ABOUT THE SPEAKER
Astro Teller - Entrepreneur, inventor, author
As "Captain of Moonshots" for X, Astro Teller oversees the secret projects that could reshape our lives in coming decades.

Why you should listen

Astro Teller oversees X, Alphabet's moonshot factory for building magical, audacious ideas that can solve concrete problems for millions of people through science and technology. As X’s head, Teller has an unmatched vantage point from which to watch possible futures unfold.

In addition to his day job shepherding Peter Pans with PhDs, Teller is on the board of several businesses including AI­-based hedge fund Cerebellum Capital, Inc., and Flux.io, a startup reinventing how buildings are designed and built. He is also the author of two novels and co-author (with Danielle Teller) of Sacred Cows, a non­fiction work analyzing society’s attitudes on divorce.

More profile about the speaker
Astro Teller | Speaker | TED.com
TED2016

Astro Teller: The unexpected benefit of celebrating failure

Filmed:
2,686,287 views

"Great dreams aren't just visions," says Astro Teller, "They're visions coupled to strategies for making them real." The head of X (formerly Google X), Teller takes us inside the "moonshot factory," as it's called, where his team seeks to solve the world's biggest problems through experimental projects like balloon-powered Internet and wind turbines that sail through the air. Find out X's secret to creating an organization where people feel comfortable working on big, risky projects and exploring audacious ideas.
- Entrepreneur, inventor, author
As "Captain of Moonshots" for X, Astro Teller oversees the secret projects that could reshape our lives in coming decades. Full bio

Double-click the English transcript below to play the video.

