1
00:00:19,220 --> 00:00:23,840
PATRICK WINSTON: The Uniform
Code of Military Justice

2
00:00:23,840 --> 00:00:27,530
specifies court martial for
any officer who sends a soldier

3
00:00:27,530 --> 00:00:30,620
into battle without a weapon.

4
00:00:30,620 --> 00:00:32,995
There ought to be a similar
protection for students

5
00:00:32,995 --> 00:00:34,370
because students
shouldn't go out

6
00:00:34,370 --> 00:00:38,070
into life without the
ability to communicate,

7
00:00:38,070 --> 00:00:40,340
and that's because
your success in life

8
00:00:40,340 --> 00:00:43,760
will be determined largely
by your ability to speak,

9
00:00:43,760 --> 00:00:47,840
your ability to write, and
the quality of your ideas,

10
00:00:47,840 --> 00:00:50,840
in that order.

11
00:00:50,840 --> 00:00:52,670
I know that I can be
successful in this

12
00:00:52,670 --> 00:00:57,050
because the quality
of communication,

13
00:00:57,050 --> 00:00:59,450
your speaking, your
writing, is largely

14
00:00:59,450 --> 00:01:01,880
determined by this formula.

15
00:01:01,880 --> 00:01:05,150
It's a matter of how
much knowledge you have,

16
00:01:05,150 --> 00:01:07,670
how much you practice
with that knowledge,

17
00:01:07,670 --> 00:01:10,100
and your inherent
talent, and notice

18
00:01:10,100 --> 00:01:12,230
that the T is very small.

19
00:01:12,230 --> 00:01:15,860
What really matters
is what you know.

20
00:01:15,860 --> 00:01:19,970
This point came to me suddenly
a few decades ago when

21
00:01:19,970 --> 00:01:22,610
I was skiing at Sun Valley.

22
00:01:22,610 --> 00:01:25,430
I had heard that it was
Celebrity Weekend, and one

23
00:01:25,430 --> 00:01:30,800
of the celebrities
was Mary Lou Retton,

24
00:01:30,800 --> 00:01:35,480
famous Olympic gymnast,
perfect 10s in the vault.

25
00:01:35,480 --> 00:01:38,510
And I heard that she
was a novice at skiing,

26
00:01:38,510 --> 00:01:40,070
so when the opportune
moment arrived,

27
00:01:40,070 --> 00:01:42,470
I looked over on
the novice slope

28
00:01:42,470 --> 00:01:46,270
and saw this young woman who,
when she became unbalanced,

29
00:01:46,270 --> 00:01:48,350
went like that.

30
00:01:48,350 --> 00:01:50,330
And I said that's got to be her.

31
00:01:50,330 --> 00:01:53,570
That must be the gymnast.

32
00:01:53,570 --> 00:01:56,210
But then, it occurred to
me, I'm a much better skier

33
00:01:56,210 --> 00:01:58,820
than she is, and she's an
Olympic athlete-- not only

34
00:01:58,820 --> 00:02:03,230
an ordinary Olympic
athlete, an outstanding one.

35
00:02:03,230 --> 00:02:08,120
And I was a better skier because
I had the K, and I had the P,

36
00:02:08,120 --> 00:02:10,729
and all she had
was the T. So you

37
00:02:10,729 --> 00:02:12,860
can get a lot better
than people who

38
00:02:12,860 --> 00:02:15,140
may have inherent
talents if you have

39
00:02:15,140 --> 00:02:17,460
the right amount of knowledge.

40
00:02:17,460 --> 00:02:20,960
So that's what my objective is
today, and here's my promise.

41
00:02:20,960 --> 00:02:24,830
Today, you will
see some examples

42
00:02:24,830 --> 00:02:28,400
of what you can put in your
armamentarium of speaking

43
00:02:28,400 --> 00:02:32,810
techniques, and it
will be the case

44
00:02:32,810 --> 00:02:37,640
that some one of those examples,
some heuristic, some technique,

45
00:02:37,640 --> 00:02:40,040
maybe only one, will make--

46
00:02:40,040 --> 00:02:43,050
will be the one that
gets you the job.

47
00:02:43,050 --> 00:02:45,600
And so this is a very
non-linear process.

48
00:02:45,600 --> 00:02:47,752
You never know when
it's going to happen,

49
00:02:47,752 --> 00:02:48,710
but that is my promise.

50
00:02:48,710 --> 00:02:52,350
By the end of the
next 60 minutes,

51
00:02:52,350 --> 00:02:55,490
you'll have been exposed to
a lot of ideas, some of which

52
00:02:55,490 --> 00:02:58,180
you'll incorporate into
your own repertoire,

53
00:02:58,180 --> 00:03:03,080
and they will ensure that you
get the maximum opportunity

54
00:03:03,080 --> 00:03:09,410
to have your ideas
valued and accepted

55
00:03:09,410 --> 00:03:11,900
by the people you speak with.

56
00:03:11,900 --> 00:03:14,300
Now, in order to
do that, we have

57
00:03:14,300 --> 00:03:17,150
to have a rule of
engagement, and that

58
00:03:17,150 --> 00:03:22,710
is no laptops, no cell phones.

59
00:03:22,710 --> 00:03:24,600
So if you could
close those, I'll

60
00:03:24,600 --> 00:03:27,300
start up as soon as you're done.

61
00:03:27,300 --> 00:03:31,800
Some people ask why that
is a rule of engagement,

62
00:03:31,800 --> 00:03:34,830
and the answer
is, we humans only

63
00:03:34,830 --> 00:03:36,990
have one language processor.

64
00:03:36,990 --> 00:03:38,760
And if your language
processor is enga--

65
00:03:38,760 --> 00:03:40,590
could you shut the
laptop, please?

66
00:03:40,590 --> 00:03:44,850
If your language processor
is engaged browsing the web

67
00:03:44,850 --> 00:03:48,030
or reading your email,
you're distracted.

68
00:03:48,030 --> 00:03:50,880
And worse yet, you distract
all of the people around you.

69
00:03:50,880 --> 00:03:52,570
Studies have shown that.

70
00:03:52,570 --> 00:03:56,160
And worse yet, if I see a
open laptop somewhere back

71
00:03:56,160 --> 00:04:02,470
there or up here, it drives
me nuts, and I do a worse job.

72
00:04:02,470 --> 00:04:04,710
And so that ensures
that all of your friends

73
00:04:04,710 --> 00:04:08,520
who are paying attention
don't get the performance

74
00:04:08,520 --> 00:04:11,230
that they came to have.

75
00:04:11,230 --> 00:04:13,830
So that's it for preamble.

76
00:04:13,830 --> 00:04:16,620
Let's get started.

77
00:04:16,620 --> 00:04:19,670
First thing we talk about,
of course, is how to start.

78
00:04:25,470 --> 00:04:28,740
Some people think
the right thing to do

79
00:04:28,740 --> 00:04:32,140
is to start a talk with a joke.

80
00:04:39,530 --> 00:04:44,300
I don't recommend
it, and the reason

81
00:04:44,300 --> 00:04:46,820
is that, in the
beginning of a talk,

82
00:04:46,820 --> 00:04:49,420
people are still putting
their laptops away.

83
00:04:49,420 --> 00:04:52,250
They're becoming adjusted
to your speaking parameters,

84
00:04:52,250 --> 00:04:58,670
to your vocal parameters, and
they're not ready for a joke.

85
00:04:58,670 --> 00:05:00,090
So it doesn't work very well.

86
00:05:00,090 --> 00:05:02,420
They usually fall flat.

87
00:05:02,420 --> 00:05:06,110
What you want to do instead is
start with empowerment promise.

88
00:05:09,560 --> 00:05:11,060
You want to tell
people what they're

89
00:05:11,060 --> 00:05:13,660
going to know at
the end of the hour

90
00:05:13,660 --> 00:05:16,560
that they didn't know at
the beginning of the hour.

91
00:05:16,560 --> 00:05:17,808
It's an empowerment promise.

92
00:05:17,808 --> 00:05:19,100
It's the reason for being here.

93
00:05:21,660 --> 00:05:22,910
What would be an example?

94
00:05:22,910 --> 00:05:24,230
Oh, I see.

95
00:05:24,230 --> 00:05:28,820
At the end of this 60
minutes, you will know things

96
00:05:28,820 --> 00:05:30,952
about speaking you
don't know now,

97
00:05:30,952 --> 00:05:32,660
and something among
those things you know

98
00:05:32,660 --> 00:05:34,310
will make a difference
in your life.

99
00:05:34,310 --> 00:05:36,200
Yeah, that's an
empowerment promise,

100
00:05:36,200 --> 00:05:38,960
so that's the best way to start.

101
00:05:38,960 --> 00:05:41,540
So now that I've talked a
little bit about how to start,

102
00:05:41,540 --> 00:05:44,900
what I want to do is give you
some samples of heuristics

103
00:05:44,900 --> 00:05:48,970
that are always on my
mind when I give a talk,

104
00:05:48,970 --> 00:05:53,185
and first of these heuristics is
that it's a good idea to cycle

105
00:05:53,185 --> 00:05:53,810
on the subject.

106
00:05:56,480 --> 00:05:57,980
Go around it.

107
00:05:57,980 --> 00:05:59,360
Go round it again.

108
00:05:59,360 --> 00:06:01,250
Go round it again.

109
00:06:01,250 --> 00:06:05,900
Some people say, tell him
what you want to tell him.

110
00:06:05,900 --> 00:06:08,690
Tell him again, and then
tell him a third time,

111
00:06:08,690 --> 00:06:11,900
as if people
weren't intelligent.

112
00:06:11,900 --> 00:06:12,830
But the point is--

113
00:06:12,830 --> 00:06:16,190
the reason is-- well, there
are many reasons, one of which

114
00:06:16,190 --> 00:06:19,340
is, at any given
moment, about 20% of you

115
00:06:19,340 --> 00:06:22,030
will be fogged out no
matter what the lecture is.

116
00:06:22,030 --> 00:06:24,530
So if you want to ensure that
the probability that everybody

117
00:06:24,530 --> 00:06:29,120
gets it is high, you need
to say it three times.

118
00:06:29,120 --> 00:06:31,250
So cycling is one of
the things that I always

119
00:06:31,250 --> 00:06:33,500
think about when I give a talk.

120
00:06:33,500 --> 00:06:38,190
Another thing I think about
is, in explaining my idea,

121
00:06:38,190 --> 00:06:45,590
I want to build a fence around
it so that it's not confused

122
00:06:45,590 --> 00:06:48,030
with somebody else's idea.

123
00:06:48,030 --> 00:06:50,900
So if you were from Mars,
and I was teaching you

124
00:06:50,900 --> 00:06:53,510
about what an arch is,
I might say to you,

125
00:06:53,510 --> 00:06:55,333
well, that's an arch.

126
00:06:55,333 --> 00:06:57,500
And that's not to be confused
with some other things

127
00:06:57,500 --> 00:06:58,970
that other people
might think is--

128
00:06:58,970 --> 00:06:59,950
this is not an arch.

129
00:07:02,600 --> 00:07:04,010
That's not an arch.

130
00:07:04,010 --> 00:07:05,780
I'm building a
fence around my idea

131
00:07:05,780 --> 00:07:09,020
so that it can be distinguished
from somebody else's idea.

132
00:07:09,020 --> 00:07:12,330
So in a more technical
sense, I might say, well,

133
00:07:12,330 --> 00:07:13,990
my algorithm might similar--

134
00:07:13,990 --> 00:07:15,740
might seem similar
to Jones's algorithm,

135
00:07:15,740 --> 00:07:18,285
except his is exponential,
and mine's linear.

136
00:07:18,285 --> 00:07:19,910
That's putting a
fence around your idea

137
00:07:19,910 --> 00:07:22,998
so that people can not
be confused about how it

138
00:07:22,998 --> 00:07:24,290
might relate to something else.

139
00:07:26,960 --> 00:07:30,230
The third thing on
this list of samples

140
00:07:30,230 --> 00:07:33,440
is the idea of
verbal punctuation.

141
00:07:44,780 --> 00:07:48,050
And the idea here is
that, because people

142
00:07:48,050 --> 00:07:52,730
will occasionally fog out and
need to get back on the bus,

143
00:07:52,730 --> 00:07:56,210
you need to provide some
landmark places where

144
00:07:56,210 --> 00:08:00,260
you're announcing that it's
a good time to get back on.

145
00:08:00,260 --> 00:08:02,480
So I might, in this
talk, say something

146
00:08:02,480 --> 00:08:05,840
about this being my outline.

