TED1994
Danny Hillis: Back to the future (of 1994)
Filmed:
Readability: 3.8
686,810 views
From deep in the TED archive, Danny Hillis outlines an intriguing theory of how and why technological change seems to be accelerating, by linking it to the very evolution of life itself. The presentation techniques he uses may look dated, but the ideas are as relevant as ever.
Danny Hillis - Computer theorist
Inventor, scientist, author, engineer -- over his broad career, Danny Hillis has turned his ever-searching brain on an array of subjects, with surprising results. Full bio
Inventor, scientist, author, engineer -- over his broad career, Danny Hillis has turned his ever-searching brain on an array of subjects, with surprising results. Full bio
Double-click the English transcript below to play the video.
00:15
Because I usually take the role
0
0
3000
00:18
of trying to explain to people
1
3000
2000
00:20
how wonderful the new technologies
2
5000
3000
00:23
that are coming along are going to be,
3
8000
2000
00:25
and I thought that, since I was among friends here,
4
10000
3000
00:28
I would tell you what I really think
5
13000
4000
00:32
and try to look back and try to understand
6
17000
2000
00:34
what is really going on here
7
19000
3000
00:37
with these amazing jumps in technology
8
22000
5000
00:42
that seem so fast that we can barely keep on top of it.
9
27000
3000
00:45
So I'm going to start out
10
30000
2000
00:47
by showing just one very boring technology slide.
11
32000
3000
00:50
And then, so if you can just turn on the slide that's on.
12
35000
3000
00:56
This is just a random slide
13
41000
2000
00:58
that I picked out of my file.
14
43000
2000
01:00
What I want to show you is not so much the details of the slide,
15
45000
3000
01:03
but the general form of it.
16
48000
2000
01:05
This happens to be a slide of some analysis that we were doing
17
50000
3000
01:08
about the power of RISC microprocessors
18
53000
3000
01:11
versus the power of local area networks.
19
56000
3000
01:14
And the interesting thing about it
20
59000
2000
01:16
is that this slide,
21
61000
2000
01:18
like so many technology slides that we're used to,
22
63000
3000
01:21
is a sort of a straight line
23
66000
2000
01:23
on a semi-log curve.
24
68000
2000
01:25
In other words, every step here
25
70000
2000
01:27
represents an order of magnitude
26
72000
2000
01:29
in performance scale.
27
74000
2000
01:31
And this is a new thing
28
76000
2000
01:33
that we talk about technology
29
78000
2000
01:35
on semi-log curves.
30
80000
2000
01:37
Something really weird is going on here.
31
82000
2000
01:39
And that's basically what I'm going to be talking about.
32
84000
3000
01:42
So, if you could bring up the lights.
33
87000
3000
01:47
If you could bring up the lights higher,
34
92000
2000
01:49
because I'm just going to use a piece of paper here.
35
94000
3000
01:52
Now why do we draw technology curves
36
97000
2000
01:54
in semi-log curves?
37
99000
2000
01:56
Well the answer is, if I drew it on a normal curve
38
101000
3000
01:59
where, let's say, this is years,
39
104000
2000
02:01
this is time of some sort,
40
106000
2000
02:03
and this is whatever measure of the technology
41
108000
3000
02:06
that I'm trying to graph,
42
111000
3000
02:09
the graphs look sort of silly.
43
114000
3000
02:12
They sort of go like this.
44
117000
3000
02:15
And they don't tell us much.
45
120000
3000
02:18
Now if I graph, for instance,
46
123000
3000
02:21
some other technology, say transportation technology,
47
126000
2000
02:23
on a semi-log curve,
48
128000
2000
02:25
it would look very stupid, it would look like a flat line.
49
130000
3000
02:28
But when something like this happens,
50
133000
2000
02:30
things are qualitatively changing.
51
135000
2000
02:32
So if transportation technology
52
137000
2000
02:34
was moving along as fast as microprocessor technology,
53
139000
3000
02:37
then the day after tomorrow,
54
142000
2000
02:39
I would be able to get in a taxi cab
55
144000
2000
02:41
and be in Tokyo in 30 seconds.
56
146000
2000
02:43
It's not moving like that.
