SaylorCorpus

Michael Saylor Keynote

Saylor University · 2017-06-29 · 38m · View on YouTube →

0:00

[Jeffery Davidson] Okay thank you everybody, I'm going to get

0:02

started again. It's my pleasure to introduce the trustee of the Saylor Foundation, Michael

0:07

Saylor. For his day job, he actually runs a very large company called MicroStrategy.

0:12

He's the chairman and CEO, they're a global enterprise -- leader in enterprise analytics

0:16

and mobility software. They have over 2,000 employees worldwide. Since we're changing

0:22

up the program, a little short on time, Michael's full bio is in your program but you'll note

0:27

he's a graduate of MIT with two degrees, both earned on a full ROTC scholarship, and upon

0:32

graduation he was commissioned as a Second Lieutenant in the United States Air Force.

0:36

When I've heard him speak, Michael often cites his ROTC scholarship as inspiration for his

0:42

willingness and passion for paying it forward and help finding a way to help others get

0:48

a first-class education without having to take a lot of debt in order to do so. His

0:52

philanthropy extends to many areas but his commitment to education is extraordinary,

0:58

giving not just his resources but his time and his energy for new ideas as well. So with

1:02

that, please join me in welcoming Michael Saylor. Thank you.

1:06

[Michael Saylor] I want to thank everybody for being here today

1:13

and thanks for the time and commitment you've given to open education. As I know, everybody's

1:19

got their own passions but it's so exciting to see everybody coming together at this event.

1:26

I know to a certain extent I'll be preaching to the choir, but Jeff asked me to share a

1:33

few thoughts on my observations about technology trends in the marketplace and the views of

1:40

various employers in the tech business toward education and toward open education. And as

1:50

I have observed, things are changing very rapidly this year; in fact they're changing

1:55

rapidly every quarter. It seems like for the past three or four years, the pace of change

2:01

has accelerated and of course it's been accelerating for quite a while. Every day I get up and

2:06

I see a new and interesting thing that is just another brick in the architecture of

2:15

open education so I thought I'd share a few observations with you. For those of you who

2:20

know, I wrote a book called The Mobile Wave in 2012 and one of the themes of the book

2:25

was the dematerialization of products and services from the physical world into the

2:34

cyber world. So many things that used to be a product like a camera, like a printer, like

2:43

a typewriter, right, now there are actually software applications or icons on your iPhone

2:52

and oftentimes as they dematerialize they became not just a piece of software but they

2:58

became entire software networks. So something like the Rand McNally atlas which once was

3:04

a product that we printed eventually became Google Maps but then Google Maps went from

3:11

dematerialization of the atlas to 200,000 pages of satellite images which you never

3:17

could add to an atlas and then it went to real-time traffic and then it went to an intelligent

3:26

advisor telling you how to drive to work. And so what started as a physical product

3:31

in the real world became an intelligent service, which is software running on a network, with

3:38

hundreds of billions of lines of telemetry and we never go back. Now that, I mean, that's

3:44

a really interesting trend, I mean this dematerialization, and I mean the world continues to digitize

3:51

at a more rapid rate. Why is it important? Because in a world of physical things if I

3:56

have a wind tunnel and I want to teach someone how to build an airplane I need to produce

4:01

the wind tunnel and there's a physical limit to the number of people that can fit into

4:05

the wind tunnel and use the facility. As an undergraduate at MIT, we didn't get to use

4:10

the wind tunnel it was too expensive. Now, in the world where you create a cyber wind

4:16

tunnel, you can produce a million copies of the wind tunnel and give it away to people

4:21

in a hundred thousand places and as the cost of hardware gets cheaper and the cost of RAM

4:27

gets cheaper -- and RAM's 10,000 times cheaper that was 15 years ago, computers are a 1,000

4:33

to 10,000 times more powerful than they were 20 years ago -- the advent of things like

4:39

AWS have resulted in a world where we have software you can punch a button at 9:00 a.m.