147
00:08:05,840 --> 00:08:09,030
The first thing we're going to
do is talk about how to start.

148
00:08:09,030 --> 00:08:11,480
Then we're going to deal
with these four samples,

149
00:08:11,480 --> 00:08:13,880
and among these
four samples, I've

150
00:08:13,880 --> 00:08:16,550
talked about the first
idea-- that's cycling.

151
00:08:16,550 --> 00:08:18,200
The second idea,
building-- and now,

152
00:08:18,200 --> 00:08:22,040
the third idea is
verbal punctuation.

153
00:08:22,040 --> 00:08:24,770
So I'm enumerating
and providing numbers.

154
00:08:24,770 --> 00:08:28,040
I'm giving you a sense that
there's a seam in the talk,

155
00:08:28,040 --> 00:08:29,360
and you can get back on.

156
00:08:33,299 --> 00:08:38,130
So now, we're on a roll,
and since we're on a roll,

157
00:08:38,130 --> 00:08:42,270
can you guess what fourth
idea might be here--

158
00:08:42,270 --> 00:08:44,480
an idea that helps people
get back on the bus?

159
00:08:48,800 --> 00:08:51,405
AUDIENCE: Ask a question.

160
00:08:51,405 --> 00:08:52,280
PATRICK WINSTON: Yes?

161
00:08:52,280 --> 00:08:53,322
AUDIENCE: Ask a question.

162
00:08:53,322 --> 00:08:53,860
[INAUDIBLE]

163
00:08:53,860 --> 00:08:55,570
PATRICK WINSTON:
Ask a question, yes.

164
00:08:55,570 --> 00:08:58,220
Thank you.

165
00:08:58,220 --> 00:08:59,150
So ask a question.

166
00:09:05,420 --> 00:09:09,030
And so I will ask a question--
how much dead air can there be?

167
00:09:09,030 --> 00:09:11,010
How long can I pause?

168
00:09:11,010 --> 00:09:13,080
I counted seven seconds.

169
00:09:13,080 --> 00:09:16,300
It seemed like an eternity to
me to wait and not say anything

170
00:09:16,300 --> 00:09:19,740
for seven seconds, but that's
the standard amount of time

171
00:09:19,740 --> 00:09:20,910
you can wait for an answer.

172
00:09:20,910 --> 00:09:23,700
And of course, the question
has to be carefully chosen.

173
00:09:23,700 --> 00:09:25,620
It can't be too obvious
because then people

174
00:09:25,620 --> 00:09:28,410
will be embarrassed to say
it, but the answers can't

175
00:09:28,410 --> 00:09:32,580
be too hard because then nobody
will have anything to say.

176
00:09:32,580 --> 00:09:34,800
So here are some
sample heuristics

177
00:09:34,800 --> 00:09:37,110
you can put in
your armamentarium

178
00:09:37,110 --> 00:09:43,460
and build up your repertoire
of ideas about presentation.

179
00:09:43,460 --> 00:09:47,340
And now, if this
persuades you that there

180
00:09:47,340 --> 00:09:52,620
is something to know,
that there is knowledge,

181
00:09:52,620 --> 00:09:55,500
then I've already succeeded
because what I want to convince

182
00:09:55,500 --> 00:09:59,160
you of, is if you watch the
speakers you admire and feel

183
00:09:59,160 --> 00:10:01,140
are effective,
and ask yourselves

184
00:10:01,140 --> 00:10:03,030
why they're successful,
then you can

185
00:10:03,030 --> 00:10:06,000
build up your own
personal repertoire

186
00:10:06,000 --> 00:10:08,010
and develop your
own personal style.

187
00:10:08,010 --> 00:10:10,830
And that's my
fundamental objective,

188
00:10:10,830 --> 00:10:12,900
and the rest of this
talk is about some

189
00:10:12,900 --> 00:10:15,180
of the things that are
in my armamentarium

190
00:10:15,180 --> 00:10:17,550
that I think are effective.

191
00:10:17,550 --> 00:10:20,280
So next thing on
our agenda, as we

192
00:10:20,280 --> 00:10:22,830
start to discuss
these other things,

193
00:10:22,830 --> 00:10:25,170
is a discussion
of time and place.

194
00:10:25,170 --> 00:10:27,420
So what do you think is a
good time to have a lecture?

195
00:10:30,612 --> 00:10:31,660
11 AM?

196
00:10:31,660 --> 00:10:32,160
Yeah.

197
00:10:38,280 --> 00:10:44,670
And the reason is most people
at MIT are awake by then,

198
00:10:44,670 --> 00:10:48,170
and hardly anyone has
gone back to sleep.

199
00:10:48,170 --> 00:10:50,100
It's not right after a meal.

200
00:10:50,100 --> 00:10:52,560
People aren't fatigued
from this or that.

201
00:10:52,560 --> 00:10:54,970
It's a great time
to have a lecture.

202
00:10:54,970 --> 00:10:56,970
So that brings me, next,
to the question of what

203
00:10:56,970 --> 00:10:58,546
about the place?

204
00:10:58,546 --> 00:11:01,440
And the most important
thing about the place

205
00:11:01,440 --> 00:11:02,830
is that it be well lit.

206
00:11:09,320 --> 00:11:11,990
This room is well lit.

207
00:11:11,990 --> 00:11:16,370
Problem with other kinds
of rooms is that we humans,

208
00:11:16,370 --> 00:11:18,740
whenever the lights
go down, or whenever

209
00:11:18,740 --> 00:11:21,410
the room is dimly
lighted, it signals

210
00:11:21,410 --> 00:11:23,350
that we should go to sleep.

211
00:11:23,350 --> 00:11:25,100
So whenever I go
somewhere to give a talk,

212
00:11:25,100 --> 00:11:27,440
even today, the
first thing I do when

213
00:11:27,440 --> 00:11:29,220
I speak to the
audio-visual people

214
00:11:29,220 --> 00:11:31,240
is say, keep the lights full up.

215
00:11:34,400 --> 00:11:39,000
Oh, they might reply, people
will see the slides better

216
00:11:39,000 --> 00:11:42,610
if we turn the lights
off, and then I reply,

217
00:11:42,610 --> 00:11:44,985
it's extremely hard to see
slides through closed eyelids.

218
00:11:50,350 --> 00:11:53,070
What else can you
say about the place?

219
00:11:53,070 --> 00:11:58,710
Well, the place
should be cased, and I

220
00:11:58,710 --> 00:12:01,680
mean that in a colloquial
sense, of like if you're

221
00:12:01,680 --> 00:12:05,520
robbing a bank, you
would go to the bank

222
00:12:05,520 --> 00:12:08,400
some occasions before
to see what it's like,

223
00:12:08,400 --> 00:12:13,170
so there are no surprises
when you do your robbery.

224
00:12:13,170 --> 00:12:15,750
So whenever I go
somewhere to speak,

225
00:12:15,750 --> 00:12:18,600
the first thing I ask my host
to do is to take me to the place

226
00:12:18,600 --> 00:12:22,320
where I'll be speaking so that
if there are any weirdnesses,

227
00:12:22,320 --> 00:12:26,520
I'll be able to deal with it.

228
00:12:26,520 --> 00:12:29,010
Sometimes, it might
require some intervention.

229
00:12:29,010 --> 00:12:31,800
Sometimes, it just might
require me to understand

230
00:12:31,800 --> 00:12:33,660
what the challenges are.

231
00:12:33,660 --> 00:12:36,960
So when I came
here this morning,

232
00:12:36,960 --> 00:12:39,203
I did what I typically do.

233
00:12:39,203 --> 00:12:40,620
I imagined that
all the seats were

234
00:12:40,620 --> 00:12:44,340
filled with disinterested
farm animals,

235
00:12:44,340 --> 00:12:46,830
and that way, I knew that,
no matter how bad it was,

236
00:12:46,830 --> 00:12:48,120
it wouldn't be as bad as that.

237
00:12:51,460 --> 00:12:55,540
So finally, it should
be reasonably--

238
00:12:55,540 --> 00:12:57,460
it should be
reasonably populated.

239
00:13:03,950 --> 00:13:07,340
It should be the case that--

240
00:13:07,340 --> 00:13:09,370
if there were 10
people in this hall,

241
00:13:09,370 --> 00:13:11,050
everyone would be
wondering, what's

242
00:13:11,050 --> 00:13:14,230
going on that's so much more
interesting that nobody's here.

243
00:13:14,230 --> 00:13:17,055
So you want to get a
right sized place that's--

244
00:13:17,055 --> 00:13:18,430
doesn't have to
be packed, but it

245
00:13:18,430 --> 00:13:20,950
has to be more than half full.

246
00:13:20,950 --> 00:13:24,550
So those are some thoughts
about a time and place.

247
00:13:24,550 --> 00:13:27,640
Next thing I want to
talk about is a subject

248
00:13:27,640 --> 00:13:30,385
of boards and props and slides.

249
00:13:38,230 --> 00:13:41,230
Well, these are the
tools of the trade.

250
00:13:41,230 --> 00:13:49,420
I believe that this is the
right tool for speaking

251
00:13:49,420 --> 00:13:52,330
when your purpose is informing.

252
00:13:52,330 --> 00:13:56,290
The slides are good when
your purpose is exposing,

253
00:13:56,290 --> 00:14:00,610
but this is what I use when I'm
informing, teaching, lecturing,

254
00:14:00,610 --> 00:14:02,440
and there's several
reasons why I use it.

255
00:14:02,440 --> 00:14:04,480
For one thing, when
you use the board,

256
00:14:04,480 --> 00:14:05,605
you have a graphic quality.

257
00:14:10,490 --> 00:14:14,440
It's the case that, when
you have a board, then

258
00:14:14,440 --> 00:14:19,990
you can easily exploit the
fact that you can use graphics

259
00:14:19,990 --> 00:14:22,668
in your presentation.

260
00:14:22,668 --> 00:14:25,210
So that's the graphic quality
that I like, and the next thing

261
00:14:25,210 --> 00:14:26,755
I like is the speed property.

262
00:14:30,310 --> 00:14:32,440
The speed with which you
write on the blackboard

263
00:14:32,440 --> 00:14:36,220
is approximately the speed at
which people can absorb ideas.

264
00:14:36,220 --> 00:14:39,040
If you go flipping
through a bunch of slides,

265
00:14:39,040 --> 00:14:40,210
nobody can go that fast.

266
00:14:44,030 --> 00:14:47,260
Finally, one great
property of a board

267
00:14:47,260 --> 00:14:48,750
is that it can be a target.

268
00:14:53,830 --> 00:14:59,200
Many people who are
novices at speaking

269
00:14:59,200 --> 00:15:04,000
find themselves suddenly
aware of their hands.

270
00:15:04,000 --> 00:15:06,370
It's as if their hands were
private parts that shouldn't

271
00:15:06,370 --> 00:15:10,820
be exposed in public, so right
away, they go into the pockets,

272
00:15:10,820 --> 00:15:14,290
and this is considered insulting
in some parts of the world.

273
00:15:14,290 --> 00:15:16,120
Or alternatively,
maybe the hands

274
00:15:16,120 --> 00:15:19,690
will go in back like this.

275
00:15:19,690 --> 00:15:27,820
I was once in a convent
in Serbia, and my host--

276
00:15:27,820 --> 00:15:30,370
as soon as we entered,
a nun came up to us

277
00:15:30,370 --> 00:15:32,260
and offered us a refreshment.

278
00:15:32,260 --> 00:15:33,980
And I was about to
say, no, thank you,

279
00:15:33,980 --> 00:15:38,410
and he said, eat
that stuff or die.

280
00:15:38,410 --> 00:15:42,370
It's a question of local
custom and politeness.

281
00:15:42,370 --> 00:15:46,150
But then before
anything happened there,

282
00:15:46,150 --> 00:15:49,360
the nun pulled my
hands out like this

283
00:15:49,360 --> 00:15:52,330
because it was extraordinarily
insulting in that culture

284
00:15:52,330 --> 00:15:54,880
to have your hands
behind your back.

285
00:15:54,880 --> 00:15:56,080
So why is that?

286
00:15:56,080 --> 00:15:59,210
Well, it's usually
supposed that that's

287
00:15:59,210 --> 00:16:02,170
that it has to do with whether
you're concealing a weapon.

288
00:16:02,170 --> 00:16:05,860
So if your hands are in your
pockets or behind your back,

289
00:16:05,860 --> 00:16:09,810
then it looks like you
might have a weapon,

290
00:16:09,810 --> 00:16:12,010
and that's what I
mean by the virtue--

291
00:16:12,010 --> 00:16:13,583
one of these virtues
of the board.