57
148000
2000
02:45
And there's nothing precedented
58
150000
2000
02:47
in the history of technology development
59
152000
2000
02:49
of this kind of self-feeding growth
60
154000
2000
02:51
where you go by orders of magnitude every few years.
61
156000
3000
02:54
Now the question that I'd like to ask is,
62
159000
3000
02:57
if you look at these exponential curves,
63
162000
3000
03:00
they don't go on forever.
64
165000
3000
03:03
Things just can't possibly keep changing
65
168000
3000
03:06
as fast as they are.
66
171000
2000
03:08
One of two things is going to happen.
67
173000
3000
03:11
Either it's going to turn into a sort of classical S-curve like this,
68
176000
4000
03:15
until something totally different comes along,
69
180000
4000
03:19
or maybe it's going to do this.
70
184000
2000
03:21
That's about all it can do.
71
186000
2000
03:23
Now I'm an optimist,
72
188000
2000
03:25
so I sort of think it's probably going to do something like that.
73
190000
3000
03:28
If so, that means that what we're in the middle of right now
74
193000
3000
03:31
is a transition.
75
196000
2000
03:33
We're sort of on this line
76
198000
2000
03:35
in a transition from the way the world used to be
77
200000
2000
03:37
to some new way that the world is.
78
202000
3000
03:40
And so what I'm trying to ask, what I've been asking myself,
79
205000
3000
03:43
is what's this new way that the world is?
80
208000
3000
03:46
What's that new state that the world is heading toward?
81
211000
3000
03:49
Because the transition seems very, very confusing
82
214000
3000
03:52
when we're right in the middle of it.
83
217000
2000
03:54
Now when I was a kid growing up,
84
219000
3000
03:57
the future was kind of the year 2000,
85
222000
3000
04:00
and people used to talk about what would happen in the year 2000.
86
225000
4000
04:04
Now here's a conference
87
229000
2000
04:06
in which people talk about the future,
88
231000
2000
04:08
and you notice that the future is still at about the year 2000.
89
233000
3000
04:11
It's about as far as we go out.
90
236000
2000
04:13
So in other words, the future has kind of been shrinking
91
238000
3000
04:16
one year per year
92
241000
3000
04:19
for my whole lifetime.
93
244000
3000
04:22
Now I think that the reason
94
247000
2000
04:24
is because we all feel
95
249000
2000
04:26
that something's happening there.
96
251000
2000
04:28
That transition is happening. We can all sense it.
97
253000
2000
04:30
And we know that it just doesn't make too much sense
98
255000
2000
04:32
to think out 30, 50 years
99
257000
2000
04:34
because everything's going to be so different
100
259000
3000
04:37
that a simple extrapolation of what we're doing
101
262000
2000
04:39
just doesn't make any sense at all.
102
264000
3000
04:42
So what I would like to talk about
103
267000
2000
04:44
is what that could be,
104
269000
2000
04:46
what that transition could be that we're going through.
105
271000
3000
04:49
Now in order to do that
106
274000
3000
04:52
I'm going to have to talk about a bunch of stuff
107
277000
2000
04:54
that really has nothing to do
108
279000
2000
04:56
with technology and computers.
109
281000
2000
04:58
Because I think the only way to understand this
110
283000
2000
05:00
is to really step back
111
285000
2000
05:02
and take a long time scale look at things.
112
287000
2000
05:04
So the time scale that I would like to look at this on
113
289000
3000
05:07
is the time scale of life on Earth.
114
292000
3000
05:13
So I think this picture makes sense
115
298000
2000
05:15
if you look at it a few billion years at a time.
116
300000
4000
05:19
So if you go back
117
304000
2000
05:21
about two and a half billion years,
118
306000
2000
05:23
the Earth was this big, sterile hunk of rock
119
308000
3000
05:26
with a lot of chemicals floating around on it.
120
311000
3000
05:29
And if you look at the way
121
314000
2000
05:31
that the chemicals got organized,
122
316000
2000
05:33
we begin to get a pretty good idea of how they do it.