4:45

and at 9:23 a.m. you can spin up a dedicated environment to support 87,000 users in Singapore

4:54

and you can run it for the next four hours and turn it off. Now, four years ago, that

5:00

would have cost you 150 full-time employees and 30 million dollars in capital. Today,

5:07

you could do it in 37 minutes and the significance is as we get to that world where you've got

5:15

infinite capacity on the iPad and infinite capacity on the laptop and massive capacity

5:23

in the backend servers -- as everything gets 10,000 times cheaper, better, faster, and

5:29

as we digitize, we're able to move from a world of fragile to a world of agile. And

5:38

in the fragile world, you wanted to try something, it cost you forty million dollars and four

5:42

years and if you screwed it up you're out four years and you're out forty million dollars.

5:46

In the agile world, you wanted to try it you tried it in four hours or in 40 hours, you

5:51

spent no capital, and if you screwed it up, you throw it away and you try something else.

5:56

And if I can try 150 things fast, then I don't have to be perfect, I just have to find one

6:04

out of 150 things that works. I discard the other 149 things, civilization advances. In

6:10

the world of the wind tunnel, you can't, right? Things that have physical physical capital

6:18

involved and bricks and mortar involved, they're inherently more fragile, it's going to be

6:22

much harder for us to do things. And one thing that's fragile inexpensive is a traditional

6:28

education at a bricks and mortar institution in a traditional fashion, it just takes too

6:34

long and costs too much money. This digitization, digital revolution, you know, we tend to -- we

6:44

were big on it in the year 2000, but there's a better story every single year that goes

6:50

by. I was sitting with a friend of mine the other day, a very successful DC businessman,

6:55

and he just took on a job running a drone company. And I said, "What do you do?" He

7:00

said, "Well, we've got these drones." I said, "Drones like the ones that fly around and

7:04

take like a photo?" And he said, "Well, no, our drones are flying LIDAR arrays." What

7:08

is LIDAR? Anybody know what LIDAR is? Anybody here? You're advanced, right? Light detection

7:15

and ranging. It's the technology that lets cars drive, self-drive, but what it means

7:20

is I put this array on the bottom of a drone and it flashes the room and it takes a perfect

7:27

image of everything. You put it on a drone above a beach, it'll tell you everything.

7:32

What is everything? It means you now have data that will tell you where the sharks are,

7:36

you can put it over agriculture, I can tell you where all the plants are, what's living,

7:40

what's dying, how fast it's growing, is it safe, is it not safe, count the number of

7:44

cats running through the field. Now, why is it interesting? Well because it's like a million

7:51

times more data than a photo. Like, we think, oh digital photo, that's a lot. Well this

7:56

LIDAR thing is this array and now we've got these drones that are a thousand dollars flying

8:02

around doing this array, that are collecting -- we're not talking about gigabytes or terabytes

8:08

of data anymore. I don't even -- you know exabytes, exabytes a minute of data, and it's

8:15

a business that didn't exist two years ago. And he goes yeah, well I just had someone,

8:21

an insurance company wanted to hire me to fly over every single roof -- whatever -- in

8:27

the country or something and figure out, you know, what's dangerous or not. Now, why is

8:33

that interesting? In every single field, there's something like that that's driving the digital

8:37

revolution and the world that was physical is digitizing and that means all of the skills

8:43

and all of the goods and services that were physical are on this path to digitizing. And

8:49

whereas it used to be I had to put you in a lab and hand you some tools to learn to

8:54

do something or to prove you had the capability to do something, now I can fly a drone over

9:00

a beach, I can hand the data to a programmer, and the programmer can write algorithms to

9:06

spot sharks. And you don't have to fly in a plane and you don't have to be on the beach

9:12

you just simply need to have a set of skills, right, and what are they? Math, science, engineering,

9:19

coding, a whole set of technical skills. And every single industry that rolls over to become

9:24

digitized creates a ton of jobs for people that actually have these kind of STEM skills

9:32

and techniques. Now, why is that interesting? Well, because in a world of digital the machine

9:39

is now software. Software is an information machine and there are more and more information

9:48

machines that we need to build. We don't have enough people to build them. You know there's

9:54

a tendency to think well we're going linearly, we used to do that we're doing this, but in

9:59

fact my observation in my industry is we're in an expanding universe. And it's expanding

10:05

-- it's an expanding universe of expectations, of demands, of requirements, of aspirations.

