0:00:00.699,0:00:02.757
Today actually is[br]a very special day for me,

0:00:02.757,0:00:04.917
because it is my birthday.

0:00:04.917,0:00:08.944
(Applause)

0:00:09.202,0:00:12.596
And so thanks for all of you[br]for joining the party.

0:00:12.596,0:00:13.787
(Laughter)

0:00:13.787,0:00:18.431
But every time you throw a party,[br]there's someone there to spoil it. Right?

0:00:18.431,0:00:19.601
(Laughter)

0:00:19.601,0:00:21.076
And I'm a physicist,

0:00:21.076,0:00:25.509
and this time I brought[br]another physicist along to do so.

0:00:25.509,0:00:27.473
His name is Albert Einstein, also Albert,

0:00:27.473,0:00:30.751
and he's the one who said that the person

0:00:30.751,0:00:35.943
who has not made his main contributions[br]to science by the age of 30

0:00:35.943,0:00:38.332
will never do so.

0:00:38.332,0:00:40.644
Now you don't need to check Wikipedia

0:00:40.644,0:00:42.695
that I'm being 30.

0:00:42.695,0:00:44.135
(Laughter)

0:00:44.135,0:00:47.262
So effectively what[br]he is telling me, and us,

0:00:47.262,0:00:50.333
is that when it comes to my science,

0:00:50.333,0:00:52.446
I'm a dead wood.

0:00:52.446,0:00:58.324
Well, luckily, I had my share[br]of luck within my career.

0:00:58.324,0:01:02.604
Around age 28, I became[br]very interested in networks,

0:01:02.969,0:01:06.284
and a few years later we managed[br]to kind of publish a few key papers

0:01:06.284,0:01:10.048
that reported the discovery[br]of scale-free networks

0:01:10.048,0:01:12.551
and really gave birth to a new discipline[br]that we call network science today.

0:01:12.551,0:01:18.464
And if you really care about it,[br]you can get a PhD now in network science

0:01:18.464,0:01:20.799
in Budapest, in Boston,

0:01:20.799,0:01:23.320
and you can study it all over the world.

0:01:23.546,0:01:25.017
A few years later,

0:01:25.017,0:01:28.149
when I moved to Harvard[br]first as a sabbatical,

0:01:28.149,0:01:30.964
I became interested[br]in another type of networks,

0:01:30.964,0:01:34.713
and that time the networks[br]within ourselves,

0:01:34.713,0:01:38.632
how the genes and the proteins[br]and the metabolites link to each other,

0:01:38.632,0:01:41.072
and how they connect to disease.

0:01:41.072,0:01:46.291
And that also, that interest led[br]to a major explosion within medicine,

0:01:46.291,0:01:50.153
including the Network Medicine[br]Division at Harvard

0:01:50.153,0:01:53.577
that has more than 300 researchers[br]who are using this perspective

0:01:53.577,0:01:56.777
to treat patients and develop new cures.

0:01:56.777,0:02:01.926
And a few years ago, I thought that I[br]would take this idea of networks

0:02:01.926,0:02:04.996
and expertise we had in networks[br]in a different area,

0:02:04.996,0:02:07.841
that is to understand success.

0:02:07.841,0:02:09.584
And why did we do that?

0:02:09.584,0:02:12.956
Well, we thought that to some degree,

0:02:12.956,0:02:14.904
our success is determined[br]by the networks we are part of,

0:02:14.904,0:02:19.463
that our networks can push us forward,[br]they can pull us back,

0:02:19.463,0:02:23.154
and I was curious if we could use[br]the knowledge in big data and expertise

0:02:23.154,0:02:28.578
that we developed on networks to really[br]quantify how these things happen.

0:02:28.774,0:02:30.220
This is a result from that.

0:02:30.220,0:02:32.914
What you see here is a network[br]of galleries in museums

0:02:32.914,0:02:34.570
that connect to each other,

0:02:34.570,0:02:38.758
and through this map[br]that we mapped out last year,

0:02:38.758,0:02:41.760
we are able to predict very accurately

0:02:41.760,0:02:44.631
the success of an artist

0:02:44.631,0:02:48.819
if you give me the first five exhibits[br]that he or she had in her career.