00:12
In 1962 at Rice University,
0
760
3320
00:16
JFK told the country about a dream he had,
1
4760
4416
00:21
a dream to put a person on the moon
by the end of the decade.
2
9200
3896
00:25
The eponymous moonshot.
3
13120
1400
00:28
No one knew if it was possible to do
4
16120
2576
00:30
but he made sure a plan was put in place
to do it if it was possible.
5
18720
4360
00:35
That's how great dreams are.
6
23560
1776
00:37
Great dreams aren't just visions,
7
25360
2456
00:39
they're visions coupled to strategies
for making them real.
8
27840
3680
00:44
I have the incredible good fortune
to work at a moonshot factory.
9
32080
5416
00:49
At X -- formerly called Google X --
10
37520
2296
00:51
you'll find an aerospace engineer
working alongside a fashion designer
11
39840
4696
00:56
and former military ops commanders
brainstorming with laser experts.
12
44560
4720
01:02
These inventors, engineers and makers
are dreaming up technologies
13
50360
5176
01:07
that we hope can make the world
a wonderful place.
14
55560
3760
01:12
We use the word "moonshots"
to remind us to keep our visions big --
15
60200
4536
01:16
to keep dreaming.
16
64760
1696
01:18
And we use the word "factory"
to remind ourselves
17
66480
4456
01:22
that we want to have concrete visions --
18
70960
3096
01:26
concrete plans to make them real.
19
74080
3296
01:29
Here's our moonshot blueprint.
20
77400
1936
01:31
Number one:
21
79360
1216
01:32
we want to find
a huge problem in the world
22
80600
2416
01:35
that affects many millions of people.
23
83040
2256
01:37
Number two:
24
85320
1496
01:38
we want to find or propose a radical
solution for solving that problem.
25
86840
3800
01:43
And then number three:
26
91480
1376
01:44
there has to be some reason to believe
27
92880
2976
01:47
that the technology
for such a radical solution
28
95880
3056
01:50
could actually be built.
29
98960
1440
01:53
But I have a secret for you.
30
101880
1616
01:55
The moonshot factory is a messy place.
31
103520
3240
01:59
But rather than avoid the mess,
32
107600
1816
02:01
pretend it's not there,
33
109440
1536
02:03
we've tried to make that our strength.
34
111000
1840
02:06
We spend most of our time breaking things
35
114080
2816
02:08
and trying to prove that we're wrong.
36
116920
1960
02:12
That's it, that's the secret.
37
120320
1640
02:14
Run at all the hardest
parts of the problem first.
38
122680
3536
02:18
Get excited and cheer,
39
126240
2055
02:20
"Hey! How are we going
to kill our project today?"
40
128320
2439
02:25
We've got this interesting balance going
41
133920
2720
02:29
where we allow our unchecked
optimism to fuel our visions.
42
137560
5536
02:35
But then we also harness
enthusiastic skepticism
43
143120
4456
02:39
to breathe life, breathe reality
into those visions.
44
147600
3040
02:43
I want to show you a few of the projects
45
151320
1936
02:45
that we've had to leave behind
on the cutting room floor,
46
153280
2696
02:48
and also a few of the gems
47
156000
1576
02:49
that at least so far,
have not only survived that process,
48
157600
3696
02:53
but have been accelerated by it.
49
161320
2760
02:56
Last year we killed a project
in automated vertical farming.
50
164760
4416
03:01
This is some of the lettuce that we grew.
51
169200
2016
03:03
One in nine people in the world
suffers from undernourishment.
52
171240
3576
03:06
So this is a moonshot
that needs to happen.
53
174840
3096
03:09
Vertical farming uses 10 times less water
54
177960
2656
03:12
and a hundred times less land
than conventional farming.
55
180640
4136
03:16
And because you can grow the food
close to where it's consumed,
56
184800
4160
03:22
you don't have to
transport it large distances.
57
190200
2680
03:27
We made progress in a lot of the areas
58
195160
2296
03:29
like automated harvesting
and efficient lighting.
59
197480
3616
03:33
But unfortunately,
60
201120
1456
03:34
we couldn't get staple crops
like grains and rice to grow this way.
61
202600
4056
03:38
So we killed the project.
62
206680
1520
03:41
Here's another huge problem.
63
209120
1360
03:43
We pay enormous costs in resources
and environmental damage
64
211240
5176
03:48
to ship goods worldwide.
65
216440
2216
03:50
Economic development
of landlocked countries
66
218680
2576
03:53
is limited by lack
of shipping infrastructure.
67
221280
3576
03:56
The radical solution?
68
224880
1616
03:58
A lighter-than-air,
variable-buoyancy cargo ship.
69
226520
4200
04:04
This has the potential to lower,
70
232240
3376
04:07
at least overall,
71
235640
1616
04:09
the cost, time and carbon