292
00:16:13,583 --> 00:16:15,500
Now, you have something
to do with your hands.

293
00:16:15,500 --> 00:16:17,050
You can point out the stuff.

294
00:16:20,150 --> 00:16:22,310
I was once watching Seymour
Papert give a lecture,

295
00:16:22,310 --> 00:16:24,110
and I thought it was terrific.

296
00:16:24,110 --> 00:16:25,850
So I went a second time--

297
00:16:25,850 --> 00:16:28,170
first time to absorb
the content, second time

298
00:16:28,170 --> 00:16:29,880
to note style.

299
00:16:29,880 --> 00:16:32,090
And what I discovered is
that Papert was constantly

300
00:16:32,090 --> 00:16:33,568
pointing at the board.

301
00:16:33,568 --> 00:16:35,360
And then I thought
about it a little while,

302
00:16:35,360 --> 00:16:38,638
and I noted that none of
the stuff he was pointing to

303
00:16:38,638 --> 00:16:40,430
had anything to do with
what he was saying.

304
00:16:43,190 --> 00:16:46,371
Nevertheless, it was
an effective technique.

305
00:16:49,200 --> 00:16:52,070
So that's just a little
bit about the virtue

306
00:16:52,070 --> 00:16:53,260
of blackboards.

307
00:16:53,260 --> 00:16:56,850
Now, I want to talk about props.

308
00:16:56,850 --> 00:16:59,820
The custodians of
knowledge about props

309
00:16:59,820 --> 00:17:01,890
are the playwrights.

310
00:17:01,890 --> 00:17:05,730
Many decades ago, I saw
a play by Henrik Ibsen.

311
00:17:05,730 --> 00:17:07,589
It was Hedda Gabler.

312
00:17:07,589 --> 00:17:11,670
I remember vaguely that it was
about a woman in an unhappy

313
00:17:11,670 --> 00:17:16,349
marriage, and her husband was in
competition for an academic job

314
00:17:16,349 --> 00:17:17,599
with somebody else.

315
00:17:17,599 --> 00:17:20,910
And he was going to lose
partly because he was boring,

316
00:17:20,910 --> 00:17:23,220
and partly because the
competitor had just

317
00:17:23,220 --> 00:17:25,650
written a magnificent book.

318
00:17:25,650 --> 00:17:27,150
By the way, this
is back in the days

319
00:17:27,150 --> 00:17:30,710
before there were copying
machines and computers.

320
00:17:30,710 --> 00:17:43,590
Anyhow, as the play opens,
there's a potbellied stove,

321
00:17:43,590 --> 00:17:47,000
and in the beginning
of the play,

322
00:17:47,000 --> 00:17:50,130
the potbelly stove,
with its open door,

323
00:17:50,130 --> 00:17:54,330
just has some slightly
glowing embers.

324
00:17:54,330 --> 00:17:57,000
But the potbellied
stove is always there,

325
00:17:57,000 --> 00:17:59,640
and as tension
mounts in the play,

326
00:17:59,640 --> 00:18:01,170
and you see this
manuscript, this

327
00:18:01,170 --> 00:18:06,210
prop that Ibsen so artfully
used, you just know

328
00:18:06,210 --> 00:18:08,520
that something's
going to happen,

329
00:18:08,520 --> 00:18:13,890
because as the play goes on,
the fire gets bigger and hotter

330
00:18:13,890 --> 00:18:15,750
and finally all
consuming, and you just

331
00:18:15,750 --> 00:18:21,210
know that that manuscript is
going to go into that fire.

332
00:18:21,210 --> 00:18:26,070
This memorable thing is what
I remember about the play.

333
00:18:26,070 --> 00:18:29,590
So playwrights have got
this all figured out,

334
00:18:29,590 --> 00:18:31,860
but on the other hand,
they're not the only people

335
00:18:31,860 --> 00:18:34,380
who can use props.

336
00:18:34,380 --> 00:18:38,010
Here's an example of
the use of a prop, also

337
00:18:38,010 --> 00:18:39,750
due to Seymour Papert.

338
00:18:39,750 --> 00:18:42,540
He was talking about how
it's important to look

339
00:18:42,540 --> 00:18:45,260
at the problem in the
right way, and here's

340
00:18:45,260 --> 00:18:50,015
an example that not only teaches
that, but makes it possible

341
00:18:50,015 --> 00:18:51,390
for you to embarrass
your friends

342
00:18:51,390 --> 00:18:53,190
in mechanical engineering.

343
00:18:53,190 --> 00:18:55,110
So here's what you do.

344
00:18:55,110 --> 00:18:58,140
Take a bicycle wheel, and
you start it spinning.

345
00:18:58,140 --> 00:19:02,010
And then you put some
torque on the axle,

346
00:19:02,010 --> 00:19:06,240
or equivalently, you
blow on the edge.

347
00:19:06,240 --> 00:19:09,570
And the issue is,
does it go that way,

348
00:19:09,570 --> 00:19:12,790
or does it go that way?

349
00:19:12,790 --> 00:19:16,590
Now, the mechanical engineers
will immediately say, oh, yes,

350
00:19:16,590 --> 00:19:18,250
I see-- right hand screw wheel.

351
00:19:18,250 --> 00:19:20,970
And they'll put their
fingers in this position,

352
00:19:20,970 --> 00:19:24,090
but forget exactly how
to align their fingers

353
00:19:24,090 --> 00:19:27,870
with the various
aspects of the problem.

354
00:19:27,870 --> 00:19:30,420
And so it's usually the
case that they get it right

355
00:19:30,420 --> 00:19:34,260
with about a 50% probability.

356
00:19:34,260 --> 00:19:38,670
So their very fancy
education gets them up

357
00:19:38,670 --> 00:19:43,060
to the point where they're
equivalent to flipping a coin.

358
00:19:43,060 --> 00:19:45,720
But it doesn't
have to be that way

359
00:19:45,720 --> 00:19:49,030
because you can think about the
problem a little differently.

360
00:19:49,030 --> 00:19:50,980
So here's what you do.

361
00:19:50,980 --> 00:19:54,080
You take some duct
tape, and you put it

362
00:19:54,080 --> 00:19:57,450
around the part of
the wheel like that.

363
00:19:57,450 --> 00:20:00,250
And now, you start to think
about, not the whole wheel,

364
00:20:00,250 --> 00:20:03,920
but just a little piece that's
underneath the duct tape.

365
00:20:03,920 --> 00:20:06,720
So here, that piece comes
rolling over the top,

366
00:20:06,720 --> 00:20:09,455
and at this point, you blow
on it with a puff of air.

367
00:20:09,455 --> 00:20:11,080
Forgetting about the
rest of the wheel,

368
00:20:11,080 --> 00:20:12,747
what happens to that
little piece that's

369
00:20:12,747 --> 00:20:13,655
under the duct tape?

370
00:20:13,655 --> 00:20:15,280
It must want to go
that way because you

371
00:20:15,280 --> 00:20:17,040
banged on it like that.

372
00:20:17,040 --> 00:20:19,420
It's already going
down like that.

373
00:20:19,420 --> 00:20:20,910
And what about the next piece?

374
00:20:20,910 --> 00:20:21,830
Same thing.

375
00:20:21,830 --> 00:20:22,330
Next piece?

376
00:20:22,330 --> 00:20:23,200
Same thing.

377
00:20:23,200 --> 00:20:25,620
So the only thing
that can happen

378
00:20:25,620 --> 00:20:28,090
is that the wheel
goes over like that.

379
00:20:28,090 --> 00:20:30,370
And so now, you'll
never wonder again

380
00:20:30,370 --> 00:20:33,000
because you're thinking about
the problem in the right way,

381
00:20:33,000 --> 00:20:36,000
and it's demonstrated
by the use of a prop.

382
00:20:38,703 --> 00:20:40,120
You can try this
after we're done.

383
00:20:44,690 --> 00:20:47,980
Another example I
like to remember

384
00:20:47,980 --> 00:20:50,880
is one from when
I was taking 8.01.

385
00:20:50,880 --> 00:20:53,960
Alan Lazarus was the
instructor at the time,

386
00:20:53,960 --> 00:20:57,220
and he was talking about the
conservation of energy, kinetic

387
00:20:57,220 --> 00:20:59,450
and potential.

388
00:20:59,450 --> 00:21:04,420
And there was a long wire in
a ceiling in 26-100 attached

389
00:21:04,420 --> 00:21:07,800
to a much bigger
steel ball, but one--

390
00:21:07,800 --> 00:21:09,270
not one like this.

391
00:21:09,270 --> 00:21:15,100
And Lazarus took the ball up
against the wall like this.

392
00:21:15,100 --> 00:21:19,210
He put his head flat against
the wall to steady himself,

393
00:21:19,210 --> 00:21:22,390
and then he let go,
and the pendulum

394
00:21:22,390 --> 00:21:25,750
takes many seconds
to go over and back,

395
00:21:25,750 --> 00:21:31,300
and then gently
kisses Lazarus's nose.

396
00:21:31,300 --> 00:21:33,700
And so you have many
seconds to think,

397
00:21:33,700 --> 00:21:36,550
this guy really believes in
the conservation of energy.

398
00:21:43,270 --> 00:21:46,600
Do not try this at home.

399
00:21:46,600 --> 00:21:50,180
The problem is that, the
first time you do this,

400
00:21:50,180 --> 00:21:51,820
you may not just let go.

401
00:21:51,820 --> 00:21:54,640
There's a natural
human tendency to push.

402
00:22:01,510 --> 00:22:04,510
So that's a little bit
on a subject of props.

403
00:22:24,612 --> 00:22:25,320
It's interesting.

404
00:22:25,320 --> 00:22:29,810
Whenever surveys are taken,
students always say more chalk,

405
00:22:29,810 --> 00:22:31,800
less PowerPoint.

406
00:22:31,800 --> 00:22:34,230
And why would that be?

407
00:22:34,230 --> 00:22:37,080
Props are also very effective.

408
00:22:37,080 --> 00:22:39,420
Why would that be?

409
00:22:39,420 --> 00:22:42,870
I'll give you my lunatic
fringe view on this.

410
00:22:42,870 --> 00:23:02,350
It has to do with what I would
call empathetic mirroring.

411
00:23:02,350 --> 00:23:05,020
When you're sitting up there
watching me write on the board,

412
00:23:05,020 --> 00:23:08,950
all those little mirror neurons
in your head, I believe,

413
00:23:08,950 --> 00:23:11,560
become actuated, and
you can feel yourself

414
00:23:11,560 --> 00:23:13,060
writing on the blackboard.

415
00:23:13,060 --> 00:23:18,560
And even more so, when I talk
about this steel ball going

416
00:23:18,560 --> 00:23:22,270
that way and this way,
you can feel the ball

417
00:23:22,270 --> 00:23:26,380
as if you were me, and you
can't do that with a slide.

418
00:23:26,380 --> 00:23:28,190
You can't do it with a picture.

419
00:23:28,190 --> 00:23:31,820
You need to see it in
the physical world.

420
00:23:31,820 --> 00:23:34,090
That's why I think that--

421
00:23:34,090 --> 00:23:35,803
oh, yes, of course, it's--

422
00:23:35,803 --> 00:23:37,470
there are speed
questions involved, too,

423
00:23:37,470 --> 00:23:39,640
that have to be separated out.

424
00:23:39,640 --> 00:23:41,140
But I think the
empathetic mirroring

425
00:23:41,140 --> 00:23:47,890
is why props and the use of a
blackboard are so effective.

426
00:23:47,890 --> 00:23:53,990
Well, let's see-- oh, yes,
there is one more thing

427
00:23:53,990 --> 00:23:57,100
by way of the
tools, and that has

428
00:23:57,100 --> 00:23:59,500
to do with the use of slides.

429
00:24:02,260 --> 00:24:05,140
I repeat, I think they're
for exposing ideas,

430
00:24:05,140 --> 00:24:08,440
not for teaching ideas, but
that's what we do in a job talk

431
00:24:08,440 --> 00:24:11,230
or conference talk--

432
00:24:11,230 --> 00:24:12,130
expose ideas.

433
00:24:12,130 --> 00:24:14,090
We don't teach them.

434
00:24:14,090 --> 00:24:16,840
So let me tell you a little
bit about my views on that.

435
00:24:20,510 --> 00:24:29,060
I remember, once, I was in
Terminal A at Logan Airport.

436
00:24:29,060 --> 00:24:33,240
I'd just come back from a
really miserable conference,

437
00:24:33,240 --> 00:24:35,155
and the flight was
really horrible.