0:02:48.819,0:02:51.261
Well, as we thought about success,

0:02:51.261,0:02:55.482
we realized that success[br]is not only about networks.

0:02:55.482,0:02:57.979
There are so many[br]other dimensions to that.

0:02:57.979,0:03:01.245
And, you know, one of the things[br]we need for success obviously

0:03:01.245,0:03:02.520
is performance.

0:03:02.520,0:03:06.588
So let's define what's the difference[br]between performance and success.

0:03:06.588,0:03:08.725
Well, performance is what you do,

0:03:08.725,0:03:11.612
how fast you run,[br]what kind of paintings you paint,

0:03:11.612,0:03:13.897
what kind of papers you publish.

0:03:13.897,0:03:16.328
However, in our working definition,

0:03:16.328,0:03:21.326
success is about what does the community[br]notices from what you did,

0:03:21.326,0:03:22.641
from your performance,

0:03:22.641,0:03:26.691
how does it acknowledge it,[br]and how does it reward you for it?

0:03:26.691,0:03:27.715
In other terms,

0:03:27.715,0:03:32.346
your performance is about you,[br]but your success is about all of us.

0:03:33.421,0:03:37.422
And this was a very[br]important shift for us,

0:03:37.422,0:03:41.248
because in the moment we defined success[br]as being a collective measure

0:03:41.248,0:03:43.192
that the community provides to us,

0:03:43.192,0:03:44.997
it became measurable,

0:03:44.997,0:03:49.295
because if it's in the community,[br]there are multiple data points about that.

0:03:49.295,0:03:52.020
So now we believe,[br]and we go to school, we exercise,

0:03:52.020,0:03:56.968
we practice because we believe[br]that performance leads to success.

0:03:58.062,0:04:00.123
But the way we actually[br]started to explore,

0:04:00.123,0:04:03.792
we realized that performance and success[br]are very, very different animals

0:04:03.792,0:04:06.690
when it comes to[br]the mathematics of the problem.

0:04:06.879,0:04:08.599
And let me illustrate that.

0:04:08.599,0:04:12.432
So what you see here is[br]the fastest man on Earth, Usain Bolt.

0:04:12.432,0:04:17.657
And of course, he wins most of[br]the competitions that he enters.

0:04:17.657,0:04:21.517
And we know he's the fastest on Earth[br]because we have a chronometer

0:04:21.517,0:04:23.064
to measure his speed.

0:04:23.064,0:04:27.190
Well, what is interesting about him[br]is that when he wins,

0:04:27.190,0:04:32.226
he doesn't do so by really significantly[br]outrunning his competition.

0:04:32.667,0:04:37.785
He's running at most a percent faster[br]than the one who loses the race.

0:04:37.785,0:04:41.156
And not only does he run only[br]one percent faster than the second one,

0:04:41.156,0:04:44.419
but he doesn't run[br]10 times faster than I do,

0:04:44.419,0:04:46.542
and I'm not a good runner,[br]trust me on that.

0:04:46.542,0:04:47.907
(Laughter)

0:04:47.907,0:04:51.767
And every time we are able[br]to measure performance,

0:04:51.767,0:04:53.405
we notice something very interesting.

0:04:53.405,0:04:56.101
That is, performance is bounded.

0:04:56.101,0:04:59.834
What it means is that there are[br]no huge variations in human performance.

0:04:59.834,0:05:03.449
It varies only in a narrow range,

0:05:03.449,0:05:06.577
and we do need the chronometer[br]to measure the differences.

0:05:06.577,0:05:09.557
This is not to say that we cannot[br]see the good from the best ones,

0:05:09.557,0:05:12.264
but the best ones[br]are very hard to distinguish,

0:05:12.264,0:05:15.400
and the problem with that[br]is that most of us work in areas

0:05:15.400,0:05:19.105
where we do not have a chronometer[br]to gauge our performance.

0:05:19.401,0:05:21.446
Alright, performance is bounded,

0:05:21.446,0:05:24.429
there are no huge differences between us[br]when it comes to our performance.

0:05:24.429,0:05:25.991
How about success?

0:05:25.991,0:05:29.082
Well, let's switch to[br]a different topic like books, right?

0:05:29.082,0:05:32.518
One measure of success for writers[br]is how many people read your work?