10:12

Everybody wants more of everything in every direction, right? It's like, Elon Musk wants

10:16

to go to Mars and now people don't laugh at that. They're like, they're wanting to do

10:20

it. We want to cure every cancer and people don't laugh. It used to be people wanted to

10:24

make you comfortable. Now people think you can cure the cancer, right? We want to get

10:28

rid of sharks on the beach. We want everything to run perfectly. There is no no tolerance

10:33

for imperfection. I say there's -- in my industry we used to, like, build software running on

10:38

top of three databases: Oracle, SQL server, Teradata, or something. Now they want to run

10:44

software on top of 48 different relational databases, 10 different OLAP databases, 50

10:50

different big data databases, 100 different enterprise applications simultaneously, and

10:55

also splice in Wikipedia, Google, you know, Facebook, Twitter, everything else. So you

11:01

go to the customer, you say, "Well which of these new things do you want to do?" They're

11:03

like, "Well, I just want to do 187 things simultaneously." Okay, we used to deploy it

11:09

to DOS and then it was Windows. Do you want it on Windows? No, it's the Web. Okay, so

11:15

you're going from Windows to the Web. No we want Windows and the Web. What's new? Well

11:18

now we want it Windows and the Web and iOS. So do you want to go from Windows to iOS?

11:22

No, I want iOS and Windows and Android, and I want it to run on all these clients, and

11:27

I want to run on smartphones. Smartphones instead of Macbooks? No, smartphones and Macbooks.

11:34

I want it to run in every client. And you want it to run in -- no, I want it to run

11:39

everywhere. Well, everywhere? Everywhere. Well, everywhere means I have to -- well yeah,

11:45

you have to comply with Chinese privacy policies in Beijing. But it's different in Singapore.

11:49

And that's different in Germany, which is different in Ireland, which is different in

11:53

Brazil, which is different in the US. Which is probably different in California of the

11:57

US. So everything is -- it's an expanding universe in every direction and we've got

12:07

a million problems. We could -- and every one of them, you know, how do you find sharks

12:14

from LIDAR, you know? That's a problem. We got a million problems and it's pretty clear

12:22

in the tech world [if] you want to solve the problem you need to have mastered all of the

12:28

undergraduate skills and then probably a bunch of master skills and the definition of a PhD

12:33

-- the classical definition -- is you know someone who is capable of contributing and

12:38

making a unique seminal contribution to the body of knowledge of the civilization. If

12:43

you're able to make a unique contribution then you probably deserve a doctorate. If

12:48

you're able to do algebra you probably don't deserve a doctorate. We have about 10 million

12:54

PhDs the last time I checked, five million in the US, 5 million everywhere else. There's

13:02

like seven or eight billion people on the planet. If you think about it a bit and you

13:07

think about what skills are required to, you know, put chips underneath the skin that,

13:13

you know, solve diabetes or cure cancer or do whatever, you come to the conclusion you probably

13:19

need a billion PhDs. There's probably -- we need -- we don't need seven billion but we

13:24

need more than ten million. You probably need like a billion people on the planet that are

13:28

able to make a unique contribution, because if you want to create an airplane that flies

13:34

15,000 miles or turn it into a rocket ship, you're not going to do it with algebra. You

13:41

know, you're not going to do it without having mastered postgraduate thermodynamics and coding

13:46

and probably plugged into someone's network. And we desperately need more sources of new

13:52

engines, new propulsion, new breakthroughs in medicine, new breakthroughs everywhere.