123
318000
3000
05:36
And I think that there's theories that are beginning to understand
124
321000
3000
05:39
about how it started with RNA,
125
324000
2000
05:41
but I'm going to tell a sort of simple story of it,
126
326000
3000
05:44
which is that, at that time,
127
329000
2000
05:46
there were little drops of oil floating around
128
331000
3000
05:49
with all kinds of different recipes of chemicals in them.
129
334000
3000
05:52
And some of those drops of oil
130
337000
2000
05:54
had a particular combination of chemicals in them
131
339000
2000
05:56
which caused them to incorporate chemicals from the outside
132
341000
3000
05:59
and grow the drops of oil.
133
344000
3000
06:02
And those that were like that
134
347000
2000
06:04
started to split and divide.
135
349000
2000
06:06
And those were the most primitive forms of cells in a sense,
136
351000
3000
06:09
those little drops of oil.
137
354000
2000
06:11
But now those drops of oil weren't really alive, as we say it now,
138
356000
3000
06:14
because every one of them
139
359000
2000
06:16
was a little random recipe of chemicals.
140
361000
2000
06:18
And every time it divided,
141
363000
2000
06:20
they got sort of unequal division
142
365000
3000
06:23
of the chemicals within them.
143
368000
2000
06:25
And so every drop was a little bit different.
144
370000
3000
06:28
In fact, the drops that were different in a way
145
373000
2000
06:30
that caused them to be better
146
375000
2000
06:32
at incorporating chemicals around them,
147
377000
2000
06:34
grew more and incorporated more chemicals and divided more.
148
379000
3000
06:37
So those tended to live longer,
149
382000
2000
06:39
get expressed more.
150
384000
3000
06:42
Now that's sort of just a very simple
151
387000
3000
06:45
chemical form of life,
152
390000
2000
06:47
but when things got interesting
153
392000
3000
06:50
was when these drops
154
395000
2000
06:52
learned a trick about abstraction.
155
397000
3000
06:55
Somehow by ways that we don't quite understand,
156
400000
3000
06:58
these little drops learned to write down information.
157
403000
3000
07:01
They learned to record the information
158
406000
2000
07:03
that was the recipe of the cell
159
408000
2000
07:05
onto a particular kind of chemical
160
410000
2000
07:07
called DNA.
161
412000
2000
07:09
So in other words, they worked out,
162
414000
2000
07:11
in this mindless sort of evolutionary way,
163
416000
3000
07:14
a form of writing that let them write down what they were,
164
419000
3000
07:17
so that that way of writing it down could get copied.
165
422000
3000
07:20
The amazing thing is that that way of writing
166
425000
3000
07:23
seems to have stayed steady
167
428000
2000
07:25
since it evolved two and a half billion years ago.
168
430000
2000
07:27
In fact the recipe for us, our genes,
169
432000
3000
07:30
is exactly that same code and that same way of writing.
170
435000
3000
07:33
In fact, every living creature is written
171
438000
3000
07:36
in exactly the same set of letters and the same code.
172
441000
2000
07:38
In fact, one of the things that I did
173
443000
2000
07:40
just for amusement purposes
174
445000
2000
07:42
is we can now write things in this code.
175
447000
2000
07:44
And I've got here a little 100 micrograms of white powder,
176
449000
6000
07:50
which I try not to let the security people see at airports.
177
455000
4000
07:54
(Laughter)
178
459000
2000
07:56
But this has in it --
179
461000
2000
07:58
what I did is I took this code --
180
463000
2000
08:00
the code has standard letters that we use for symbolizing it --
181
465000
3000
08:03
and I wrote my business card onto a piece of DNA
182
468000
3000
08:06
and amplified it 10 to the 22 times.
183
471000
3000
08:09
So if anyone would like a hundred million copies of my business card,
184
474000
3000
08:12
I have plenty for everyone in the room,
185
477000
2000
08:14
and, in fact, everyone in the world,
186
479000
2000
08:16
and it's right here.
187
481000
3000
08:19
(Laughter)
188
484000
5000
08:26
If I had really been a egotist,
189
491000
2000
08:28
I would have put it into a virus and released it in the room.
190
493000
3000
08:31
(Laughter)
191
496000
5000
08:39
So what was the next step?
192
504000
2000
08:41
Writing down the DNA was an interesting step.