438
00:24:35,155 --> 00:24:37,530
It was one of those that feels
like an unbalanced washing

439
00:24:37,530 --> 00:24:39,095
machine.

440
00:24:39,095 --> 00:24:40,470
And for the only
time in my life,

441
00:24:40,470 --> 00:24:46,620
I decided to stop on my way to
my car and have a cup of coffee

442
00:24:46,620 --> 00:24:48,560
and relax a little bit.

443
00:24:48,560 --> 00:24:51,450
And as I was there for a few
minutes, someone came up to me

444
00:24:51,450 --> 00:24:54,690
and said, are you
Professor Winston?

445
00:24:54,690 --> 00:24:57,808
I think so, I said.

446
00:24:57,808 --> 00:24:58,350
I don't know.

447
00:24:58,350 --> 00:25:00,570
I guess I was
trying to be funny.

448
00:25:00,570 --> 00:25:02,700
In any event, he said,
I'm on my way to Europe

449
00:25:02,700 --> 00:25:04,320
to give a job talk.

450
00:25:04,320 --> 00:25:07,440
Would you mine
critiquing my slides?

451
00:25:07,440 --> 00:25:08,610
Not at all, I said.

452
00:25:08,610 --> 00:25:12,750
You have too many, and
they have too many words.

453
00:25:12,750 --> 00:25:15,280
How did you know, he
said, thinking perhaps I

454
00:25:15,280 --> 00:25:17,790
had seen a talk of his before.

455
00:25:17,790 --> 00:25:19,800
I hadn't.

456
00:25:19,800 --> 00:25:22,770
My reply was, because
it's always true.

457
00:25:22,770 --> 00:25:26,310
There are always too many
slides, always too many words.

458
00:25:26,310 --> 00:25:29,725
So let me show you some extreme
examples of how not to use

459
00:25:29,725 --> 00:25:30,225
slides.

460
00:25:34,110 --> 00:25:42,270
Well, for this demonstration,
I need to be way over here

461
00:25:42,270 --> 00:25:48,510
and when I get over here, then
I can start to say things like,

462
00:25:48,510 --> 00:25:51,342
one of the things you shouldn't
do is read your transparencies.

463
00:25:51,342 --> 00:25:53,050
People in your audience
know how to read,

464
00:25:53,050 --> 00:25:55,410
and reading will
just annoy them.

465
00:25:55,410 --> 00:25:56,940
Also, you should
be sure that you

466
00:25:56,940 --> 00:25:59,700
have only a few words
on each transparency,

467
00:25:59,700 --> 00:26:02,160
and that the words
are easy to read.

468
00:26:02,160 --> 00:26:05,310
And I hope I'm driving
you crazy because I'm

469
00:26:05,310 --> 00:26:08,550
committing all kinds of
crimes, the first of which

470
00:26:08,550 --> 00:26:12,540
is that there are too
many words on the slide.

471
00:26:12,540 --> 00:26:14,630
Second of which is,
I'm way over there,

472
00:26:14,630 --> 00:26:16,560
and the slide's way over there.

473
00:26:16,560 --> 00:26:20,070
So you get into this
tennis match feeling

474
00:26:20,070 --> 00:26:23,700
of shifting back and
forth between the slide

475
00:26:23,700 --> 00:26:24,600
and the speaker.

476
00:26:24,600 --> 00:26:26,790
You want the slides to
be condiments to what

477
00:26:26,790 --> 00:26:29,730
you're saying, not the main
event or the opposite way

478
00:26:29,730 --> 00:26:31,150
around.

479
00:26:31,150 --> 00:26:33,420
So how can we fix this?

480
00:26:33,420 --> 00:26:37,860
Step number one is to get
rid of the background junk.

481
00:26:37,860 --> 00:26:40,560
That's always distraction.

482
00:26:40,560 --> 00:26:43,590
Step number two is to
get rid of the words.

483
00:26:43,590 --> 00:26:46,320
When I reduced the
words to these,

484
00:26:46,320 --> 00:26:50,210
then everything I
read a previous time,

485
00:26:50,210 --> 00:26:53,270
I'm not licensed to say,
because it's not on the slide.

486
00:26:53,270 --> 00:26:55,090
I'm not reading
my slides anymore,

487
00:26:55,090 --> 00:26:58,370
but I'm saying what was
written on the slides

488
00:26:58,370 --> 00:27:02,010
in a previous example.

489
00:27:02,010 --> 00:27:05,310
So what else can we
do to simplify this?

490
00:27:05,310 --> 00:27:07,760
Well, we can get
rid of the logos.

491
00:27:07,760 --> 00:27:10,750
We don't need them.

492
00:27:10,750 --> 00:27:13,030
Simplification.

493
00:27:13,030 --> 00:27:15,400
What else can we do?

494
00:27:15,400 --> 00:27:17,410
Get rid of the title.

495
00:27:17,410 --> 00:27:20,240
Now, I want to talk to you
about some rules for slide

496
00:27:20,240 --> 00:27:21,250
preparation.

497
00:27:21,250 --> 00:27:22,840
I'm telling you the title.

498
00:27:22,840 --> 00:27:24,880
It doesn't have to be up there.

499
00:27:24,880 --> 00:27:27,370
By reducing the number
of words on the slide,

500
00:27:27,370 --> 00:27:29,350
I'm allowing you to pay
more attention to me

501
00:27:29,350 --> 00:27:32,295
and less to what's
written on the slide.

502
00:27:32,295 --> 00:27:34,420
I mentioned it before--
we-- have only one language

503
00:27:34,420 --> 00:27:38,140
processor, and we can
either use it to read stuff

504
00:27:38,140 --> 00:27:40,170
or to listen to the speaker.

505
00:27:40,170 --> 00:27:43,240
And so if we have too
many words on the slide,

506
00:27:43,240 --> 00:27:45,700
it forces people in the
audience to read this stuff

507
00:27:45,700 --> 00:27:46,330
and not listen.

508
00:27:49,270 --> 00:27:52,570
A student of mine did an
experiment a few years ago.

509
00:27:52,570 --> 00:28:00,130
He taught some students some
web-based programming ideas.

510
00:28:00,130 --> 00:28:04,580
Half the information was on
slides, he said the other half,

511
00:28:04,580 --> 00:28:07,630
and then for a control
group, he reversed it.

512
00:28:07,630 --> 00:28:10,960
And the question was,
what did the subjects--

513
00:28:10,960 --> 00:28:13,240
that is to say, freshmen
at his fraternity--

514
00:28:13,240 --> 00:28:18,040
what did the subjects
remember best, what he said,

515
00:28:18,040 --> 00:28:21,260
or what they read on the slide?

516
00:28:21,260 --> 00:28:26,432
And the answer is, what
they read on the slide.

517
00:28:26,432 --> 00:28:28,390
When their slides have
a lot of material on it,

518
00:28:28,390 --> 00:28:30,110
they don't pay attention
to the speaker.

519
00:28:30,110 --> 00:28:33,190
In fact, in the after action
report, one of the subjects

520
00:28:33,190 --> 00:28:36,100
said, I wish you
hadn't talked so much.

521
00:28:36,100 --> 00:28:37,060
It was distracting.

522
00:28:41,000 --> 00:28:44,120
Well, the last item
is eliminate clutter.

523
00:28:44,120 --> 00:28:46,230
Here's some clutter.

524
00:28:46,230 --> 00:28:49,710
No reason even
for those bullets.

525
00:28:49,710 --> 00:28:51,780
So the too many words
problem is a consequence

526
00:28:51,780 --> 00:28:55,410
of a crime Microsoft
has committed

527
00:28:55,410 --> 00:28:59,350
by allowing you to use
fonts that are too small.

528
00:28:59,350 --> 00:29:03,570
So you should all have
a sample slide like this

529
00:29:03,570 --> 00:29:06,240
that you can use to determine
what the minimum font

530
00:29:06,240 --> 00:29:09,480
size is that's easily legible.

531
00:29:09,480 --> 00:29:14,745
[INAUDIBLE],, what do
you think of those?

532
00:29:14,745 --> 00:29:16,212
AUDIENCE: Which size is right?

533
00:29:16,212 --> 00:29:17,190
PATRICK WINSTON: What's that?

534
00:29:17,190 --> 00:29:19,023
AUDIENCE: Did you ask
me what size is right?

535
00:29:19,023 --> 00:29:21,560
PATRICK WINSTON:
Yeah, minimum, maybe.

536
00:29:21,560 --> 00:29:22,770
AUDIENCE: 40 or 50.

537
00:29:22,770 --> 00:29:24,990
PATRICK WINSTON: Yeah,
he says 40 or 50.

538
00:29:24,990 --> 00:29:27,050
I think that's about right.

539
00:29:27,050 --> 00:29:29,730
35 is beginning to get
too small, not necessarily

540
00:29:29,730 --> 00:29:32,917
because you can't read it,
but because you're probably

541
00:29:32,917 --> 00:29:34,750
using it to get too
many words on the slide.

542
00:29:37,380 --> 00:29:39,940
What other crimes do we have?

543
00:29:39,940 --> 00:29:43,755
Well, we have the
laser pointer crime.

544
00:29:43,755 --> 00:29:46,890
And for that-- in
the old days, when

545
00:29:46,890 --> 00:29:49,120
we didn't have laser
pointers, we used wooden ones,

546
00:29:49,120 --> 00:29:52,590
and people would go waving
these things around.

547
00:29:52,590 --> 00:29:59,450
And pretty soon it became almost
like a baton twirling contest,

548
00:29:59,450 --> 00:30:02,400
so here's what I
recommended in the old days

549
00:30:02,400 --> 00:30:04,080
for dealing with
this kind of pointer.

550
00:30:08,970 --> 00:30:10,530
This is an example
of use of a prop.

551
00:30:14,070 --> 00:30:17,090
Jim Glass up there saw this
talk about 20 years ago,

552
00:30:17,090 --> 00:30:19,380
and said, oh, yeah,
I remember that talk.

553
00:30:19,380 --> 00:30:23,250
That's the one where
you broke the pointer.

554
00:30:23,250 --> 00:30:25,200
It's amazing how
props tend to be

555
00:30:25,200 --> 00:30:27,480
the things that are remembered.

556
00:30:27,480 --> 00:30:32,215
Well, now, we don't have
physical pointers anymore.

557
00:30:32,215 --> 00:30:35,880
We've got laser pointers.

558
00:30:35,880 --> 00:30:39,090
It's a wonder more people aren't
driven into epileptic fits

559
00:30:39,090 --> 00:30:41,190
over this sort of stuff.

560
00:30:41,190 --> 00:30:43,848
Well, here's what
tends to happen.

561
00:30:43,848 --> 00:30:44,390
Look at that.

562
00:30:44,390 --> 00:30:49,170
It's a lovely recursive picture,
and I can become part of it

563
00:30:49,170 --> 00:30:52,026
by putting that laser beam
right on the back of my head

564
00:30:52,026 --> 00:30:52,526
up there.

565
00:30:56,100 --> 00:30:58,720
Then what do you see?

566
00:30:58,720 --> 00:31:01,440
You see the back of my head.

567
00:31:01,440 --> 00:31:04,830
I have no eye contact,
no engagement, nothing.

568
00:31:04,830 --> 00:31:08,300
I was sitting with a student
watching a talk one day,

569
00:31:08,300 --> 00:31:11,790
and she said, you know
what, we could all leave,

570
00:31:11,790 --> 00:31:12,710
and he wouldn't know.

571
00:31:16,065 --> 00:31:17,940
So what happens when
you use a laser pointer?

572
00:31:17,940 --> 00:31:21,090
You can't use a laser pointer
without turning your head

573
00:31:21,090 --> 00:31:25,050
and pointing it at something,
and when you do that,

574
00:31:25,050 --> 00:31:27,570
you lose contact
with the audience.

575
00:31:27,570 --> 00:31:29,590
You don't want to do it.

576
00:31:29,590 --> 00:31:32,705
So what do you do if you
need to identify something

577
00:31:32,705 --> 00:31:34,080
in your image,
and you don't want

578
00:31:34,080 --> 00:31:35,280
to point at it with a laser?

579
00:31:35,280 --> 00:31:36,840
This is what you do.

580
00:31:36,840 --> 00:31:38,340
Put a little arrow
on there and say,

581
00:31:38,340 --> 00:31:41,730
now, look at that guy at
the end of arrow number one.

582
00:31:41,730 --> 00:31:43,730
You don't need to have
laser pointer to do that.

583
00:31:48,933 --> 00:31:52,480
The too-heavy
crime-- when people

584
00:31:52,480 --> 00:31:55,370
ask me to review
a presentation, I

585
00:31:55,370 --> 00:31:57,940
ask them to print it out
and lay it out on a table.

