new system architectures uh like
00:15:27
basically in order to be able to train
00:15:28
the large scale AI models we had to sort
00:15:31
of completely build a new type of
00:15:32
supercomputer uh which is very different
00:15:35
than what you would do to run just say
00:15:37
regular Enterprise VMS or what have you
00:15:39
so that to me is what is really helpful
00:15:42
which is you can have different system
00:15:44
architectures like in order to think
00:15:46
about realtime streaming or and
00:15:50
visualization uh is a very different
00:15:52
type of workload but that requires
00:15:54
different type of computation and we can
00:15:56
build the cloud for that and then make
00:15:58
it elastic you can do the same thing for
00:15:59
training you can do the same thing for
00:16:01
inference so that to me is why I think
00:16:03
the cloud Revolution is still in the
00:16:05
very early stages but the way I I myself
00:16:08
having started in the Microsoft cloud
00:16:09
way back in say 2010 and to recognize it
00:16:13
now I just can't I mean it's just a very
00:16:15
different Beast yeah uh and so when we
00:16:17
say cloud what do we mean we mean
00:16:21
compute systems that are changing by the
00:16:24
nature of the workloads yep and and then
00:16:27
to to shift a little let's let's talk
00:16:29
about the future of collaboration we've
00:16:30
all come out of you know the pandemic
00:16:33
we've all kind of suddenly gotten shift
00:16:34
many of us gotten shifted in location um
00:16:37
we have this kind of question and this
00:16:39
may also be you in terms of how you're
00:16:41
leading Microsoft not just the tech but
00:16:43
it's like you know 70% of people say
00:16:45
according to Microsoft they they want
00:16:47
this flexibility but also 70% say they
00:16:49
want more human connection so how is it
00:16:52
both leading Microsoft and the tools
00:16:54
you're navigating this you know apparent
00:16:57
Paradox it's hard
00:17:01
uh yeah it's
00:17:03
it's it's let first you know because um
00:17:07
there are a lot of people uh who during
00:17:10
even all of this came to work right so
00:17:12
the first line workers you know whatever
00:17:15
60% of the workforce didn't have the
00:17:17
option of working remotely or working
00:17:19
from home so I think let's recognize
00:17:21
that but having said that I think the
00:17:25
the the knowledge workers uh this just
00:17:28
just a real structural shift um we did a
00:17:33
survey
00:17:34
recently uh or just looked at data
00:17:37
pretty broadly across sectors across
00:17:39
Goos um and there were three findings
00:17:43
that at least we are staring at
00:17:46
including ourselves at Microsoft the
00:17:48
first one is what we describe as this uh
00:17:51
productivity Paradox right I mean to 85%
00:17:54
of the managers think their employees
00:17:57
are slacking off and 85% of the
00:18:00
employees think that they're working too
00:18:02
hard and they're burnt up right so I
00:18:05
mean and it's real data um and so we
00:18:08
have this Paradox as to how can you sort
00:18:12
of see the same thing in two different
00:18:14
ways um and so the only way around it
00:18:18
read to me is you got to use data in
00:18:20
other words Dogma is not going to help
00:18:22
um and you now need to ground yourself
00:18:25
on how do you act for example have a l
00:18:29
goals if anything I I think as Leaders
00:18:31
we have to learn how to bring Clarity to
00:18:34
what is the output uh that you would
00:18:36
like you like to see what are the
00:18:38
measures of it and then using that to
00:18:41
see whether it's working or not and then
00:18:42
create the Norms the other piece that's
00:18:45
very interesting is I mean think about
00:18:47
it right in the introduction of the
00:18:49
summit you sort of talked about people
00:18:51
came here not because you sort of set a
00:18:54
policy and mandated it they came for
00:18:57
other people right people come for
00:18:59
people not for policy yes and u i I like
00:19:02
to say that all of us are now event
00:19:04
managers right it's not just the uh
00:19:07
believe me if I say I have a meeting
00:19:08
nobody's going to show up uh but if I
00:19:11
say oh I have an
00:19:14
event people are going to come yes at
00:19:17
least once you know and then and so you
00:19:21
really have to learn I think as a leader
00:19:23
new soft skills around how to create
00:19:26
occasions what does it mean uh to hold
00:19:29
an event um and then I would say the
00:19:32
last thing is I think every one of us
00:19:35
you know grew a lot during the pandemic
00:19:38
uh just because you had one onboarding
00:19:41
which was a remote onboarding means
00:19:43
nothing right you have to re-recruit
00:19:45
re-energize so there's a lot of tough
00:19:48
tough work but there are lots of tools
00:19:51
there's lots of Technologies but the
00:19:53
interesting scarce commodity I think is
00:19:56
a little bit of what is the new set of
00:19:58
in new set of soft skills and Leadership
00:20:01
skills we have to learn just because you
00:20:04
were a great leader in
00:20:05
2019 you can't go keep going back and
00:20:08
saying I just want the world to look
00:20:10
like 2019 it just I mean maybe it will
00:20:13
but I don't think so yeah wishes were
00:20:15
fishes all right so we are actually in
00:20:16
overtime so very quickly leaders in this
00:20:19
room leaders in a live feed run other
00:20:21
teams one message about how the world is
00:20:24
changing technology Workforce Etc what
00:20:27
would it be
00:20:29
I think the the thing the world needs
00:20:34
today more than anything else from
00:20:36
leaders is optimism and energy uh it's
00:20:42
easy to sort of be down on everything
00:20:45
well the way you started the
00:20:48
conversation uh that's my
00:20:52
role not that you are referencing the
00:20:54
world yes so what does a leader do I I
00:20:58
think my true measure of any leader is
00:21:01
who can come into a situation bring
00:21:04
Clarity generate energy and solve over
00:21:07
constrainted problems and as long as you
00:21:09
do those three things then I think we'll
00:21:11
be a better place SAA always
00:21:15
[Applause]
00:21:23
amazing what's said at the summit stays
00:21:27
at the summit everybody will have to pay
00:21:29
attention almost no one sees this coming
00:21:33
those Founders are becoming more
00:21:35
globally ambitious they're coming after
00:21:38
your talent your Capital your market
00:21:41
share the difficulty is when a team has
00:21:44
been working so hard on one
00:21:48
thing if you are trying to use your time
00:21:51
and your talent and your money in the
00:21:53
service of others the work will never
00:21:55
end we have to deal with inequities
00:21:59
it's different if you're a Founder but
00:22:00
when you walk into something so big and
00:22:02
so perfect and so successful you know
00:22:04
it's not about me it's about me becoming
00:22:07
a part of it and just evolving it and
00:22:08
making it better the thing that I love
00:22:10
about it and it makes me so optimistic
00:22:12
is that there's it's just creating more
00:22:14
dialogue I'm here to tell you that wow
00:22:16
human innovation's incredible this is
00:22:19
masters of scale Summit