193
506000
2000
08:43
And that caused these cells --
194
508000
2000
08:45
that kept them happy for another billion years.
195
510000
2000
08:47
But then there was another really interesting step
196
512000
2000
08:49
where things became completely different,
197
514000
3000
08:52
which is these cells started exchanging and communicating information,
198
517000
3000
08:55
so that they began to get communities of cells.
199
520000
2000
08:57
I don't know if you know this,
200
522000
2000
08:59
but bacteria can actually exchange DNA.
201
524000
2000
09:01
Now that's why, for instance,
202
526000
2000
09:03
antibiotic resistance has evolved.
203
528000
2000
09:05
Some bacteria figured out how to stay away from penicillin,
204
530000
3000
09:08
and it went around sort of creating its little DNA information
205
533000
3000
09:11
with other bacteria,
206
536000
2000
09:13
and now we have a lot of bacteria that are resistant to penicillin,
207
538000
3000
09:16
because bacteria communicate.
208
541000
2000
09:18
Now what this communication allowed
209
543000
2000
09:20
was communities to form
210
545000
2000
09:22
that, in some sense, were in the same boat together;
211
547000
2000
09:24
they were synergistic.
212
549000
2000
09:26
So they survived
213
551000
2000
09:28
or they failed together,
214
553000
2000
09:30
which means that if a community was very successful,
215
555000
2000
09:32
all the individuals in that community
216
557000
2000
09:34
were repeated more
217
559000
2000
09:36
and they were favored by evolution.
218
561000
3000
09:39
Now the transition point happened
219
564000
2000
09:41
when these communities got so close
220
566000
2000
09:43
that, in fact, they got together
221
568000
2000
09:45
and decided to write down the whole recipe for the community
222
570000
3000
09:48
together on one string of DNA.
223
573000
3000
09:51
And so the next stage that's interesting in life
224
576000
2000
09:53
took about another billion years.
225
578000
2000
09:55
And at that stage,
226
580000
2000
09:57
we have multi-cellular communities,
227
582000
2000
09:59
communities of lots of different types of cells,
228
584000
2000
10:01
working together as a single organism.
229
586000
2000
10:03
And in fact, we're such a multi-cellular community.
230
588000
3000
10:06
We have lots of cells
231
591000
2000
10:08
that are not out for themselves anymore.
232
593000
2000
10:10
Your skin cell is really useless
233
595000
3000
10:13
without a heart cell, muscle cell,
234
598000
2000
10:15
a brain cell and so on.
235
600000
2000
10:17
So these communities began to evolve
236
602000
2000
10:19
so that the interesting level on which evolution was taking place
237
604000
3000
10:22
was no longer a cell,
238
607000
2000
10:24
but a community which we call an organism.
239
609000
3000
10:28
Now the next step that happened
240
613000
2000
10:30
is within these communities.
241
615000
2000
10:32
These communities of cells,
242
617000
2000
10:34
again, began to abstract information.
243
619000
2000
10:36
And they began building very special structures
244
621000
3000
10:39
that did nothing but process information within the community.
245
624000
3000
10:42
And those are the neural structures.
246
627000
2000
10:44
So neurons are the information processing apparatus
247
629000
3000
10:47
that those communities of cells built up.
248
632000
3000
10:50
And in fact, they began to get specialists in the community
249
635000
2000
10:52
and special structures
250
637000
2000
10:54
that were responsible for recording,
251
639000
2000
10:56
understanding, learning information.
252
641000
3000
10:59
And that was the brains and the nervous system
253
644000
2000
11:01
of those communities.
254
646000
2000
11:03
And that gave them an evolutionary advantage.
255
648000
2000
11:05
Because at that point,
256
650000
3000
11:08
an individual --
257
653000
3000
11:11
learning could happen
258
656000
2000
11:13
within the time span of a single organism,
259
658000
2000
11:15
instead of over this evolutionary time span.
260
660000
3000
11:18
So an organism could, for instance,
261
663000
2000
11:20
learn not to eat a certain kind of fruit
262
665000
2000
11:22
because it tasted bad and it got sick last time it ate it.