586
00:32:01,060 --> 00:32:04,450
When they do that, it's easy
to see whether the talk is

587
00:32:04,450 --> 00:32:08,740
too heavy, too much
text, not enough air, not

588
00:32:08,740 --> 00:32:12,950
enough white space,
not enough imagery.

589
00:32:12,950 --> 00:32:16,540
This is a good example
of such a talk--

590
00:32:16,540 --> 00:32:19,300
way too heavy.

591
00:32:19,300 --> 00:32:22,982
The presenter has taken
advantage of a small font sizes

592
00:32:22,982 --> 00:32:24,690
to get as much on the
slide as he wanted.

593
00:32:27,600 --> 00:32:31,080
Lots of other crimes here,
but the too-heavy-- the fact

594
00:32:31,080 --> 00:32:33,800
that it's too heavy is what
I wanted to illustrate.

595
00:32:33,800 --> 00:32:36,750
So here, by contrast,
another talk--

596
00:32:36,750 --> 00:32:37,890
one I gave a few years ago.

597
00:32:37,890 --> 00:32:41,520
It's not-- it wasn't a
deeply technical talk,

598
00:32:41,520 --> 00:32:46,800
but I show it to you
because there's air in it.

599
00:32:46,800 --> 00:32:49,110
It's mostly pictures of things.

600
00:32:49,110 --> 00:32:52,930
There are three or four
slides that have text on them,

601
00:32:52,930 --> 00:32:56,282
but when I come to those, I give
the audience time to read them.

602
00:32:56,282 --> 00:32:57,990
And they're there
because they might have

603
00:32:57,990 --> 00:33:00,613
some historical significance.

604
00:33:00,613 --> 00:33:02,280
The first slide with
a lot of text on it

605
00:33:02,280 --> 00:33:06,120
is an extraction from the 1957--

606
00:33:06,120 --> 00:33:10,830
from the proposal for the 1957
AI conference at Dartmouth.

607
00:33:10,830 --> 00:33:13,190
Extraordinarily
interesting event,

608
00:33:13,190 --> 00:33:17,790
and that historical
extraction from the proposal

609
00:33:17,790 --> 00:33:20,960
helps drive that point home.

610
00:33:20,960 --> 00:33:23,080
What else have we got here?

611
00:33:23,080 --> 00:33:25,680
Oh, yeah, your vocabulary
word for the day.

612
00:33:25,680 --> 00:33:28,770
This is an hapax legomenon.

613
00:33:28,770 --> 00:33:31,170
What that means is, this
is the kind of slide

614
00:33:31,170 --> 00:33:36,540
you can get away with exactly
once in your presentation.

615
00:33:36,540 --> 00:33:38,910
This is a slide that got
some currency some years

616
00:33:38,910 --> 00:33:43,500
ago because it shows the
complexity of governing

617
00:33:43,500 --> 00:33:49,640
in Afghanistan by showing
how impossibly complex it is.

618
00:33:49,640 --> 00:33:52,480
It's something you in the
audience can't understand,

619
00:33:52,480 --> 00:33:55,920
and that's the point, but
you can't have many of these.

620
00:33:55,920 --> 00:33:58,590
You can have one per work,
one per presentation,

621
00:33:58,590 --> 00:34:00,780
one per paper, one per book.

622
00:34:00,780 --> 00:34:04,620
That's what hapax legomenon is,
and this is an example of it.

623
00:34:06,725 --> 00:34:08,100
Well, I've shown
you some crimes.

624
00:34:08,100 --> 00:34:12,739
So you might be asking, do
these crimes actually occur?

625
00:34:12,739 --> 00:34:16,861
So they do.

626
00:34:16,861 --> 00:34:21,775
[LAUGHTER]

627
00:34:21,775 --> 00:34:23,400
There's the hands in
the pockets crime.

628
00:34:26,540 --> 00:34:32,210
There's a crime and time
and place selection here.

629
00:34:32,210 --> 00:34:35,837
This is how you get
to the Bartos Theater.

630
00:34:35,837 --> 00:34:37,670
First thing you do is
you get on these steps

631
00:34:37,670 --> 00:34:44,120
over at the Media Lab, then you
cross this large open space,

632
00:34:44,120 --> 00:34:45,920
then you turn right
down this corridor.

633
00:34:45,920 --> 00:34:47,201
[LAUGHTER]

634
00:34:47,201 --> 00:34:48,860
At this point,
whenever I go in there,

635
00:34:48,860 --> 00:34:51,962
I wonder if there are torture
implements around the corner.

636
00:34:51,962 --> 00:34:52,462
[LAUGHTER]

637
00:34:52,462 --> 00:34:54,530
And then when you
get in there, you

638
00:34:54,530 --> 00:34:57,680
get into this
dark, gloomy place.

639
00:34:57,680 --> 00:35:00,418
So it's well named when they
call it the Bartos Theater

640
00:35:00,418 --> 00:35:02,460
because it's a place where
you can watch a movie,

641
00:35:02,460 --> 00:35:04,418
but it's not a place
where you can give a talk.

642
00:35:07,330 --> 00:35:10,700
Now, on a subject
of does it happen,

643
00:35:10,700 --> 00:35:14,870
here's a talk I attended
a while back in Stata.

644
00:35:14,870 --> 00:35:19,150
Notice that the speaker is
far away from the slides.

645
00:35:19,150 --> 00:35:22,750
Speaker's using a laser pointer.

646
00:35:22,750 --> 00:35:26,110
And you say to me, well,
what's happening here?

647
00:35:26,110 --> 00:35:29,500
It's, by the way,
the 80th - 80th!

648
00:35:29,500 --> 00:35:31,460
slide of the presentation.

649
00:35:31,460 --> 00:35:33,320
Notice that it extends
with the words,

650
00:35:33,320 --> 00:35:35,810
this is the first of
10 conclusions slides.

651
00:35:35,810 --> 00:35:37,960
[LAUGHTER]

652
00:35:37,960 --> 00:35:41,740
So what's the audience reaction?

653
00:35:41,740 --> 00:35:43,572
That's the sponsor
of the meeting.

654
00:35:43,572 --> 00:35:44,840
[LAUGHTER]

655
00:35:44,840 --> 00:35:47,320
He's reading his email.

656
00:35:47,320 --> 00:35:50,240
This is the co-sponsor
of the meeting.

657
00:35:50,240 --> 00:35:52,012
He's examining the lunch menu.

658
00:35:52,012 --> 00:35:54,552
[LAUGHTER]

659
00:35:54,552 --> 00:35:55,510
What about this person?

660
00:35:58,830 --> 00:36:02,250
This person looks like
he's paying attention,

661
00:36:02,250 --> 00:36:04,863
but just because
it's a still picture.

662
00:36:04,863 --> 00:36:06,780
If you were to see a
video, what you would see

663
00:36:06,780 --> 00:36:10,555
is something like this [YAWNS].

664
00:36:10,555 --> 00:36:14,350
[LAUGHTER]

665
00:36:14,350 --> 00:36:17,960
So yeah, it does happen.

666
00:36:17,960 --> 00:36:23,080
Well, now, that's a
quick review of tools.

667
00:36:23,080 --> 00:36:27,700
Now, I want to talk
about some special cases.

668
00:36:30,300 --> 00:36:32,830
We could talk a little
bit about the informing

669
00:36:32,830 --> 00:36:38,010
or to say another way,
doing what I'm doing now.

670
00:36:38,010 --> 00:36:40,360
But I'll just say a
few words about that.

671
00:36:40,360 --> 00:36:42,340
In that kind of
presentation, you

672
00:36:42,340 --> 00:36:46,870
want to start with a promise
like I did for this hour

673
00:36:46,870 --> 00:36:48,255
that we're going through now.

674
00:36:48,255 --> 00:36:49,630
And then it comes
to the question

675
00:36:49,630 --> 00:36:52,888
of how do you inspire people?

676
00:36:52,888 --> 00:36:54,430
I've given this talk
for a long time,

677
00:36:54,430 --> 00:36:58,300
and a few years ago,
our department chairman

678
00:36:58,300 --> 00:37:02,260
said, would you please give
this talk to a new faculty,

679
00:37:02,260 --> 00:37:06,732
and be sure to emphasize what
it takes to inspire students.

680
00:37:06,732 --> 00:37:09,190
And strangely, I hadn't thought
about that question before.

681
00:37:09,190 --> 00:37:11,290
So I started a survey.

682
00:37:11,290 --> 00:37:14,490
I'd talked to some of my
incoming freshmen advisees,

683
00:37:14,490 --> 00:37:17,290
and I talked to senior
faculty and everything

684
00:37:17,290 --> 00:37:21,520
in between about how
they've been inspired.

685
00:37:21,520 --> 00:37:23,963
What I found from
the incoming freshmen

686
00:37:23,963 --> 00:37:26,380
is that they were inspired by
some high school teacher who

687
00:37:26,380 --> 00:37:29,050
told them they could do it.

688
00:37:29,050 --> 00:37:32,140
What I found in
the senior faculty,

689
00:37:32,140 --> 00:37:35,280
they were inspired
by someone who

690
00:37:35,280 --> 00:37:38,310
helped them see a
problem in a new way.

691
00:37:38,310 --> 00:37:44,050
And what I saw from everyone
is that they were inspired when

692
00:37:44,050 --> 00:37:49,680
someone exhibited passion
about what they were doing,

693
00:37:49,680 --> 00:37:52,920
exhibited passion about
what they were doing.

694
00:37:52,920 --> 00:37:57,720
So that's one way
to be inspiring.

695
00:37:57,720 --> 00:38:03,310
It's easy for me because I
do artificial intelligence.

696
00:38:03,310 --> 00:38:06,030
And how can you
not be interested

697
00:38:06,030 --> 00:38:07,230
in artificial intelligence?

698
00:38:07,230 --> 00:38:08,040
[LAUGHTER]

699
00:38:08,040 --> 00:38:10,842
I mean, if you're not interested
in artificial intelligence,

700
00:38:10,842 --> 00:38:13,050
you're probably not interested
in interesting things.

701
00:38:15,630 --> 00:38:20,640
So when I'm lecturing in my
AI class, it's natural for me

702
00:38:20,640 --> 00:38:23,550
to talk about what
I think is cool

703
00:38:23,550 --> 00:38:27,320
and how exciting
some new idea is.

704
00:38:27,320 --> 00:38:31,080
So that's the kind of
expression of passion

705
00:38:31,080 --> 00:38:34,950
that makes a difference
while informing with respect

706
00:38:34,950 --> 00:38:37,080
to this question of inspiring.

707
00:38:39,960 --> 00:38:43,540
Oh, yeah and of course,
during this promise phase,

708
00:38:43,540 --> 00:38:47,070
you can also express
how cool stuff is.

709
00:38:47,070 --> 00:38:50,670
Let me give you an example of
a lecture that starts this way.

710
00:38:50,670 --> 00:38:53,060
I'm talking about
resource allocation.

711
00:38:53,060 --> 00:38:55,760
It's the same sort of stuff
you would think of when your--

712
00:38:55,760 --> 00:38:57,302
it's the same sort
of ideas you would

713
00:38:57,302 --> 00:39:01,290
need if you're allocating
aircraft to a flight schedule

714
00:39:01,290 --> 00:39:04,000
or trying to schedule a
factory or something like that.

715
00:39:04,000 --> 00:39:06,060
But the example
is putting colors

716
00:39:06,060 --> 00:39:08,760
on the states in
the United States

717
00:39:08,760 --> 00:39:11,380
without any bordering states
having the same color.

718
00:39:11,380 --> 00:39:12,050
So here it goes.

719
00:39:15,390 --> 00:39:18,600
This is what I show at the
beginning of the class.

720
00:39:18,600 --> 00:39:21,520
This is a way of
doing that coloring.

721
00:39:21,520 --> 00:39:25,440
And you might say, well, why
don't we wait till it finishes?

722
00:39:25,440 --> 00:39:27,190
Would you like to do that?

723
00:39:27,190 --> 00:39:29,283
No?

724
00:39:29,283 --> 00:39:31,200
Well, we're not going
to wait till it finishes

725
00:39:31,200 --> 00:39:33,990
because the sun will have
exploded and consumed the earth

726
00:39:33,990 --> 00:39:35,430
before this program finishes.

727
00:39:35,430 --> 00:39:36,960
[LAUGHTER]

728
00:39:36,960 --> 00:39:41,520
But with a slight adjustment
to how the program works, which

729
00:39:41,520 --> 00:39:44,070
I tell my students you will
understand in the next 50

730
00:39:44,070 --> 00:39:45,660
minutes, this is what you get.