13:57

And that'll be done with a lot of very sophisticated, educated people. Human capital. We don't have

14:04

enough human capital, right? I mean the cost to create a PhD is a million dollars a year.

14:10

I mean, sorry, a million dollars total. It's just expensive, right? It's a quarter million

14:16

dollars to get through college, it's a quarter million dollars to get through K through 12,

14:19

it's a quarter million dollars for the rest, right? Maybe someone can figure out how to

14:24

do it for two hundred thousand dollars. But two hundred thousand to a million dollars

14:28

to take a person and convert them into a PhD, if you do it the traditional way, and so I

14:35

don't think we can do it the traditional way. I mean we -- the only way we're going to create

14:39

the kind of human capital we need to solve all of these problems and move the civilization

14:43

forward is if we provide 100X more education even if -- even if I said I got a hundred

14:52

billion dollars and I'll pay everybody's education everywhere on the planet, we still don't have

14:58

the capacity and the bricks and mortar institutions to train these people, right? So if I paid

15:03

everybody's Harvard education, Harvard still will cap you at whatever the number of students

15:07

is they want on the campus. So we need 100x more capacity and we need 100x cheaper price,

15:14

right? The cost per education unit delivered has got to go way down, the capacity's got

15:22

to go way up. To do that, you need to create a machine and in this case a machine to manufacture

15:28

education. And software that teaches is that machine and we know that you can create software

15:36

that will teach people. Now, we also know if I want to teach you ballet, right, if I'm

15:43

working in a skill like golf or ballet or cooking, right, you're in the physical world,

15:49

in the analog world, and all of a sudden you're dealing with fragility and capital intensity.

15:54

It's really hard to do that with a piece of software, right? So I don't think we're going

16:00

to solve the problem of how do you create a great, great cheap education in physical

16:08

world so easily; we will struggle with that. But on the other hand, if what I wanted to

16:14

do was generate someone that knows Calculus, in theory, there's no reason why you can't

16:18

learn Calculus in front of a computer just as well as learning Calculus in the classroom.

16:22

And as a practical matter -- I get a kick out of this -- on Saylor.org we have lectures

16:30

from MIT that were uploaded from the time that I was at MIT okay? But here's the joke,

16:38

right? When I was at MIT, my entire family savings for the last 200 years would have

16:43

been slurped up by the first six weeks of my education there as a student, right? It

16:49

was too expensive. And I sat in a room bigger than this where the lecturer was as far away

16:56

as you are my dear in the back, right, and I squinted to figure out what was going on

17:01

on the chalkboard and it was very uncomfortable. Today you would get a better education for

17:06

nothing with a $500 computer logging into saylor.org, and I know because I was there,

17:13

right, I lived it. So just because something is expensive it doesn't mean it's better.

17:20

The world is full of products that are manufactured, the iPhone for example, and there's nobody

17:25

that thinks that if they spent a hundred thousand dollars and built their own phone it would

17:29

be better than the one that they got from Apple. And if you spent a hundred million

17:33

dollars to build your own phone it still wouldn't be better [than] the one you got from Apple.

17:38

So productizing education, I think, is critical for us, and I am optimistic because it's a

17:48

lot easier to productize the teaching of algebra or Calculus or coding. You can objectively

17:54

determine that someone knows two plus two equals four. You don't need the opinion of

17:59

someone that smelled them in the classroom. They don't need to show up and sit in the

18:04

front row, on time. All they need to do is prove the two plus two equals four and that

18:11

hasn't changed for thousands and thousands of years. So one would think that you could

18:16

automate it and slip the copyright by now. I think open education plus computer power

18:25

plus cheap ubiquitous networks, right, that they lead to better, cheaper, more comprehensive

18:34

education, right, that's the magic the magic there. Now if you look at it from an employer

18:41

point of view, I think employers are becoming much more data-driven today, and I mean that's

18:47

a buzzword, so let me let me tell you what I mean by that. We've been in business 27