footprint of shipping
72
237280
4536
04:13
without needing runways.
73
241840
1280
04:15
We came up with this clever set
of technical breakthroughs
74
243880
3056
04:18
that together might make it possible
for us to lower the cost enough
75
246960
4936
04:23
that we could actually make these ships --
76
251920
2136
04:26
inexpensively enough in volume.
77
254080
1640
04:28
But however cheap they would
have been to make in volume
78
256760
3319
04:33
it turned out that it was going to cost
close to 200 million dollars
79
261320
3856
04:37
to design and build the first one.
80
265200
2120
04:40
200 million dollars
is just way too expensive.
81
268680
3656
04:44
Because X is structured
with these tight feedback loops
82
272360
3816
04:48
of making mistakes
and learning and new designs,
83
276200
3656
04:51
we can't spend 200 million dollars
84
279880
1736
04:53
to get the first data point
85
281640
1736
04:55
about whether we're
on the right track or not.
86
283400
2560
04:58
If there's an Achilles' heel
in one our projects,
87
286600
2336
05:00
we want to know it now, up front,
not way down the road.
88
288960
4296
05:05
So we killed this project, too.
89
293280
1680
05:08
Discovering a major flaw in a project
90
296880
2736
05:11
doesn't always mean
that it ends the project.
91
299640
2136
05:13
Sometimes it actually gets us
onto a more productive path.
92
301800
3176
05:17
This is our fully self-driving
vehicle prototype,
93
305000
3936
05:20
which we built without
a steering wheel or break pedal.
94
308960
3160
05:24
But that wasn't actually
our goal when we started.
95
312800
2400
05:27
With 1.2 million people dying
on the roads globally every year,
96
315760
5896
05:33
building a car that drives itself
was a natural moonshot to take.
97
321680
3816
05:37
Three and a half years ago,
98
325520
1600
05:39
when we had these Lexus,
retrofitted, self-driving cars in testing,
99
327920
5576
05:45
they were doing so well,
we gave them out to other Googlers
100
333520
3656
05:49
to find out what they thought
of the experience.
101
337200
2240
05:52
And what we discovered
102
340000
1256
05:53
was that our plan to have the cars
do almost all the driving
103
341280
4616
05:57
and just hand over to the users
in case of emergency
104
345920
3896
06:01
was a really bad plan.
105
349840
3080
06:06
It wasn't safe
106
354200
1656
06:07
because the users didn't do their job.
107
355880
3136
06:11
They didn't stay alert
108
359040
1440
06:13
in case the car needed
to hand control back to them.
109
361160
4056
06:17
This was a major crisis for the team.
110
365240
2016
06:19
It sent them back to the drawing board.
111
367280
2040
06:22
And they came up
with a beautiful, new perspective.
112
370320
3360
06:26
Aim for a car where
you're truly a passenger.
113
374600
3720
06:31
You tell the car where you want to go,
114
379280
2256
06:33
you push a button
115
381560
1256
06:34
and it takes you
from point A to point B by itself.
116
382840
4160
06:39
We're really grateful
117
387720
1296
06:41
that we had this insight
as early on in the project as we did.
118
389040
3536
06:44
And it's shaped everything
we've done since then.
119
392600
3960
06:49
And now our cars have self-driven
more than 1.4 million miles,
120
397200
6336
06:55
and they're out everyday
121
403560
1616
06:57
on the streets of Mountain View,
California and Austin, Texas.
122
405200
4040
07:02
The cars team shifted their perspective.
123
410120
3856
07:06
This is one of X's mantras.
124
414000
2056
07:08
Sometimes shifting your perspective
is more powerful than being smart.
125
416080
4800
07:13
Take wind energy.
126
421880
1496
07:15
It's one of my favorite examples
of perspective shifting.
127
423400
3336
07:18
There's no way that we're going to build
128
426760
2856
07:21
a better standard wind turbine
than the experts in that industry.
129
429640
4400
07:26
But we found a way
to get up higher into the sky,
130
434720
4896
07:31
and so get access to faster,
more consistent winds,
131
439640
3816
07:35
and so more energy without needing
hundreds of tons of steel to get there.
132
443480
5656
07:41
Our Makani energy kite
rises up from its perch
133
449160
3336
07:44
by spinning up those
propellers along its wing.
134
452520
3016
07:47
And it pulls out a tether as it rises,
135
455560
2320
07:50
pulling energy up through the tether.
136
458640
1896
07:52
Once the tether's all the way out,
137
460560
1856
07:54
it goes into crosswind circles in the sky.
138
462440
3160
07:58
And now those propellers that lifted it up
have become flying turbines.
139