731
00:39:49,850 --> 00:39:50,630
Isn't that cool?

732
00:39:53,410 --> 00:39:55,160
You got to be
amazed by stuff that

733
00:39:55,160 --> 00:39:57,830
takes a computation from
longer than the lifetime

734
00:39:57,830 --> 00:40:00,920
of the solar system
into a few seconds.

735
00:40:00,920 --> 00:40:05,757
So that's what I mean by
providing a promise upfront

736
00:40:05,757 --> 00:40:07,340
and expressing some
passion about what

737
00:40:07,340 --> 00:40:10,400
you're talking about.

738
00:40:10,400 --> 00:40:12,080
Well, the last item
in this little block

739
00:40:12,080 --> 00:40:15,110
here is it has to
do with what people

740
00:40:15,110 --> 00:40:16,640
think that they do it at MIT.

741
00:40:16,640 --> 00:40:19,880
You ask faculty what the
most important purpose is,

742
00:40:19,880 --> 00:40:23,480
and they'll say, well, the
most important thing I do

743
00:40:23,480 --> 00:40:26,060
is teach people how to think.

744
00:40:26,060 --> 00:40:29,090
And then you say
oh, that's great.

745
00:40:29,090 --> 00:40:32,030
How do you teach
people how to think?

746
00:40:32,030 --> 00:40:32,730
Blank stare.

747
00:40:35,250 --> 00:40:38,040
No one can quite
respond to that part,

748
00:40:38,040 --> 00:40:40,693
that natural next question.

749
00:40:40,693 --> 00:40:42,360
So how do you teach
people how to think?

750
00:40:42,360 --> 00:40:48,840
Well, I believe that we
are storytelling animals.

751
00:40:48,840 --> 00:40:52,518
And that we start
developing our story,

752
00:40:52,518 --> 00:40:54,060
understanding and
manipulating skills

753
00:40:54,060 --> 00:40:57,810
with fairy tales in
childhood and continue on

754
00:40:57,810 --> 00:41:02,490
through professional schools
like law, business, medicine,

755
00:41:02,490 --> 00:41:03,870
everything.

756
00:41:03,870 --> 00:41:06,690
And we continue doing
that throughout life.

757
00:41:06,690 --> 00:41:09,870
So if that is what
thinking is all about.

758
00:41:09,870 --> 00:41:11,730
And we want to teach
people how to think,

759
00:41:11,730 --> 00:41:14,490
you provide them with the
stories they need to know,

760
00:41:14,490 --> 00:41:17,070
the questions they need to
ask about those stories,

761
00:41:17,070 --> 00:41:19,350
mechanisms for
analyzing those stories,

762
00:41:19,350 --> 00:41:20,890
ways of putting
stories together,

763
00:41:20,890 --> 00:41:24,760
ways of evaluating how
reliable a story is.

764
00:41:24,760 --> 00:41:28,080
And that's what I think you
need to do when you teach people

765
00:41:28,080 --> 00:41:29,770
how to think.

766
00:41:29,770 --> 00:41:31,920
But that's all about education.

767
00:41:31,920 --> 00:41:35,790
And many of you here
are not necessarily

768
00:41:35,790 --> 00:41:39,870
for that, but rather
for this part,

769
00:41:39,870 --> 00:41:43,410
for persuading, which breaks
down into several categories,

770
00:41:43,410 --> 00:41:48,210
oral exams, not shown,
shop talks, getting famous.

771
00:41:48,210 --> 00:41:51,720
I won't say much about oral
exams other than the fact

772
00:41:51,720 --> 00:41:54,930
that they used to be a lot
scarier than they are today.

773
00:41:54,930 --> 00:41:59,280
In the old days,
reading the literature

774
00:41:59,280 --> 00:42:01,410
in a foreign language
was a part of that.

775
00:42:01,410 --> 00:42:04,350
And there was a
high failure rate.

776
00:42:04,350 --> 00:42:08,790
And when you look back
on those failures,

777
00:42:08,790 --> 00:42:12,540
the most usual reason for
people failing an oral exam

778
00:42:12,540 --> 00:42:17,460
is failure to situate and
a failure to practice.

779
00:42:17,460 --> 00:42:20,430
By situate, I mean,
it's important to talk

780
00:42:20,430 --> 00:42:22,680
about your research in context.

781
00:42:22,680 --> 00:42:25,530
This is a problem that's being
pursued all over the world.

782
00:42:25,530 --> 00:42:30,450
There hasn't been any progress
before me in the past 30 years.

783
00:42:30,450 --> 00:42:32,160
Everyone is looking
for a solution

784
00:42:32,160 --> 00:42:33,840
because it will
have impact on so

785
00:42:33,840 --> 00:42:37,380
many other things, such
situating and time and place

786
00:42:37,380 --> 00:42:38,860
and feel.

787
00:42:38,860 --> 00:42:41,970
And then as far as
practice is concerned, yes,

788
00:42:41,970 --> 00:42:43,530
practice is important.

789
00:42:43,530 --> 00:42:48,090
But that doesn't mean showing
your slides to the people

790
00:42:48,090 --> 00:42:50,370
you share an with.

791
00:42:50,370 --> 00:42:54,023
The problem with that is that if
people know what you're doing,

792
00:42:54,023 --> 00:42:55,440
they will hallucinate
that there's

793
00:42:55,440 --> 00:42:57,480
material in your
presentation that isn't there

794
00:42:57,480 --> 00:42:59,748
if it isn't there.

795
00:42:59,748 --> 00:43:01,290
A variation on the
scene, by the way,

796
00:43:01,290 --> 00:43:06,330
is your faculty supervisor
is not a very good person

797
00:43:06,330 --> 00:43:09,300
to help you debug a talk
because they, in fact, know

798
00:43:09,300 --> 00:43:10,200
what you're doing.

799
00:43:10,200 --> 00:43:11,940
And they will, in
fact, hallucinate

800
00:43:11,940 --> 00:43:14,818
there's material in your
presentation that isn't there.

801
00:43:14,818 --> 00:43:16,860
So you need to get together
with some friends who

802
00:43:16,860 --> 00:43:20,180
don't know what you're
doing and have them--

803
00:43:20,180 --> 00:43:23,760
well, you start the
practice session by saying,

804
00:43:23,760 --> 00:43:27,330
if you can't make me cry,
I won't value as a friend

805
00:43:27,330 --> 00:43:27,840
anymore.

806
00:43:27,840 --> 00:43:28,980
[LAUGHTER]

807
00:43:28,980 --> 00:43:33,600
And then when you get to
the faculty on a oral exam,

808
00:43:33,600 --> 00:43:35,350
it will be easy.

809
00:43:35,350 --> 00:43:39,000
You see, difficulty--
the amount of flak

810
00:43:39,000 --> 00:43:41,990
you'll get from somebody
is proportional to age.

811
00:43:41,990 --> 00:43:45,390
The older somebody
is, the more they

812
00:43:45,390 --> 00:43:47,160
understand where they
are in the world.

813
00:43:47,160 --> 00:43:49,470
But the young people are
trying to show the old people

814
00:43:49,470 --> 00:43:52,122
how smart they are, so
it's subtly vicious.

815
00:43:52,122 --> 00:43:53,580
So whenever you
have an opportunity

816
00:43:53,580 --> 00:43:55,140
to have an examining
committee that's

817
00:43:55,140 --> 00:43:59,490
full of people with gray
hair, that's what you want.

818
00:43:59,490 --> 00:44:01,500
Well, that's just a word
or two about something

819
00:44:01,500 --> 00:44:02,460
I haven't listed here.

820
00:44:02,460 --> 00:44:05,610
Let's get into the
subject of job talks.

821
00:44:05,610 --> 00:44:14,230
So I was sitting in a bar
many years ago in San Diego.

822
00:44:14,230 --> 00:44:16,680
I was a member of the
Navy Science Board,

823
00:44:16,680 --> 00:44:21,420
and I was sitting with a couple
of my colleagues on the board

824
00:44:21,420 --> 00:44:24,910
Delores Etter from the
University of Colorado.

825
00:44:24,910 --> 00:44:27,960
She made me so jealous I could
spit because she'd written 21

826
00:44:27,960 --> 00:44:30,660
books, and I'd only written 17.

827
00:44:30,660 --> 00:44:35,130
And then the other
one was Bill Weldon

828
00:44:35,130 --> 00:44:38,230
from the University of Texas.

829
00:44:38,230 --> 00:44:41,280
He was an electromagnetism
guy, and he

830
00:44:41,280 --> 00:44:45,030
knew how to use rail guns to
drive steel rods through tank

831
00:44:45,030 --> 00:44:45,660
armor.

832
00:44:45,660 --> 00:44:47,590
These were interesting people.

833
00:44:47,590 --> 00:44:53,880
So I said, what do you look
for in a faculty candidate?

834
00:44:53,880 --> 00:44:58,000
And within one
microsecond, Delores

835
00:44:58,000 --> 00:45:00,270
said, they have
to show us they've

836
00:45:00,270 --> 00:45:05,940
got some kind of vision, quickly
followed by Bill who said,

837
00:45:05,940 --> 00:45:08,063
they have to show us that
they've done something.

838
00:45:15,160 --> 00:45:18,550
Oh, that sounds good, I said.

839
00:45:18,550 --> 00:45:20,560
And then I said
to them, how long

840
00:45:20,560 --> 00:45:24,640
does a candidate have to
establish these two things?

841
00:45:24,640 --> 00:45:26,850
What do you think?

842
00:45:26,850 --> 00:45:28,510
Well, compare your
answer to theirs.

843
00:45:33,970 --> 00:45:36,803
Five minutes.

844
00:45:36,803 --> 00:45:38,470
So if you haven't
expressed your vision,

845
00:45:38,470 --> 00:45:42,850
if you haven't told people that
you've done something in five

846
00:45:42,850 --> 00:45:45,880
minutes, you've already lost.

847
00:45:45,880 --> 00:45:47,630
So you have to be
able to do that.

848
00:45:47,630 --> 00:45:51,250
And let me just mention a couple
of things in that connection.

849
00:45:51,250 --> 00:45:54,190
Here, the vision is
in part, a problem

850
00:45:54,190 --> 00:46:00,340
that somebody cares about and
something new in your approach.

851
00:46:06,350 --> 00:46:08,560
So the problem is
understanding the nature

852
00:46:08,560 --> 00:46:10,480
of human intelligence.

853
00:46:10,480 --> 00:46:12,640
And the approach
is asking questions

854
00:46:12,640 --> 00:46:16,510
about what makes us
different from chimpanzees

855
00:46:16,510 --> 00:46:18,130
and Neanderthals.

856
00:46:18,130 --> 00:46:19,720
Is it merely a
matter of quantity,

857
00:46:19,720 --> 00:46:23,320
or we're just a little bit
smarter in some continuous way?

858
00:46:23,320 --> 00:46:26,490
Or do we have something
that's fundamentally different

859
00:46:26,490 --> 00:46:30,550
that chimpanzees don't have
and Neanderthals either?

860
00:46:30,550 --> 00:46:34,630
And the answer is yes, we
do have something different.

861
00:46:34,630 --> 00:46:36,900
We are symbolic creatures.

862
00:46:36,900 --> 00:46:39,550
And because we're
symbolic creatures,

863
00:46:39,550 --> 00:46:44,560
we can build
symbolic descriptions

864
00:46:44,560 --> 00:46:46,330
of relations and events.

865
00:46:46,330 --> 00:46:48,835
We can string them
together and make stories.

866
00:46:48,835 --> 00:46:51,760
And because we can make stories,
that's what makes us different.

867
00:46:51,760 --> 00:46:54,190
So that's my stump speech.

868
00:46:54,190 --> 00:46:56,560
That's how I start
most of my talks

869
00:46:56,560 --> 00:46:59,260
on my own personal research.

870
00:46:59,260 --> 00:47:04,990
How do you express the notion
that you've done something?

871
00:47:04,990 --> 00:47:10,240
By listing the steps that need
to be taken in order to achieve

872
00:47:10,240 --> 00:47:12,602
the solution to that problem.

873
00:47:12,602 --> 00:47:14,560
You don't have to have
done all of those steps.

874
00:47:14,560 --> 00:47:18,370
But you can say here's
what needs to be done.

875
00:47:18,370 --> 00:47:20,660
An example, here's
what needs to be done.

876
00:47:20,660 --> 00:47:23,260
We need to specify
some behavior.