18:52

years, I've hired 20,000 people over the course of 27 years. In the year 2016 for the first

19:03

time we started to administer diagnostics and we used a platform called eSkill and HackerRank

19:12

and that that -- we use a product called SmartRecruiters that recruit the people. So here's how it

19:17

works. For every single person that goes to an interview with our company -- and this

19:21

means the last 5,000 people -- and we probably hired a thousand, so probably a thousand people

19:27

have been hired in this technique. Everybody goes through the company [and] we give them

19:30

an ABCDE diagnostic. A: analytics. It's like the math version the SAT; it takes about 20

19:37

minutes to 30 minutes and we figure out what your symbolic reasoning is. B: business. We

19:43

actually -- we actually give people a business exam to figure out if they have common-sense

19:49

reflexes about how to do business in the enterprise, like you have a meeting how early should you

19:54

show up for the meeting -- five minutes before, 15 minutes before, or..etc. A lot of a standard

20:01

business test is just questions to see if someone has common sense. C: coding, can they

20:08

code? And we used HackerRank to see if they could code, it's a quick evaluation. D: design.

20:14

We actually give people about 25 different systems, you say here's some information,

20:20

here are five different ways to present the information, pick the best way. And you know

20:25

what? Like, I can administer that in half an hour and I can tell -- in one second I

20:31

can tell whether someone has the aptitude, the talent, to be able to design and create

20:39

application interfaces. And I -- and that's more important to me than knowing that they

20:44

have a degree in fine arts from Harvard, right? In fact, the last one of those is English.

20:52

E. But we also have an F for French or G for German and M is Mandarin, right. We check

20:57

some other languages. But the general idea is, let's just figure out do they have these

21:04

basic aptitudes and it's like the SAT, but you know what, it used to be people would

21:10

say, well, the SAT, that's no longer relevant and you've been out of work -- in the workforce

21:15

for 15 years and so we won't look at that. We'll just do a bunch of interviews and we'll

21:19

look at your resume. And we have a lot of things we take into account when we hire someone.

21:25

We do look at their resume, we do interview them. But we actually found a better strategy

21:30

is we put the diagnostics up front. We screen, and then we interview, then we look at the

21:36

resume. I could tell you -- you could have a master's degree in computer science from

21:41

Harvard or from MIT. You get a 50 on the HackerRank and then I have to compare you to someone

21:48

from a university in China I never met before in my life, who's 8,000 miles away I will

21:53

never meet. And in one second -- they're 98, I will hire them. And so it's a very interesting

21:59

thing. We're lurching from a subjective, traditional set of credentials to objective, universal

22:06

set of credentials. And by the way, these aren't perfect, but I'll tell you what's interesting

22:11

here, is we applied them and we found that the leading indicator of failure of all executives

22:18

in the company is a low analytic score. Like -- like if they if they rated below 50 -- and

22:25

it's zero to a hundred by the way -- and if you, if you come out of school and you're

22:30

a bright whippersnapper you're going to be 90 to 100 and then over time those skills

22:36

kind of start to languish a little bit, and they'll fade. And 25 years out, maybe you'll

22:41

be 80 instead of 100, but you'll still be 80 and if 25 years out you're 50 or 40 or

22:47

30 or 20 or whatever the number is, we found that those people come into the company, they

22:52

just don't understand what's going on or they can't communicate fast enough. They can't

22:57

learn fast enough. Ao they fail at their job or they get frustrated or people get frustrated

23:02

with them and they get pushed out. And in the last -- say I have 20 executives I hire,

23:10

10 of them fail and all 10 of them, you know, had -- had diagnostic scores which were really

23:15

low and we looked past it to the resume and to the credential. Everybody's got good credentials,

23:21

by the way, they've all got a lot of experience. And so the problem with experience is if you're

23:25

really bad at your job you end up having a lot of experience at a lot of different places.