0:05:32.518,0:05:39.391
And so when my previous book[br]came out in 2009,

0:05:39.391,0:05:42.263
I was in Europe talking with my editor,[br]and I was interested,

0:05:42.263,0:05:44.269
who is the competition?

0:05:44.269,0:05:47.136
And I had some fabulous one.

0:05:47.136,0:05:49.144
That week (Laughter)

0:05:49.144,0:05:52.354
Dan Brown came out with "The Lost Symbol"

0:05:52.354,0:05:55.011
and "The Last Song" also came out,

0:05:55.011,0:05:57.271
Nicholas Sparks.

0:05:57.271,0:06:01.189
And when you just look at the list,

0:06:01.189,0:06:02.651
you realize, you know, performance-wise

0:06:02.651,0:06:05.550
there is hardly any difference[br]between these books or mine.

0:06:05.550,0:06:06.510
Right?

0:06:06.510,0:06:13.348
So maybe if Nicholas Sparks' team[br]works a little harder,

0:06:13.348,0:06:14.186
he could easily be number one,

0:06:14.186,0:06:16.732
because it's almost by accident[br]who ended up at the top.

0:06:16.936,0:06:19.539
So I said, let's look at the numbers.[br]I am a data person, right?

0:06:19.539,0:06:24.315
So let's see what were[br]the sales for Nicholas Sparks.

0:06:24.315,0:06:26.157
And it turns out that[br]that opening weekend,

0:06:26.157,0:06:28.908
Nicholas Sparks sold more than[br]a hundred thousand copies,

0:06:28.908,0:06:30.468
which is an amazing number.

0:06:30.468,0:06:33.587
You can actually get to the top[br]of the New York Times Bestseller List

0:06:33.587,0:06:36.501
by selling 10,000 copies a week,

0:06:36.501,0:06:40.423
so he tenfold overcame[br]what he needed to be number one.

0:06:40.423,0:06:42.523
Yet he wasn't number one. Why?

0:06:42.523,0:06:46.721
Because there was Dan Brown,[br]who sold 1.2 million copies that week.

0:06:46.721,0:06:48.839
(Laughter)

0:06:48.839,0:06:53.435
And the reason I like this number[br]is because it shows that really,

0:06:53.435,0:06:57.241
when it comes to success, it's unbounded,

0:06:57.241,0:07:02.887
that the best doesn't only get[br]slightly more than the second best,

0:07:02.887,0:07:05.718
but gets orders of magnitude more,

0:07:05.718,0:07:08.605
because success is a collective measure.

0:07:08.605,0:07:12.931
We give it to them, rather than[br]we earn it through our performance.

0:07:12.931,0:07:17.267
So one of things we realized[br]is that performance,

0:07:17.267,0:07:21.104
what we do is bounded but success,[br]which is collective, is unbounded,

0:07:21.104,0:07:24.534
which makes you wonder,

0:07:24.534,0:07:25.589
how do you get these[br]huge differences in success

0:07:25.589,0:07:28.829
when you have so tiny[br]differences in performance?

0:07:28.829,0:07:32.527
And recently I published a book[br]that I devoted to that very question,

0:07:32.527,0:07:35.337
and they didn't give enough time[br]to go over all of that,

0:07:35.337,0:07:38.259
so I'm going to go back[br]to the question of alright,

0:07:38.259,0:07:40.844
you have success, when should that appear?

0:07:40.844,0:07:43.259
So let's go back to the party spoiler,

0:07:43.259,0:07:45.028
and ask ourselves,

0:07:45.028,0:07:48.805
why did Einstein make[br]this ridiculous statement

0:07:48.805,0:07:52.078
that only before 30[br]you could actually be creative?

0:07:52.078,0:07:56.732
Well, because he looked around himself[br]and he saw all these fabulous physicists

0:07:56.732,0:07:59.361
that created quantum mechanics[br]and modern physics,

0:07:59.361,0:08:03.825
and they were all in their 20s[br]and early 30s when they did so.

0:08:03.825,0:08:07.140
And it's not only him.