877
00:47:23,260 --> 00:47:25,840
We need to enumerate
the constraints that

878
00:47:25,840 --> 00:47:28,690
make it possible to
deal with that behavior.

879
00:47:28,690 --> 00:47:31,030
We have to implement a system
because we're engineers,

880
00:47:31,030 --> 00:47:33,190
and we don't think that
we've understood something

881
00:47:33,190 --> 00:47:35,333
unless we can build it.

882
00:47:35,333 --> 00:47:37,000
And we've built such
a system, and we're

883
00:47:37,000 --> 00:47:38,542
about to demonstrate
it to you today.

884
00:47:38,542 --> 00:47:42,040
That would be an example of
enumerating a series of steps

885
00:47:42,040 --> 00:47:45,290
needed to realize the vision.

886
00:47:45,290 --> 00:47:49,280
So then blah, blah, blah,
blah, blah, blah, blah, blah.

887
00:47:49,280 --> 00:47:51,250
And then you conclude by--

888
00:47:51,250 --> 00:47:54,550
you conclude by enumerating
your contributions.

889
00:48:01,110 --> 00:48:03,930
It's kind of mirror
of these steps.

890
00:48:03,930 --> 00:48:06,610
And it helps to establish
that you've done something.

891
00:48:06,610 --> 00:48:08,730
So that's a kind of
general purpose framework

892
00:48:08,730 --> 00:48:12,180
for doing a technical talk.

893
00:48:12,180 --> 00:48:16,930
Now, only a few more
things left to do today.

894
00:48:16,930 --> 00:48:19,360
Getting famous is the
next item on our agenda

895
00:48:19,360 --> 00:48:20,920
because once you've
got the job, you

896
00:48:20,920 --> 00:48:22,420
need to think a
little bit about how

897
00:48:22,420 --> 00:48:25,790
you're going to be
recognized for what you do.

898
00:48:25,790 --> 00:48:38,140
So oh, first of all, why should
you care about getting famous?

899
00:48:38,140 --> 00:48:40,360
I thought about
this in connection

900
00:48:40,360 --> 00:48:44,920
with a fundraising
event I attended

901
00:48:44,920 --> 00:48:49,920
once, a fundraising
event for raising money

902
00:48:49,920 --> 00:48:52,680
to save Venice from going under
water and having all of its art

903
00:48:52,680 --> 00:48:53,820
destroyed.

904
00:48:53,820 --> 00:49:00,210
Anyway, I was sitting here,
and JC was sitting here.

905
00:49:00,210 --> 00:49:03,210
That was Julia, the
late Julia Child.

906
00:49:08,190 --> 00:49:10,950
And as the evening wore
on, more and more people

907
00:49:10,950 --> 00:49:14,760
would come up and ask Julia
to autograph something

908
00:49:14,760 --> 00:49:19,590
or express a feeling that
she had changed their life.

909
00:49:19,590 --> 00:49:23,380
And it just happened
over and over again.

910
00:49:23,380 --> 00:49:27,810
So eventually, I turned to
Julia, and I said, Ms. Child,

911
00:49:27,810 --> 00:49:30,157
is it fun to be famous?

912
00:49:30,157 --> 00:49:31,740
And she thought about
it for a second.

913
00:49:31,740 --> 00:49:34,752
And she said, you
get used to it.

914
00:49:34,752 --> 00:49:35,900
[LAUGHTER]

915
00:49:35,900 --> 00:49:38,070
But you know what occured to me?

916
00:49:38,070 --> 00:49:41,590
You never get used
to being ignored.

917
00:49:41,590 --> 00:49:45,240
So it's-- here's a
way to think about it.

918
00:49:45,240 --> 00:49:47,100
Your ideas are
like your children.

919
00:49:47,100 --> 00:49:50,370
And you don't want them to
go into the world in rags.

920
00:49:50,370 --> 00:49:51,870
So what you want
to do is to be sure

921
00:49:51,870 --> 00:49:55,390
that you have these
techniques, these mechanisms,

922
00:49:55,390 --> 00:49:58,470
these thoughts about how to
present ideas that you have so

923
00:49:58,470 --> 00:50:02,280
that they're recognized for
the value that is in them.

924
00:50:02,280 --> 00:50:05,370
So that's why it's a legitimate
thing to concern yourself

925
00:50:05,370 --> 00:50:08,010
with packaging.

926
00:50:08,010 --> 00:50:11,070
Now, how do you get remembered?

927
00:50:11,070 --> 00:50:14,720
Well, there's something I
like to call Winston's star.

928
00:50:14,720 --> 00:50:20,130
And every one of the items
I'm about to articulate starts

929
00:50:20,130 --> 00:50:24,510
with an S. So if you want
your presentation ideas

930
00:50:24,510 --> 00:50:27,240
to be remembered, one of
the things you need to do

931
00:50:27,240 --> 00:50:32,820
is to make sure that you have
some kind of symbol associated

932
00:50:32,820 --> 00:50:34,590
with your work.

933
00:50:34,590 --> 00:50:37,980
So this arch example is
actually from my PhD thesis

934
00:50:37,980 --> 00:50:39,780
many, many years ago.

935
00:50:39,780 --> 00:50:45,000
And in the course of
my work at that time,

936
00:50:45,000 --> 00:50:47,820
this work on arch learning
became mildly famous,

937
00:50:47,820 --> 00:50:49,770
and I didn't know why.

938
00:50:49,770 --> 00:50:51,630
It was only many years
later that I realized

939
00:50:51,630 --> 00:50:54,660
that that work accidentally
had all of the elements

940
00:50:54,660 --> 00:50:56,220
on this star.

941
00:50:56,220 --> 00:50:59,910
So the first element is that
there was a kind of symbol.

942
00:50:59,910 --> 00:51:00,900
It's the arch itself.

943
00:51:05,510 --> 00:51:10,480
Next thing you need is
some kind of slogan,

944
00:51:10,480 --> 00:51:13,740
a kind of phrase that
provides a handle on the work.

945
00:51:13,740 --> 00:51:17,160
And in this case, the phrase
was one shot learning.

946
00:51:20,120 --> 00:51:23,340
And it was one shot
because the program I wrote

947
00:51:23,340 --> 00:51:27,030
learned something definite
from every example that

948
00:51:27,030 --> 00:51:28,860
was presented to us.

949
00:51:28,860 --> 00:51:32,970
So in going from a model
based on this configuration

950
00:51:32,970 --> 00:51:35,940
to something that isn't an arch
base on that configuration,

951
00:51:35,940 --> 00:51:38,640
the program learned that
it has to be on top,

952
00:51:38,640 --> 00:51:41,490
one shot learning.

953
00:51:41,490 --> 00:51:45,210
So that's a symbol, slogan.

954
00:51:45,210 --> 00:51:46,575
And now we need a surprise.

955
00:51:51,330 --> 00:51:54,290
Yeah, the surprise is you
don't need a million examples

956
00:51:54,290 --> 00:51:55,910
of something to learn.

957
00:51:55,910 --> 00:51:58,010
You can do it with
one example if you're

958
00:51:58,010 --> 00:52:02,180
smart enough to make use of
that example appropriately.

959
00:52:02,180 --> 00:52:03,230
So that was the surprise.

960
00:52:03,230 --> 00:52:07,160
You can learn something
definite from each example.

961
00:52:07,160 --> 00:52:08,840
Next item was a salient idea.

962
00:52:13,280 --> 00:52:16,785
Now, when I say salient
idea, I don't mean important.

963
00:52:16,785 --> 00:52:18,410
What I mean is an
idea that sticks out.

964
00:52:21,260 --> 00:52:24,080
Some theses, funnily enough,
have too many good ideas,

965
00:52:24,080 --> 00:52:27,110
and you don't know what it's all
about because which one is it?

966
00:52:27,110 --> 00:52:29,290
So you need an idea
that sticks out.

967
00:52:29,290 --> 00:52:33,890
And the idea that stuck out here
was the notion of a near miss.

968
00:52:39,280 --> 00:52:43,650
You see, this is not an arch,
but it doesn't miss by much.

969
00:52:43,650 --> 00:52:45,490
So it's a near miss.

970
00:52:45,490 --> 00:52:50,920
And finally, you need
to tell the story

971
00:52:50,920 --> 00:52:56,160
of how you did it, how it
works, why it's important.

972
00:52:56,160 --> 00:53:01,720
So that's a bit on how to
not so much get famous,

973
00:53:01,720 --> 00:53:06,240
but how to ensure that
your work is recognized.

974
00:53:06,240 --> 00:53:08,890
Well, we're almost
finished because now we're

975
00:53:08,890 --> 00:53:13,470
down to this last item,
which is how to stop.

976
00:53:13,470 --> 00:53:14,998
And when we come
to that, there's

977
00:53:14,998 --> 00:53:17,290
a question of all right,
well, what is the final slide?

978
00:53:22,412 --> 00:53:23,620
And what are the final words?

979
00:53:30,610 --> 00:53:32,830
So for the final
slide, let me give you

980
00:53:32,830 --> 00:53:34,750
some examples of possibilities.

981
00:53:37,588 --> 00:53:38,380
How about this one?

982
00:53:42,150 --> 00:53:50,610
Well, you might see that
slide and think to yourself,

983
00:53:50,610 --> 00:53:53,730
there are 1,000 faculty at MIT.

984
00:53:53,730 --> 00:53:55,770
Nice piece of work,
but not so much,

985
00:53:55,770 --> 00:54:01,050
but it's only a tiny piece of
work if you divide by 1,000.

986
00:54:01,050 --> 00:54:03,810
So when you show a whole
gigantic list of collaborators

987
00:54:03,810 --> 00:54:07,380
at the end of a talk,
it's a kind of let down

988
00:54:07,380 --> 00:54:09,150
because it suggests
that nobody knows.

989
00:54:09,150 --> 00:54:12,650
Well, did you do
anything significant?

990
00:54:12,650 --> 00:54:15,180
Now, you've got to recognize
your collaborators, right?

991
00:54:15,180 --> 00:54:16,970
So where do you do that?

992
00:54:16,970 --> 00:54:19,877
Not on the last slide,
on the first slide.

993
00:54:19,877 --> 00:54:21,210
All this was on the first slide.

994
00:54:21,210 --> 00:54:23,263
These are the
collaborators, so you don't

995
00:54:23,263 --> 00:54:24,430
want to put them at the end.

996
00:54:24,430 --> 00:54:25,805
You don't want a
slide like this.

997
00:54:28,290 --> 00:54:31,060
How about this one?

998
00:54:31,060 --> 00:54:33,050
This is the worst possible
way to end a talk.

999
00:54:33,050 --> 00:54:34,700
[LAUGHTER]

1000
00:54:34,700 --> 00:54:39,410
Because this slide can be
up there for 20 minutes.

1001
00:54:39,410 --> 00:54:40,660
I've seen it happen.

1002
00:54:40,660 --> 00:54:43,820
It squanders real estate.

1003
00:54:43,820 --> 00:54:47,420
It squanders an opportunity
to tell people who you are.

1004
00:54:47,420 --> 00:54:48,200
It's just--

1005
00:54:51,777 --> 00:54:52,610
What about this one?

1006
00:54:56,780 --> 00:54:57,800
I often see it.

1007
00:54:57,800 --> 00:55:01,300
I never see anybody
write it down.

1008
00:55:01,300 --> 00:55:03,991
Also, it wastes opportunity.

1009
00:55:06,580 --> 00:55:11,440
Oh my God, even worse.

1010
00:55:14,970 --> 00:55:18,180
All of these lines
do nothing for you.

1011
00:55:18,180 --> 00:55:21,840
They waste an opportunity
for you to tell people--

1012
00:55:21,840 --> 00:55:23,730
to leave people with what you--

1013
00:55:23,730 --> 00:55:26,360
with who you are.

1014
00:55:26,360 --> 00:55:29,000
Well, what about this?

1015
00:55:29,000 --> 00:55:31,450
Is this a good one?

1016
00:55:31,450 --> 00:55:34,660
It might seem so at first,
but here's the problem.

1017
00:55:38,340 --> 00:55:41,040
If you say these
are my conclusions,

1018
00:55:41,040 --> 00:55:42,960
these are perfectly
legitimate conclusions

1019
00:55:42,960 --> 00:55:44,910
that nobody cares about.

1020
00:55:44,910 --> 00:55:47,557
What they care about
is what you have done.

1021
00:55:47,557 --> 00:55:49,140
And that's why your
final slide should

1022
00:55:49,140 --> 00:55:53,460
have this label, contributions.