23:32

Yeah. So it's -- that's really difficult and there's an interview bias which is, you know,

23:39

I meet with someone -- we all like the people we meet with because it's a human nature thing

23:44

to like them. But I, you know, there's a phrase in aviation -- I learned to fly when I was

23:49

in the Air Force -- you know, it's like trust your instruments, don't trust what your brain

23:53

is telling you. Trust your instruments because --because your brain is going to kill you.

23:57

The instrument tells you that you're upside down, right, pay attention to the horizon

24:01

instrumentation. The same is true in our business. We're recruiting so fast that we're better

24:07

off to trust our instrumentation. Now of course we're always tuning it and these aren't perfect

24:13

diagnostics so we augment them with capability assessments once people join and performance

24:19

assessments. I tell you -- fascinating -- with performance assessment we used to we used

24:25

to take one assessment per year. Your boss gives you a rating once a year. And then we

24:30

said well how about once a quarter. Your boss gives your rating once a quarter. And then

24:35

we thought, well wouldn't it be good to see what the rest of the people on your team think?

24:38

So then we roll out 360-degree where we ask all the people that work for you whether or

24:43

not they find the engagement constructive. And what about all your peers and your counterparties,

24:48

did they find -- what we'll call collaborators -- did they find it constructive? And then

24:53

what about the two or three superiors, did they find it constructive? Now for two thousand

24:58

people, instead of two thousand data points a year, I end up with two thousand data points

25:03

times ten, or twenty thousand data points a quarter. Eighty thousand a year, okay, and

25:11

if you come into the company and you're just driving everybody crazy, it's not going to

25:16

matter what your academic credentials were right? Maybe you have a degree in how to get

25:19

along with people from Harvard. But then everybody hates you, right? Well, so we now have data

25:26

that tells us. We don't -- we don't have to guess, we don't have to take everybody else's

25:31

word for it. You know, we take this to the extreme. There's engagement analytics now

25:38

where where we actually start to get ratings every time you teach a class. We ask every

25:42

student whether they enjoyed it. Like, we would ask every one of you what do you -- enjoyed

25:46

my speech or didn't enjoy my speech. Every time we have a meeting, we ask people what

25:51

they thought the meeting was constructive. You know, there's this concept Ray Dalio has

25:55

called radical transparency: what if we rate every single meeting? You have 200 meetings

26:00

a year and everybody -- and everybody finds they get a lot from each one of them or you

26:05

have 200 meetings a year and you were doing great in January, in February, and in March

26:10

all your meetings went negative. Now that's -- it's a very interesting world, but here's

26:17

my point. In a world where we didn't collect the data and we couldn't manipulate the data,

26:21

we had to rely upon traditional credentials that came from political -- the established

26:26

institutions, right, and my best bet is I get someone with a master's degree in computer

26:30

science from MIT. Well that's the old world, that's a fragile world. And it's a fragile

26:35

world because MIT only produces a few hundred of those things every single year and that

26:40

cuts out seven billion people who just don't have a chance. The new world is -- we, we

26:46

put in place a more objective set of diagnostics. I would love, right, for someone to put out

26:53

like a branded thermodynamics diagnostic where they tell me this person certainly knows thermodynamics

26:59

as of yesterday. Not -- I know thermodynamics as of when I was at MIT, but I can tell you

27:05

right now, you would want to trust your life to my thermodynamic calculations because I

27:11

forgot, okay, so how do you actually figure out in five seconds that someone didn't forget?

27:18

Now, it might take the, the applicant half an hour or an hour or whatever, but it only

27:26

takes the employer a second. And this is -- this is the important point. Because we -- we want

27:35

to sift through 18,000 people in 30 seconds. Like, there are massive pools of capital right

27:46

now controlled by executives who are trying to solve a big problem in the market or pursue

27:52

an opportunity in the market. Now I -- I am a little executive with a little pool of capital,

27:56

but to put this in perspective, I would hire 100 people like in 30 seconds if you gave

28:03

me that certain objective capability that I want. And I would hire them in Warsaw or

28:09

in China or in some place I'd never been, I'm not going to go. But I could instantly

28:14

create a hundred jobs. I could probably create a hundred jobs to 300 jobs with nothing more

28:20

than a few numbers. What we need is liquid competence information. And it is -- it is

28:29

forming,right, hence the rise of eSkill and hence the rise of HackerRank and the like.