0:08:07.140,0:08:08.818
It's not only observational bias,

0:08:08.818,0:08:11.007
because there's actually[br]a whole field of genius research

0:08:11.007,0:08:15.778
that has documented the fact[br]that if we look at the people

0:08:15.778,0:08:16.975
we admire from the past

0:08:16.975,0:08:19.938
and then we look at what age[br]did they make their biggest contribution,

0:08:19.938,0:08:21.802
whether that's music,[br]whether that's science,

0:08:21.802,0:08:23.461
whether that's engineering,

0:08:23.461,0:08:27.462
most of them tend to do so[br]in their 20s, 30s, early 40s at most.

0:08:27.462,0:08:33.521
But there's a problem[br]with this genius research.

0:08:33.521,0:08:36.799
Well, first of all, it created[br]the impression to us

0:08:36.799,0:08:40.315
that creativity equals youth,

0:08:40.315,0:08:43.413
which is painful, right?

0:08:43.413,0:08:44.484
(Laughter)

0:08:44.484,0:08:47.863
And it also has an observational bias,

0:08:47.863,0:08:50.827
because it only looks at geniuses

0:08:50.827,0:08:52.811
and doesn't look at ordinary scientists,

0:08:52.811,0:08:55.787
and doesn't look at all of us and asking,

0:08:55.787,0:08:58.478
is it really true that creativity[br]vanishes as we age?

0:08:58.478,0:09:01.776
So that's exactly what we tried to do,[br]and this is important for that

0:09:01.776,0:09:04.434
to actually have references.

0:09:04.434,0:09:07.119
So let's look at an ordinary[br]scientist like myself,

0:09:07.119,0:09:08.753
and let's look at my career.

0:09:08.753,0:09:13.177
So what you see here is all the papers[br]that I published from my first paper,

0:09:13.177,0:09:16.821
it's in 1989 that I was still[br]in Romania when I did so,

0:09:16.821,0:09:19.208
til kind of this year.

0:09:19.208,0:09:21.932
And vertically you see[br]the impact of the paper,

0:09:21.932,0:09:24.401
that is, how many citations,

0:09:24.401,0:09:27.612
how many other papers[br]have been written that cited that work?

0:09:27.612,0:09:31.644
And when you look at that,[br]you see that my career

0:09:31.644,0:09:32.810
has three roughly different stages.

0:09:32.810,0:09:35.556
I had the first 10 years[br]where I work a lot

0:09:35.556,0:09:36.508
and I don't achieve much.

0:09:36.508,0:09:37.754
No one seems to care[br]about what I do, right?

0:09:37.754,0:09:40.926
There's hardly any impact.

0:09:40.926,0:09:44.012
That time, I was doing material science,

0:09:44.012,0:09:47.457
and then I kind of discovered[br]for myself networks,

0:09:47.457,0:09:49.338
and then started publishing in network,

0:09:49.338,0:09:52.736
and that led one high-impact paper[br]to the other one.

0:09:52.736,0:09:55.159
And it really felt good.[br]That was that stage of my career.

0:09:55.159,0:09:57.426
(Laughter)

0:09:57.426,0:10:00.766
So the question is,[br]what happens right now?

0:10:00.766,0:10:08.088
And we don't know, because there[br]hasn't been enough time passed yet

0:10:08.088,0:10:09.182
to actually figure out how much impact[br]those papers will get.

0:10:09.182,0:10:09.924
It takes time to acquire.

0:10:09.924,0:10:10.924
Well, when you look at the data,

0:10:10.924,0:10:12.779
it seems to be that Einstein,[br]the genius research, is right,

0:10:12.779,0:10:14.716
and I'm at that stage of my career.

0:10:14.716,0:10:16.551
(Laughter)

0:10:17.010,0:10:24.535
So we said, OK, let's figure out[br]how does this really happen

0:10:24.535,0:10:28.729
first in science, right?

0:10:28.729,0:10:30.528
And in order not to have[br]the selection bias,

0:10:30.528,0:10:31.368
to look only at geniuses,

0:10:31.368,0:10:33.410
we ended up reconstructing the career[br]of every single scientist

0:10:33.410,0:10:36.057
from 1900 til today

0:10:36.057,0:10:38.395
and finding for all scientists[br]what was their personal best,

0:10:38.395,0:10:42.322
whether they got the Nobel Prize[br]or they never did,

0:10:42.322,0:10:45.720
or no one knows what they did,[br]even their personal best.