466320
5360
08:04
And that sends energy
back down the tether.
140
472400
2680
08:07
We haven't yet found
a way to kill this project.
141
475880
2856
08:10
And the longer it survives that pressure,
the more excited we get
142
478760
4096
08:14
that this could become
a cheaper and more deployable form
143
482880
4456
08:19
of wind energy for the world.
144
487360
1920
08:23
Probably the craziest sounding project
we have is Project Loon.
145
491080
3480
08:27
We're trying to make
balloon-powered Internet.
146
495320
3536
08:30
A network of balloons in the stratosphere
147
498880
3080
08:34
that beam an internet connection down
to rural and remote areas of the world.
148
502800
5096
08:39
This could bring online
as many as four billion more people,
149
507920
3896
08:43
who today have little
or no internet connection.
150
511840
2240
08:47
But you can't just take a cell tower,
151
515120
2895
08:50
strap it to a balloon
and stick it in the sky.
152
518039
2897
08:52
The winds are too strong,
it would be blown away.
153
520960
2376
08:55
And the balloons are too high up
to tie it to the ground.
154
523360
4400
09:00
Here comes the crazy moment.
155
528800
1440
09:03
What if, instead,
156
531000
2096
09:05
we let the balloons drift
157
533120
2216
09:07
and we taught them how to sail the winds
to go where the needed to go?
158
535360
3840
09:13
It turns out the stratosphere has winds
159
541560
3336
09:16
that are going in quite different
speeds and directions in thin strata.
160
544920
4096
09:21
So we hoped that using smart algorithms
and wind data from around the world,
161
549040
4880
09:26
we could maneuver the balloons a bit,
162
554800
2456
09:29
getting them to go up and down
just a tiny bit in the stratosphere
163
557280
3096
09:32
to grab those winds going
in those different directions and speeds.
164
560400
3616
09:36
The idea is to have enough balloons
165
564040
2296
09:38
so as one balloon floats out of your area,
166
566360
2856
09:41
there's another balloon
ready to float into place,
167
569240
3736
09:45
handing off the internet connection,
168
573000
2616
09:47
just like your phone
hands off between cell towers
169
575640
3456
09:51
as you drive down the freeway.
170
579120
1640
09:53
We get how crazy that vision sounds --
171
581760
4416
09:58
there's the name of the project
to remind us of that.
172
586200
2736
10:00
So since 2012,
173
588960
2336
10:03
the Loon team has prioritized
the work that seems the most difficult
174
591320
5776
10:09
and so the most likely
to kill their project.
175
597120
3840
10:13
The first thing that they did
176
601920
1536
10:15
was try to get a Wi-Fi connection
from a balloon in the stratosphere
177
603480
4216
10:19
down to an antenna on the ground.
178
607720
2120
10:22
It worked.
179
610720
1216
10:23
And I promise you there were bets
that it wasn't going to.
180
611960
2715
10:26
So we kept going.
181
614699
1240
10:28
Could we get the balloon
to talk directly to handsets,
182
616880
3376
10:32
so that we didn't need the antenna
as an intermediary receiver?
183
620280
3360
10:36
Yeah.
184
624720
1200
10:38
Could we get the balloon
bandwidth high enough
185
626680
2616
10:41
so it was a real Internet connection?
186
629320
2736
10:44
So that people could have
something more than just SMS?
187
632080
4960
10:49
The early tests weren't even
a megabit per second,
188
637640
3816
10:53
but now we can do
up to 15 megabits per second.
189
641480
3416
10:56
Enough to watch a TED Talk.
190
644920
1320
10:59
Could we get the balloons
to talk to each other through the sky
191
647360
3576
11:02
so that we could reach our signal
deeper into rural areas?
192
650960
4856
11:07
Check.
193
655840
1536
11:09
Could we get balloons the size of a house
to stay up for more than 100 days,
194
657400
6656
11:16
while costing less than five percent
195
664080
2296
11:18
of what traditional, long-life
balloons have cost to make?
196
666400
3560
11:23
Yes. In the end.
197
671640
1896
11:25
But I promise you, you name it,
we had to try it to get there.
198
673560
5040
11:31
We made round, silvery balloons.
199
679480
2840
11:35
We made giant pillow-shaped balloons.
200
683200
3440
11:40
We made balloons the size of a blue whale.
201
688400
2880
11:44
We busted a lot of balloons.
202
692440
2936
11:47
(Laughter)
203
695400
1616
11:49
Since one of the things that was
most likely to kill the Loon project
204
697040
4056
11:53
was whether we could guide
the balloons through the sky,
205
701120
3056
11:56
one of our most important experiments
was putting a balloon inside a balloon.