263
667000
4000
11:26
That could happen within the lifetime of a single organism,
264
671000
3000
11:29
whereas before they'd built these special information processing structures,
265
674000
4000
11:33
that would have had to be learned evolutionarily
266
678000
2000
11:35
over hundreds of thousands of years
267
680000
3000
11:38
by the individuals dying off that ate that kind of fruit.
268
683000
3000
11:41
So that nervous system,
269
686000
2000
11:43
the fact that they built these special information structures,
270
688000
3000
11:46
tremendously sped up the whole process of evolution.
271
691000
3000
11:49
Because evolution could now happen within an individual.
272
694000
3000
11:52
It could happen in learning time scales.
273
697000
3000
11:55
But then what happened
274
700000
2000
11:57
was the individuals worked out,
275
702000
2000
11:59
of course, tricks of communicating.
276
704000
2000
12:01
And for example,
277
706000
2000
12:03
the most sophisticated version that we're aware of is human language.
278
708000
3000
12:06
It's really a pretty amazing invention if you think about it.
279
711000
3000
12:09
Here I have a very complicated, messy,
280
714000
2000
12:11
confused idea in my head.
281
716000
3000
12:14
I'm sitting here making grunting sounds basically,
282
719000
3000
12:17
and hopefully constructing a similar messy, confused idea in your head
283
722000
3000
12:20
that bears some analogy to it.
284
725000
2000
12:22
But we're taking something very complicated,
285
727000
2000
12:24
turning it into sound, sequences of sounds,
286
729000
3000
12:27
and producing something very complicated in your brain.
287
732000
4000
12:31
So this allows us now
288
736000
2000
12:33
to begin to start functioning
289
738000
2000
12:35
as a single organism.
290
740000
3000
12:38
And so, in fact, what we've done
291
743000
3000
12:41
is we, humanity,
292
746000
2000
12:43
have started abstracting out.
293
748000
2000
12:45
We're going through the same levels
294
750000
2000
12:47
that multi-cellular organisms have gone through --
295
752000
2000
12:49
abstracting out our methods of recording,
296
754000
3000
12:52
presenting, processing information.
297
757000
2000
12:54
So for example, the invention of language
298
759000
2000
12:56
was a tiny step in that direction.
299
761000
3000
12:59
Telephony, computers,
300
764000
2000
13:01
videotapes, CD-ROMs and so on
301
766000
3000
13:04
are all our specialized mechanisms
302
769000
2000
13:06
that we've now built within our society
303
771000
2000
13:08
for handling that information.
304
773000
2000
13:10
And it all connects us together
305
775000
3000
13:13
into something
306
778000
2000
13:15
that is much bigger
307
780000
2000
13:17
and much faster
308
782000
2000
13:19
and able to evolve
309
784000
2000
13:21
than what we were before.
310
786000
2000
13:23
So now, evolution can take place
311
788000
2000
13:25
on a scale of microseconds.
312
790000
2000
13:27
And you saw Ty's little evolutionary example
313
792000
2000
13:29
where he sort of did a little bit of evolution
314
794000
2000
13:31
on the Convolution program right before your eyes.
315
796000
3000
13:34
So now we've speeded up the time scales once again.
316
799000
3000
13:37
So the first steps of the story that I told you about
317
802000
2000
13:39
took a billion years a piece.
318
804000
2000
13:41
And the next steps,
319
806000
2000
13:43
like nervous systems and brains,
320
808000
2000
13:45
took a few hundred million years.
321
810000
2000
13:47
Then the next steps, like language and so on,
322
812000
3000
13:50
took less than a million years.
323
815000
2000
13:52
And these next steps, like electronics,
324
817000
2000
13:54
seem to be taking only a few decades.
325
819000
2000
13:56
The process is feeding on itself
326
821000
2000
13:58
and becoming, I guess, autocatalytic is the word for it --
327
823000
3000
14:01
when something reinforces its rate of change.
328
826000
3000
14:04
The more it changes, the faster it changes.
329
829000
3000
14:07
And I think that that's what we're seeing here in this explosion of curve.
330
832000
3000
14:10
We're seeing this process feeding back on itself.