0:10:46.071,0:10:48.100
And that's what you see[br]actually in this slide.

0:10:48.100,0:10:50.172
Each line is a career,

0:10:50.172,0:10:52.804
and when you have a light blue dot[br]on the top of that career,

0:10:52.804,0:10:55.425
it says that was their personal best.

0:10:55.425,0:10:56.437
And the question is,

0:10:56.437,0:10:59.451
when did they actually make[br]their biggest discovery?

0:10:59.451,0:11:02.118
And to quantify that, we look at[br]what's the probability

0:11:02.118,0:11:05.018
that you make your biggest discovery[br]let's say, one, two, three,

0:11:05.018,0:11:06.821
or 10 years into your career?

0:11:06.821,0:11:09.945
We're not looking at real age.

0:11:09.945,0:11:12.099
We're looking at[br]what we call academic age.

0:11:12.099,0:11:14.632
Your academic age starts[br]when you publish your first papers.

0:11:14.632,0:11:15.935
I know some of you are still babies.

0:11:15.935,0:11:16.797
(Laughter)

0:11:16.797,0:11:21.704
So let's look at the probability[br]that you publish your highest impact paper

0:11:21.704,0:11:24.898
and what you see is indeed[br]the genius research is right.

0:11:24.898,0:11:28.188
Most scientists tend to publish[br]their highest impact paper

0:11:28.188,0:11:30.666
in the first 10, 15 years in their career,

0:11:30.666,0:11:33.973
and it tanks after that.

0:11:33.973,0:11:38.986
It tanks so fast that I'm about,[br]I'm exactly 30 years into my career,

0:11:38.986,0:11:42.629
and the chance that I will publish a paper[br]that would have higher impact

0:11:42.629,0:11:44.553
than anything that I did before

0:11:44.553,0:11:46.083
is less than one percent.

0:11:46.083,0:11:49.939
I am in that stage of my career,[br]according to this data.

0:11:49.939,0:11:52.054
But there's a problem with that.

0:11:52.054,0:11:55.614
We're not doing controls properly.

0:11:55.614,0:11:57.163
So the control would be,

0:11:57.163,0:12:01.549
how would a scientist look like[br]who makes random contribution to science?

0:12:01.549,0:12:04.966
Or what is the productivity[br]of the scientist?

0:12:04.966,0:12:06.591
When do they write papers?

0:12:06.591,0:12:09.097
So we measured the productivity,

0:12:09.097,0:12:11.717
and amazingly, the productivity,

0:12:11.717,0:12:15.186
your likelihood of writing a paper[br]in year one, 10, or 20 in your career

0:12:15.186,0:12:18.876
is indistinguishable from the likelihood[br]of having the impact

0:12:18.876,0:12:20.582
in that part of your career.

0:12:21.178,0:12:23.429
And to make a long story short,

0:12:23.429,0:12:27.445
after lots of statistical tests,[br]there's only one explanation for that,

0:12:27.445,0:12:30.331
that really the way we scientists work

0:12:30.331,0:12:34.110
is that every single paper we write,[br]every project we do,

0:12:34.110,0:12:37.901
has exactly the same chance[br]of being our personal best.

0:12:37.901,0:12:40.768
That is, discovery is like a lottery,

0:12:40.768,0:12:42.689
and like a lottery ticket,

0:12:42.689,0:12:47.040
and then the more lottery tickets we buy,[br]the higher is the chance,

0:12:47.040,0:12:49.567
and it happens to be so[br]that most scientists

0:12:49.567,0:12:53.848
buy most of their lottery tickets[br]in the first 10, 15 years of their career,

0:12:53.848,0:12:57.762
and after that,[br]their productivity decreases.

0:12:57.762,0:12:59.741
They're not buying[br]any more lottery tickets.

0:12:59.741,0:13:03.323
So it looks as if[br]they would not be creative.

0:13:03.323,0:13:05.877
In reality, they stopped trying.

0:13:05.877,0:13:08.153
So when we actually put the data together,[br]the conclusion is very simple:

0:13:08.153,0:13:09.805
success can come at any time.

0:13:09.805,0:13:15.597
It could be your very first[br]or very last paper of your career.