206
704200
4760
12:02
So there are two compartments here,
one with air and then one with helium.
207
710280
3800
12:07
The balloon pumps air in
to make itself heavier,
208
715360
3856
12:11
or lets air out to make it lighter.
209
719240
2360
12:14
And these weight changes
allow it to rise or fall,
210
722520
3296
12:17
and that simple movement of the balloon
is its steering mechanism.
211
725840
3736
12:21
It floats up or down,
212
729600
1656
12:23
hoping to grab winds going in the speed
and direction that it wants.
213
731280
3920
12:27
But is that good enough
for it to navigate through the world?
214
735920
4200
12:32
Barely at first,
215
740800
2536
12:35
but better all the time.
216
743360
1376
12:36
This particular balloon,
our latest balloon,
217
744760
2080
12:39
can navigate a two-mile
vertical stretch of sky
218
747680
3560
12:44
and can sail itself to within 500 meters
of where it wants to go
219
752280
4056
12:48
from 20,000 kilometers away.
220
756360
2600
12:56
We have lots more to do
221
764360
2216
12:58
in terms of fine-tuning
the system and reducing costs.
222
766600
4576
13:03
But last year,
a balloon built inexpensively
223
771200
3016
13:06
went around the world
19 times over 187 days.
224
774240
6136
13:12
So we're going to keep going.
225
780400
1440
13:15
(Applause)
226
783120
3320
13:21
Our balloons today
227
789520
1256
13:22
are doing pretty much everything
a complete system needs to do.
228
790800
3360
13:27
We're in discussions
with telcos around the world,
229
795160
2856
13:30
and we're going to fly
over places like Indonesia
230
798040
2336
13:32
for real service testing this year.
231
800400
2736
13:35
This probably all sounds
too good to be true,
232
803160
2560
13:38
and you're right.
233
806800
1200
13:41
Being audacious
234
809320
1736
13:43
and working on big, risky things
235
811080
2656
13:45
makes people inherently uncomfortable.
236
813760
2480
13:49
You cannot yell at people
and force them to fail fast.
237
817040
5080
13:55
People resist. They worry.
238
823440
2096
13:57
"What will happen to me if I fail?
239
825560
3056
14:00
Will people laugh at me?
240
828640
1480
14:02
Will I be fired?"
241
830880
1440
14:05
I started with our secret.
242
833880
2696
14:08
I'm going to leave you
with how we actually make it happen.
243
836600
2880
14:12
The only way to get people
to work on big, risky things --
244
840920
5016
14:17
audacious ideas --
245
845960
1736
14:19
and have them run at all
the hardest parts of the problem first,
246
847720
4936
14:24
is if you make that the path
of least resistance for them.
247
852680
3880
14:29
We work hard at X to make it safe to fail.
248
857360
3880
14:34
Teams kill their ideas
as soon as the evidence is on the table
249
862000
3936
14:37
because they're rewarded for it.
250
865960
2560
14:41
They get applause from their peers.
251
869160
2816
14:44
Hugs and high fives
from their manager, me in particular.
252
872000
3880
14:49
They get promoted for it.
253
877160
2496
14:51
We have bonused every single person
on teams that ended their projects,
254
879680
6256
14:57
from teams as small as two
to teams of more than 30.
255
885960
3016
15:01
We believe in dreams
at the moonshot factory.
256
889000
2920
15:04
But enthusiastic skepticism
257
892840
2496
15:07
is not the enemy of boundless optimism.
258
895360
3456
15:10
It's optimism's perfect partner.
259
898840
2680
15:14
It unlocks the potential in every idea.
260
902280
4320
15:19
We can create the future
that's in our dreams.
261
907720
3760
15:24
Thank you very much.
262
912400
1296
15:25
(Applause)
263
913720
6022

▲Back to top

ABOUT THE SPEAKER
Astro Teller - Entrepreneur, inventor, author
As "Captain of Moonshots" for X, Astro Teller oversees the secret projects that could reshape our lives in coming decades.

Why you should listen

Astro Teller oversees X, Alphabet's moonshot factory for building magical, audacious ideas that can solve concrete problems for millions of people through science and technology. As X’s head, Teller has an unmatched vantage point from which to watch possible futures unfold.

In addition to his day job shepherding Peter Pans with PhDs, Teller is on the board of several businesses including AI­-based hedge fund Cerebellum Capital, Inc., and Flux.io, a startup reinventing how buildings are designed and built. He is also the author of two novels and co-author (with Danielle Teller) of Sacred Cows, a non­fiction work analyzing society’s attitudes on divorce.

More profile about the speaker
Astro Teller | Speaker | TED.com