28:35

And when you consider the Googles of the world and the Amazons of the world, they will hire

28:40

a million people in a heartbeat if they actually can get that kind of objective talent. So

28:48

that, I mean, the trend -- the trend is going to continue to capture this mega amount of

28:53

data and as the data forms and the world digitizes, the skills that are economically feasible

29:02

or manipulating that data with these information machines -- those skills you can't -- you

29:08

can test. You know, like, there's a big dataset from Uber. They'll just send you the data

29:13

set. Okay, you've got the dataset. Fix traffic. Okay. A million people can take the dataset

29:20

and it's like they're owning the traffic network of the United States. Fix the traffic. What

29:24

do you need? It used to be you needed a country to fix the traffic and you needed to be able

29:28

experiment on the country. Today, you know, maybe you might not need more than a thousand

29:34

dollar computer to fix the traffic and maybe, maybe nineteen thousand people will try but

29:39

100 will succeed and they'll be really good. So our ability to find the talent is much

29:47

greater. Our need to find it is greater. These, these approaches to finding talent, they're

29:55

all about agile and agile is all about speed, and speed is life. And if I can do things

30:01

fast and I can do them everywhere, you get -- you can see a world where we want to trade

30:07

more, we want to tap into labor pools everywhere, in China and India, wherever the labor might

30:13

be, we want to create the talent and then we want to repurpose the talent. I have services

30:18

I could sell here in the US or in Europe but I don't have the human capital necessary to

30:25

create it and maybe the human capital doesn't exist anywhere. And we need to create it,

30:30

right, so I think as we go toward more automation we'll be able to create that human capital

30:36

and we can then manipulate it or distribute it in order to solve the world's problems.

30:41

So I think in that regard, now, I'll end with this thought, right, the architecture for

30:46

success in the civilization is we start with cheap slash free computer technology and sensing

30:58

technology which just creates the world awash of data, and then we augment that with cheap,

31:04

free digital education. Create software that -- digital software that will provide education,

31:12

right, automate that, the education machine. Once we've done that, if we've got -- if we've

31:17

got an education software apparatus that can be made freely available or cheaply available,

31:23

the next element is free and precise digital certifications of capability and that's what's

31:31

been missing, right, if we produce all these people but our certification is very imprecise.

31:39

Everybody from MIT has the same degree in aeronautics and yet they don't all have the

31:43

same skills in aeronautics. Some are really good at some things, some are not good at

31:46

some things. So it used to be this idea of I'm just going to give you -- you know there

31:52

are business schools that they have this presumption that, well, we're not going to give grades,

31:57

okay, we're too good for that. If you went to whatever -- Stanford Business School -- you

32:02

know, you ought to just accept the Stanford brand and let us do anything we want with

32:05

any of your money. And I'll tell you the problem with that. I have found that there's a theme

32:13

and the theme is business school grads that have been -- that have training in product

32:18

management or training in project development all these things -- they're all failing at

32:23

a massive rate in the real world because that person at Google wants the car to drive via

32:31

LIDAR. They don't want like a theoretical plan. If you don't know what LIDAR is you're

32:37

going to fail no matter how good your b-school degree is. We don't need general skills and

32:42

general problem-solving. What we need is someone that knows that Amazon hasn't deployed its

32:49

full stack of services in Beijing and so your product will absolutely crash and burn in

32:54

Beijing until they do that. There are very particular things that we have to do. And

33:00

if you're going to actually solve these problems in the real world you have to have technique.