331
835000
3000
14:13
Now I design computers for a living,
332
838000
3000
14:16
and I know that the mechanisms
333
841000
2000
14:18
that I use to design computers
334
843000
3000
14:21
would be impossible
335
846000
2000
14:23
without recent advances in computers.
336
848000
2000
14:25
So right now, what I do
337
850000
2000
14:27
is I design objects at such complexity
338
852000
3000
14:30
that it's really impossible for me to design them in the traditional sense.
339
855000
3000
14:33
I don't know what every transistor in the connection machine does.
340
858000
4000
14:37
There are billions of them.
341
862000
2000
14:39
Instead, what I do
342
864000
2000
14:41
and what the designers at Thinking Machines do
343
866000
3000
14:44
is we think at some level of abstraction
344
869000
2000
14:46
and then we hand it to the machine
345
871000
2000
14:48
and the machine takes it beyond what we could ever do,
346
873000
3000
14:51
much farther and faster than we could ever do.
347
876000
3000
14:54
And in fact, sometimes it takes it by methods
348
879000
2000
14:56
that we don't quite even understand.
349
881000
3000
14:59
One method that's particularly interesting
350
884000
2000
15:01
that I've been using a lot lately
351
886000
3000
15:04
is evolution itself.
352
889000
2000
15:06
So what we do
353
891000
2000
15:08
is we put inside the machine
354
893000
2000
15:10
a process of evolution
355
895000
2000
15:12
that takes place on the microsecond time scale.
356
897000
2000
15:14
So for example,
357
899000
2000
15:16
in the most extreme cases,
358
901000
2000
15:18
we can actually evolve a program
359
903000
2000
15:20
by starting out with random sequences of instructions.
360
905000
4000
15:24
Say, "Computer, would you please make
361
909000
2000
15:26
a hundred million random sequences of instructions.
362
911000
3000
15:29
Now would you please run all of those random sequences of instructions,
363
914000
3000
15:32
run all of those programs,
364
917000
2000
15:34
and pick out the ones that came closest to doing what I wanted."
365
919000
3000
15:37
So in other words, I define what I wanted.
366
922000
2000
15:39
Let's say I want to sort numbers,
367
924000
2000
15:41
as a simple example I've done it with.
368
926000
2000
15:43
So find the programs that come closest to sorting numbers.
369
928000
3000
15:46
So of course, random sequences of instructions
370
931000
3000
15:49
are very unlikely to sort numbers,
371
934000
2000
15:51
so none of them will really do it.
372
936000
2000
15:53
But one of them, by luck,
373
938000
2000
15:55
may put two numbers in the right order.
374
940000
2000
15:57
And I say, "Computer,
375
942000
2000
15:59
would you please now take the 10 percent
376
944000
3000
16:02
of those random sequences that did the best job.
377
947000
2000
16:04
Save those. Kill off the rest.
378
949000
2000
16:06
And now let's reproduce
379
951000
2000
16:08
the ones that sorted numbers the best.
380
953000
2000
16:10
And let's reproduce them by a process of recombination
381
955000
3000
16:13
analogous to sex."
382
958000
2000
16:15
Take two programs and they produce children
383
960000
3000
16:18
by exchanging their subroutines,
384
963000
2000
16:20
and the children inherit the traits of the subroutines of the two programs.
385
965000
3000
16:23
So I've got now a new generation of programs
386
968000
3000
16:26
that are produced by combinations
387
971000
2000
16:28
of the programs that did a little bit better job.
388
973000
2000
16:30
Say, "Please repeat that process."
389
975000
2000
16:32
Score them again.
390
977000
2000
16:34
Introduce some mutations perhaps.
391
979000
2000
16:36
And try that again and do that for another generation.
392
981000
3000
16:39
Well every one of those generations just takes a few milliseconds.
393
984000
3000
16:42
So I can do the equivalent
394
987000
2000
16:44
of millions of years of evolution on that
395
989000
2000
16:46
within the computer in a few minutes,
396
991000
3000
16:49
or in the complicated cases, in a few hours.
397
994000
2000
16:51
At the end of that, I end up with programs
398
996000
3000
16:54
that are absolutely perfect at sorting numbers.