1023
00:55:53,460 --> 00:55:56,510
It's a mirror of what I said
over there about how job talks

1024
00:55:56,510 --> 00:55:57,830
ought to be like a sandwich.

1025
00:55:57,830 --> 00:55:59,460
And the final slide,
the one that's

1026
00:55:59,460 --> 00:56:02,240
up there while people are
asking questions and filing out,

1027
00:56:02,240 --> 00:56:05,470
it ought to be the one that
has your contributions on it.

1028
00:56:05,470 --> 00:56:09,140
Here's an example from
my own stump speech.

1029
00:56:09,140 --> 00:56:13,020
Yeah, this is what
I talk about a lot.

1030
00:56:13,020 --> 00:56:16,912
Yes, here are the things
that I typically demonstrate.

1031
00:56:21,160 --> 00:56:23,146
And I wait for
people to read it.

1032
00:56:23,146 --> 00:56:27,616
And the final element there is
this is what we get out of it,

1033
00:56:27,616 --> 00:56:32,780
so that's an example of
a contribution slide.

1034
00:56:32,780 --> 00:56:37,670
All right, now, what
about the other part?

1035
00:56:37,670 --> 00:56:39,740
You got your final
slide up there.

1036
00:56:39,740 --> 00:56:41,150
It's a contribution slide.

1037
00:56:41,150 --> 00:56:44,430
Somehow you have to tell
people you're finished.

1038
00:56:44,430 --> 00:56:47,551
So let's see it, check
out a few possibilities.

1039
00:56:50,140 --> 00:56:54,700
One thing you could
do in the final words

1040
00:56:54,700 --> 00:56:56,000
is you could tell a joke.

1041
00:56:59,840 --> 00:57:02,232
It's OK.

1042
00:57:02,232 --> 00:57:04,565
By the time you're done,
people have adjusted themselves

1043
00:57:04,565 --> 00:57:05,728
to your voice parameters.

1044
00:57:05,728 --> 00:57:06,770
They're ready for a joke.

1045
00:57:10,040 --> 00:57:13,010
I was sitting in another bar,
this time in Austin, Texas

1046
00:57:13,010 --> 00:57:16,310
with a colleague of
mine named Doug Lenat.

1047
00:57:16,310 --> 00:57:18,880
And Doug's a fantastic speaker.

1048
00:57:18,880 --> 00:57:21,920
And so I said to
Doug, Doug, you're

1049
00:57:21,920 --> 00:57:24,860
a fantastic speaker,
what's your secret?

1050
00:57:24,860 --> 00:57:29,440
And he said, oh, I always
finish with a joke,

1051
00:57:29,440 --> 00:57:32,120
and that way, people think
they've had fun the whole time.

1052
00:57:32,120 --> 00:57:35,480
[LAUGHTER]

1053
00:57:35,480 --> 00:57:39,540
So yeah, a joke will
work down there.

1054
00:57:39,540 --> 00:57:40,580
How about this one?

1055
00:57:50,770 --> 00:57:51,270
Thank you.

1056
00:57:54,030 --> 00:57:55,650
I don't recommend it.

1057
00:57:55,650 --> 00:57:58,410
It's a weak move.

1058
00:57:58,410 --> 00:58:00,330
You will not go to
hell if you conclude

1059
00:58:00,330 --> 00:58:02,730
your talk by saying thank
you, but it's a weak move,

1060
00:58:02,730 --> 00:58:04,880
and here's why.

1061
00:58:04,880 --> 00:58:08,460
When you say thank you, even
worse, thank you for listening,

1062
00:58:08,460 --> 00:58:10,350
it suggests that
everybody has stayed

1063
00:58:10,350 --> 00:58:14,310
that long out of politeness and
that they had a profound desire

1064
00:58:14,310 --> 00:58:15,632
to be somewhere else.

1065
00:58:15,632 --> 00:58:17,340
But they're so polite,
they stuck it out.

1066
00:58:17,340 --> 00:58:20,430
And that's what you're
thanking them for.

1067
00:58:20,430 --> 00:58:21,990
So once wild
applause has started,

1068
00:58:21,990 --> 00:58:24,240
you can mouth a thank
you, and there's

1069
00:58:24,240 --> 00:58:25,380
nothing wrong with that.

1070
00:58:25,380 --> 00:58:29,992
But the last thing you do
should not be saying thank you.

1071
00:58:29,992 --> 00:58:31,950
Now, you say to me, well,
doesn't everybody say

1072
00:58:31,950 --> 00:58:33,570
thank you?

1073
00:58:33,570 --> 00:58:37,800
Well, what everybody does is
not necessarily the right thing.

1074
00:58:37,800 --> 00:58:40,110
And I like to illustrate
how some talks can

1075
00:58:40,110 --> 00:58:41,580
end without saying thank you.

1076
00:58:41,580 --> 00:58:45,210
I like to draw from
political speeches,

1077
00:58:45,210 --> 00:58:47,956
but the ones that I've heard
recently aren't so good, so--

1078
00:58:47,956 --> 00:58:50,240
[LAUGHTER]

1079
00:58:50,240 --> 00:58:55,500
So I'm going to have to
go go back a little bit.

1080
00:58:55,500 --> 00:58:57,010
So here is Governor Christie.

1081
00:58:57,010 --> 00:59:03,680
He gave the Republican
keynote address one year.

1082
00:59:03,680 --> 00:59:05,040
This is the end of his talk.

1083
00:59:05,040 --> 00:59:07,214
Let's see what he does.

1084
00:59:07,214 --> 00:59:08,966
[APPLAUSE]

1085
00:59:08,966 --> 00:59:11,920
CHRIS CHRISTIE: And together,
everybody, together.

1086
00:59:11,920 --> 00:59:16,060
We will stand up once again
for American greatness

1087
00:59:16,060 --> 00:59:18,190
for our children
and grandchildren.

1088
00:59:18,190 --> 00:59:20,583
God bless you, and
God bless America.

1089
00:59:20,583 --> 00:59:23,241
[APPLAUSE]

1090
00:59:23,241 --> 00:59:27,260
PATRICK WINSTON: So that's a
classic benediction ending.

1091
00:59:27,260 --> 00:59:30,620
God bless you,
God bless America.

1092
00:59:30,620 --> 00:59:32,610
Now, I don't want to
be partisan about this.

1093
00:59:32,610 --> 00:59:35,570
So I think I'd better switch
to the keynote address

1094
00:59:35,570 --> 00:59:38,840
in the Democratic Convention.

1095
00:59:38,840 --> 00:59:42,050
It was delivered that year
by Bill Clinton, who knows

1096
00:59:42,050 --> 00:59:45,010
something about how to speak.

1097
00:59:45,010 --> 00:59:46,380
[APPLAUSE]

1098
00:59:46,380 --> 00:59:48,940
BILL CLINTON: If that
is what you want,

1099
00:59:48,940 --> 00:59:52,300
if that is what you
believe, you must vote,

1100
00:59:52,300 --> 00:59:57,046
and you must re-elect
President Barack Obama.

1101
00:59:57,046 --> 00:59:58,588
God bless you, and
God bless America.

1102
00:59:58,588 --> 01:00:00,061
[APPLAUSE]

1103
01:00:00,061 --> 01:00:03,989
[LAUGHTER]

1104
01:00:09,390 --> 01:00:11,410
PATRICK WINSTON:
Now, watch this.

1105
01:00:11,410 --> 01:00:13,297
Let's go back a little
bit and redo it.

1106
01:00:13,297 --> 01:00:15,130
What I want you to see
is that at one point,

1107
01:00:15,130 --> 01:00:18,430
he seems to be almost pressing
his lips together, forcing

1108
01:00:18,430 --> 01:00:20,315
himself not to say thank you.

1109
01:00:20,315 --> 01:00:22,690
Then there's another place
where he does a little salute.

1110
01:00:22,690 --> 01:00:24,784
So watch for those
this time around.

1111
01:00:24,784 --> 01:00:25,650
[APPLAUSE]

1112
01:00:25,650 --> 01:00:28,210
BILL CLINTON: If that
is what you want,

1113
01:00:28,210 --> 01:00:31,470
if that is what you
believe, you must vote,

1114
01:00:31,470 --> 01:00:36,310
and you must re-elect
President Barack Obama.

1115
01:00:36,310 --> 01:00:38,470
God bless you., and
God bless America.

1116
01:00:38,470 --> 01:00:39,625
[APPLAUSE]

1117
01:00:39,625 --> 01:00:42,208
PATRICK WINSTON: That's
where he's pursing his lips.

1118
01:00:42,208 --> 01:00:43,186
[LAUGHTER]

1119
01:00:43,186 --> 01:00:45,142
There's the salute.

1120
01:00:45,142 --> 01:00:49,455
[LAUGHTER]

1121
01:00:49,455 --> 01:00:50,830
Yeah, I think
that's pretty good.

1122
01:00:50,830 --> 01:00:52,640
Now, what are we gonna
take away from this?

1123
01:00:52,640 --> 01:00:57,910
Well, I suppose I could conclude
this talk by saying God bless

1124
01:00:57,910 --> 01:01:00,960
you, and God bless the
Massachusetts Institute

1125
01:01:00,960 --> 01:01:04,732
of Technology, but it
might not work so well.

1126
01:01:04,732 --> 01:01:06,190
But what you can
get out of this is

1127
01:01:06,190 --> 01:01:08,110
you don't have to say thank you.

1128
01:01:08,110 --> 01:01:09,850
There are other
things you can do.

1129
01:01:09,850 --> 01:01:13,060
And it's interesting that over
time, people figure this out,

1130
01:01:13,060 --> 01:01:15,730
and there's some stock
ways of ending things.

1131
01:01:15,730 --> 01:01:21,180
So in the Catholic church,
and the good old Latin mass,

1132
01:01:21,180 --> 01:01:27,990
it ended with ite
missa est, which

1133
01:01:27,990 --> 01:01:31,380
translates approximately
to OK, the mass is over,

1134
01:01:31,380 --> 01:01:32,718
you can go home now.

1135
01:01:32,718 --> 01:01:34,470
[LAUGHTER]

1136
01:01:34,470 --> 01:01:37,440
And of course, at
musical concerts,

1137
01:01:37,440 --> 01:01:42,230
you know that it's time to clap
not at the end of the song,

1138
01:01:42,230 --> 01:01:45,810
but rather when the conductor
goes over and shakes hands

1139
01:01:45,810 --> 01:01:47,280
with the concert master.

1140
01:01:47,280 --> 01:01:50,850
Those are conventions that tell
you that the event is over.

1141
01:01:53,470 --> 01:01:58,500
So those are all
possibilities for here.

1142
01:01:58,500 --> 01:02:00,180
But one more
possibility, and that

1143
01:02:00,180 --> 01:02:07,320
is that you can
salute the audience.

1144
01:02:07,320 --> 01:02:09,070
And by that, I mean,
you can say something

1145
01:02:09,070 --> 01:02:13,060
about how much you value
your time at a place.

1146
01:02:13,060 --> 01:02:20,000
So I could say, well, it's
been great fun being here.

1147
01:02:20,000 --> 01:02:22,780
It's been fascinating to see
what you folks are doing here

1148
01:02:22,780 --> 01:02:24,190
at MIT.

1149
01:02:24,190 --> 01:02:27,535
I've been much
stimulated and provoked

1150
01:02:27,535 --> 01:02:29,410
by the kinds of questions
you've been asking,

1151
01:02:29,410 --> 01:02:30,800
it's been really great.

1152
01:02:30,800 --> 01:02:33,380
And I look forward
to coming back

1153
01:02:33,380 --> 01:02:35,210
on many occasions in the future.

1154
01:02:35,210 --> 01:02:36,693
So that salutes the audience.

1155
01:02:36,693 --> 01:02:37,360
You can do that.

1156
01:02:40,670 --> 01:02:42,520
Well, there it is.

1157
01:02:42,520 --> 01:02:45,260
You know what?

1158
01:02:45,260 --> 01:02:46,940
I'm glad you're here.

1159
01:02:46,940 --> 01:02:48,830
And the reason is
by being here, I

1160
01:02:48,830 --> 01:02:51,590
think you have demonstrated
an understanding

1161
01:02:51,590 --> 01:02:54,770
that how you present and how
you package your ideas is

1162
01:02:54,770 --> 01:02:56,510
an important thing.

1163
01:02:56,510 --> 01:02:58,013
And I salute you for that.

1164
01:02:58,013 --> 01:03:03,500
[LAUGHTER] And I suggest that
you come back again and bring

1165
01:03:03,500 --> 01:03:05,990
your friends.

1166
01:03:05,990 --> 01:03:09,340
[APPLAUSE]