33:05

And the technique means, well do you or don't you have enough mathematics to solve this

33:10

problem? If you haven't mastered the calculus of variations you cannot -- you will never

33:15

solve this problem. Doesn't matter whether you have a prestigious degree. I don't care

33:18

if you have a -- four PhDs from the best school on earth. I need to know whether you've mastered

33:23

the calculus of variations. We need that very precise and -- and by the way, I'm using examples

33:31

that were around for a long time, right, Isaac Newton gave us the calculus of variations.

33:35

You want to solve some problems in the tech world, there's stuff that Amazon invented

33:40

last year they're putting into the market this month and if you don't master that particular

33:45

technology you will absolutely fail, right, so it's -- it's a very rapidly expanding universe,

33:52

it's getting more complicated at a rapid rate, and what we want is we want to very precisely

33:59

know what someone can do. Not what -- what they could do twenty years ago, I want to

34:02

know what they can do now. And here's the issue: I don't have eight hours to talk to

34:09

you. I -- I don't have six months to find out, right? I can't afford to take six months

34:15

to try it out. I actually need to scan 937 people in one second and find the one person

34:22

that actually can really do that. Then I'll spend the next six hours figuring out what

34:27

you need to be happy in the organization, and help you transition, and negotiate with

34:32

you, but -- but I need to find the the human capital and we need to create the human capital.

34:37

So -- so when I talk about certification, I mean, we need it precise in an area and

34:43

we -- in our case, like, we rate everybody 0 to 100. Well I -- there's a big difference

34:49

between -- do you want the doctor that was 99 or a hundred you want the doctor that was

34:53

50 on the scale of solving the problem that's about to kill you? I mean you really want

34:57

the best, right, especially if I'm going to write a piece of software that does that thing

35:03

19 million times a minute I really want the best. I don't want the average, right? I don't

35:10

need the average. The world doesn't need a hundred thousand average algebra teachers.

35:17

The world needs one really, really good algebra teacher, then to be automated and manifested

35:22

in software which then delivers an algebra education to the next 10 billion people. We

35:28

just need the best. Nobody wants the average phone, right? The iPhone that you have in

35:33

your pocket is better than every device ever created in the history of man. Nobody wants

35:38

the average one they don't want the best [sic], they want, they want the extreme and they

35:43

want to stamp out a million copies of it. And that gets problematic again as the world's

35:49

products get so diversified. So, I mean, the first step is cheap and free digital education,

35:55

the second is precise certification -- and ideally free precise certifications. That

36:02

should result in a massive increase in human capabilities. Give that to eight billion people

36:07

and we ought to be able to double, triple, quadruple the amount of talent or the amount

36:12

of capability out there. And that creates an acceleration in -- in the rate of human

36:19

capital allocation, right? If we could precisely describe what those billion people could do,

36:24

then we can move them between companies and employers. They could move faster, they can

36:28

change jobs, they can redeploy themselves and that's agile. And that will create a massive

36:36

output in goods services. Everything that we might want, we'll produce more of it. And

36:42

that creates an increase in productivity as a second-order effect, as we find the most

36:46

talented and we replace the crappy program with the better program, with the better program,

36:51

with the better chip. And and that of course is going to increase everybody's quality of

36:57

life. And that will result in a general advance in human knowledge. And if you sum it all

37:04

up, right, I mean, that's the formula for making the world a better place. And there

37:09

are other things people are doing to make the world a better place but, I mean, I think

37:12

we all share an enthusiasm for open education and I see that dynamic of cheap education,

37:22

to more human capital, to more precision, to more agility, to more -- to more engagement

37:28

-- I see that as our best route in the 21st century to make the world a better place.

37:35

I know we can't do it alone. We need to harness the power of every organization -- for-profit,

37:41

nonprofit, governmental -- we possibly can. But I want to thank everybody for being engaged

37:48

in the process and -- and let you know I do appreciate, you know, your support. And anything

37:53

that we can do to help, we will. Thank you. [Applause]

Copied!