399
999000
2000
16:56
In fact, they are programs that are much more efficient
400
1001000
3000
16:59
than programs I could have ever written by hand.
401
1004000
2000
17:01
Now if I look at those programs,
402
1006000
2000
17:03
I can't tell you how they work.
403
1008000
2000
17:05
I've tried looking at them and telling you how they work.
404
1010000
2000
17:07
They're obscure, weird programs.
405
1012000
2000
17:09
But they do the job.
406
1014000
2000
17:11
And in fact, I know, I'm very confident that they do the job
407
1016000
3000
17:14
because they come from a line
408
1019000
2000
17:16
of hundreds of thousands of programs that did the job.
409
1021000
2000
17:18
In fact, their life depended on doing the job.
410
1023000
3000
17:21
(Laughter)
411
1026000
4000
17:26
I was riding in a 747
412
1031000
2000
17:28
with Marvin Minsky once,
413
1033000
2000
17:30
and he pulls out this card and says, "Oh look. Look at this.
414
1035000
3000
17:33
It says, 'This plane has hundreds of thousands of tiny parts
415
1038000
4000
17:37
working together to make you a safe flight.'
416
1042000
4000
17:41
Doesn't that make you feel confident?"
417
1046000
2000
17:43
(Laughter)
418
1048000
2000
17:45
In fact, we know that the engineering process doesn't work very well
419
1050000
3000
17:48
when it gets complicated.
420
1053000
2000
17:50
So we're beginning to depend on computers
421
1055000
2000
17:52
to do a process that's very different than engineering.
422
1057000
4000
17:56
And it lets us produce things of much more complexity
423
1061000
3000
17:59
than normal engineering lets us produce.
424
1064000
2000
18:01
And yet, we don't quite understand the options of it.
425
1066000
3000
18:04
So in a sense, it's getting ahead of us.
426
1069000
2000
18:06
We're now using those programs
427
1071000
2000
18:08
to make much faster computers
428
1073000
2000
18:10
so that we'll be able to run this process much faster.
429
1075000
3000
18:13
So it's feeding back on itself.
430
1078000
3000
18:16
The thing is becoming faster
431
1081000
2000
18:18
and that's why I think it seems so confusing.
432
1083000
2000
18:20
Because all of these technologies are feeding back on themselves.
433
1085000
3000
18:23
We're taking off.
434
1088000
2000
18:25
And what we are is we're at a point in time
435
1090000
3000
18:28
which is analogous to when single-celled organisms
436
1093000
2000
18:30
were turning into multi-celled organisms.
437
1095000
3000
18:33
So we're the amoebas
438
1098000
2000
18:35
and we can't quite figure out what the hell this thing is we're creating.
439
1100000
3000
18:38
We're right at that point of transition.
440
1103000
2000
18:40
But I think that there really is something coming along after us.
441
1105000
3000
18:43
I think it's very haughty of us
442
1108000
2000
18:45
to think that we're the end product of evolution.
443
1110000
3000
18:48
And I think all of us here
444
1113000
2000
18:50
are a part of producing
445
1115000
2000
18:52
whatever that next thing is.
446
1117000
2000
18:54
So lunch is coming along,
447
1119000
2000
18:56
and I think I will stop at that point,
448
1121000
2000
18:58
before I get selected out.
449
1123000
2000
19:00
(Applause)
450
1125000
3000
ABOUT THE SPEAKER
Danny Hillis - Computer theoristInventor, scientist, author, engineer -- over his broad career, Danny Hillis has turned his ever-searching brain on an array of subjects, with surprising results.
Why you should listen
Danny Hillis is an inventor, scientist, author and engineer. While completing his doctorate at MIT, he pioneered the concept of parallel computers that is now the basis for graphics processors and cloud computing. He holds more than 300 US patents, covering parallel computers, disk arrays, forgery prevention methods, various electronic and mechanical devices, and the pinch-to-zoom display interface. He has recently been working on problems in medicine as well. He is also the designer of a 10,000-year mechanical clock, and he gave a TED Talk in 1994 that is practically prophetic. Throughout his career, Hillis has worked at places like Disney, and now MIT and Applied Invention, always looking for the next fascinating problem.
Danny Hillis | Speaker | TED.com