0:13:15.597,0:13:20.139
It's totally random[br]in the space of the projects.

0:13:20.139,0:13:22.147
It is the productivity that changes.

0:13:22.147,0:13:23.555
And let me illustrate that.

0:13:23.555,0:13:25.021
Here is Frank Wilczek,

0:13:25.021,0:13:28.423
who get the Nobel Prize in Physics[br]for the very first paper

0:13:28.423,0:13:31.902
he ever wrote in his career[br]as a graduate student.

0:13:31.902,0:13:35.052
More interesting is John Fenn,

0:13:35.052,0:13:39.752
who at age 70 was forcefully retired[br]by Yale university.

0:13:39.752,0:13:41.611
They shut his lab down,

0:13:41.611,0:13:45.428
and at that moment, he moved[br]to Virginia Commonwealth University,

0:13:45.428,0:13:47.321
opened another lab,

0:13:47.321,0:13:49.943
and it is there at age 72[br]that he published a paper

0:13:49.943,0:13:55.035
for which 15 years later he got[br]the Nobel Prize for Chemistry.

0:13:55.255,0:13:58.276
And you think, OK,[br]well science is special, right?

0:13:58.276,0:14:00.623
But what about other areas[br]where we need to be creative?

0:14:00.623,0:14:06.757
So let me take another[br]typical example: entrepreneurship.

0:14:06.757,0:14:08.415
Silicon Valley,

0:14:08.415,0:14:11.349
the land of the youth, right?

0:14:11.349,0:14:13.485
And indeed, when you look at it,[br]you realize that the biggest awards,

0:14:13.485,0:14:17.016
the TechCrunch Awards and other awards,

0:14:17.016,0:14:21.169
are all going to people,

0:14:21.169,0:14:24.766
an average age for them[br]is late 20s, very early 30s.

0:14:24.766,0:14:30.416
You look at who the VCs give the money to,[br]some of the biggest VC firms,

0:14:30.416,0:14:33.087
all people in their early 30s.

0:14:33.087,0:14:36.824
Which, of course, we know:

0:14:36.824,0:14:39.013
there is this ethos in Silicon Valley[br]that youth equals success.

0:14:39.013,0:14:42.201
Not when you look at the data,

0:14:42.201,0:14:44.671
because it's not only[br]about forming a company.

0:14:44.671,0:14:47.447
Forming a company is like productivity:[br]trying, trying, trying.

0:14:47.447,0:14:50.134
When you look at which[br]of these individuals actually put out

0:14:50.134,0:14:53.961
a successful company, a successful exit,

0:14:53.961,0:14:57.752
and recently some of our colleagues[br]looked at exactly that question,

0:14:57.752,0:15:00.788
and it turns out that yes,[br]those in the 20s and 30s

0:15:00.788,0:15:02.992
put out a huge number of companies,[br]form lots of companies,

0:15:02.992,0:15:04.428
but most of them go bust,

0:15:04.428,0:15:09.157
and when you look at the successful exits,[br]what you see in this particular plot,

0:15:09.157,0:15:14.879
the older you are, the more likely that[br]you will actually hit the stock market

0:15:14.879,0:15:17.140
or the sell the company successfully.

0:15:17.140,0:15:20.434
This is so strong, actually,[br]that if you are in the 50s,

0:15:20.434,0:15:23.946
you are twice as likely[br]to actually have a successful exit

0:15:23.946,0:15:26.574
than if you are in your 30s.

0:15:26.857,0:15:30.481
(Applause)

0:15:31.921,0:15:35.257
So at the end, what it is[br]that see, actually?

0:15:35.257,0:15:39.067
What we see is that creativity has no age.

0:15:39.067,0:15:41.693
Productivity does, right?

0:15:41.693,0:15:45.176
Which is telling me that[br]at the end of the day,

0:15:45.176,0:15:48.386
if you keep trying

0:15:48.386,0:15:50.366
(Laughter)

0:15:50.366,0:15:53.861
you could still succeed[br]and succeed over and over,

0:15:53.861,0:15:55.908
so my conclusion is very simple.

0:15:55.908,0:15:58.772
I am of this age, back in my lab.

0:15:58.772,0:16:01.161
Thank you.

0:16:01.161,0:16:02.898
(Applause)