WEBVTT

1
00:00:00.333 --> 00:00:04.796
A noise complaint, holiday lights, and a coffee stain.

2
00:00:05.797 --> 00:00:07.618
What could they possibly have in common?

3
00:00:08.639 --> 00:00:12.622
They are the unlikely beginnings of discoveries with unexpected benefits.

4
00:00:13.302 --> 00:00:18.306
And this year's Golden Goose Awards honors the federally funded research behind those stories.

5
00:00:19.367 --> 00:00:27.572
Junsu and her colleagues spent decades turning an unexpected lab result into a way to destroy liver tumors with sound.

6
00:00:28.903 --> 00:00:32.606
NASA's Black Marble team spotted holiday light patterns from space.

7
00:00:33.527 --> 00:00:42.075
The same tools that reveal cities brightening from celebrations can show them going dark after disasters and track power returning.

8
00:00:43.056 --> 00:00:53.044
And Sidney Nagel found deep physics inside everyday messes, shifting sand, splashing drops, and the ring left by a drying drop of coffee.

9
00:00:54.005 --> 00:01:02.528
Turning those surprises into something useful took years of experiments, sustained support, and people willing to keep asking questions.

10
00:01:03.529 --> 00:01:09.231
Today, we're exploring the science behind these discoveries and the questions that made them possible.

11
00:01:10.071 --> 00:01:14.533
This is From First Principles, and these are the Golden Goose Awards.

12
00:01:19.462 --> 00:01:20.203
Hello, Internet.

13
00:01:20.223 --> 00:01:29.030
This is your captain speaking, Lester Nare, joined as always by my co-host and our resident PhD, Krishna Chaudhary.

14
00:01:29.811 --> 00:01:42.101
This week is part one of our special coverage of the 2026 Golden Goose Awards organized by the American Association for the Advancement of Science, or AAAS.

15
00:01:44.018 --> 00:01:56.066
These awards are meant to shine a light on federally funded research that starts out as basic research and results in incredible practical applications, hence the name Golden Goose.

16
00:01:56.747 --> 00:02:03.832
By the time you are watching this, we will have attended the ceremony at the Library of Congress in Washington, D.C.,

17
00:02:04.212 --> 00:02:11.135
and have done interviews with all three of the award-winning research stories that we are going to cover today.

18
00:02:11.535 --> 00:02:19.258
Those interviews, along with our conversation with AAAS CEO Sudeep Parikh, will be coming out in part now.

19
00:02:19.538 --> 00:02:26.844
Today, we're going to be getting to know the researchers and unpacking the science behind their discoveries.

20
00:02:26.944 --> 00:02:30.067
As always, we are going to talk about the science from the ground up.

21
00:02:30.187 --> 00:02:38.814
And if you want to get the best coverage of the latest breaking science around the world, be sure to subscribe.

22
00:02:39.315 --> 00:02:40.756
Now, for our first story.

23
00:02:42.527 --> 00:02:50.530
Let's start with a noise complaint that helped set a remarkable discovery in motion.

24
00:02:51.570 --> 00:03:08.336
Our first story is From Noise Complaint to Cancer Treatment, recognizing Dr. Junshu, the Li Ka-shing Endowed Professor of Biomedical Engineering at the University of Michigan, UMICH.

25
00:03:09.156 --> 00:03:15.538
In the early 2000s, Shu was a graduate student in the late Professor Charles Kane's lab.

26
00:03:16.458 --> 00:03:18.879
Unfortunately, he passed away in 2020.

27
00:03:19.219 --> 00:03:22.520
They were investigating ways to remove heart tissue using ultrasound.

28
00:03:23.220 --> 00:03:31.023
And the idea was to help children who had congenital heart disease without open surgery, which would obviously be...

29
00:03:31.763 --> 00:03:32.304
Incredible.

30
00:03:33.164 --> 00:03:40.870
But the experimental setup was so loud that her lab mates started to make some complaints.

31
00:03:41.431 --> 00:03:45.213
So she tried changing the pulse sequence to quiet it down.

32
00:03:45.994 --> 00:03:53.640
And when she tested those pulses on pig hearts, she saw something remarkable, which was that the tissue was breaking apart.

33
00:03:55.292 --> 00:04:11.801
That observation helped launch more than two decades of work by Xu and her colleagues to develop histotripsy, one of my favorite names for an area of study, with funding support from the National Institutes of Health and the Department of Defense.

34
00:04:12.701 --> 00:04:15.743
Now, this is a great visual labeled bubble cloud.

35
00:04:17.374 --> 00:04:19.096
of what we're talking about here.

36
00:04:19.196 --> 00:04:26.106
So histotripsy uses short, intense pulses of focused ultrasound to create a cloud of tiny bubbles in targeted tissue.

37
00:04:26.447 --> 00:04:27.889
As those bubbles expand and collapse,

38
00:04:28.158 --> 00:04:32.940
They physically break the tissue apart, all without an incision.

39
00:04:33.000 --> 00:04:37.062
And that's what we're looking at here, which is just an incredible visual.

40
00:04:37.142 --> 00:04:38.082
Absolutely incredible.

41
00:04:38.583 --> 00:04:39.443
To understand.

42
00:04:39.803 --> 00:04:42.544
And the complexity of this just blows my mind.

43
00:04:42.584 --> 00:04:42.824
Yeah.

44
00:04:43.004 --> 00:04:49.027
So Shu and her colleagues co-founded Histosonics in 2009 to help bring this technology to patients.

45
00:04:49.127 --> 00:04:55.590
Again, starting from a place of just trying to make the lab mates happy because there's a little bit of loudness happening.

46
00:04:56.745 --> 00:05:05.171
Its Edison system received FDA authorization in 2023 for non-invasive destruction of liver tumors specifically.

47
00:05:05.651 --> 00:05:15.958
But the question is, how do you control something as violent as the collapsing of bubbles precisely enough to use it inside the human body?

48
00:05:16.498 --> 00:05:23.923
Yeah, I mean, it's an incredible question that first you have to, like, control the collapsing bubbles and then you have to move it around.

49
00:05:23.963 --> 00:05:26.085
If you go back to that video that you just showed, right?

50
00:05:27.122 --> 00:05:33.067
There's a bubble that you've somehow engineered using sound inside of the human body, which has tissue.

51
00:05:33.587 --> 00:05:37.770
So it's not like water where the speed of sound is the same everywhere, right?

52
00:05:38.090 --> 00:05:39.892
You've got to control where the bubble is.

53
00:05:39.972 --> 00:05:40.913
It can't be too big.

54
00:05:41.153 --> 00:05:44.375
Otherwise, it messes up all the other tissue that's nearby.

55
00:05:44.675 --> 00:05:46.157
You've got to control where it is.

56
00:05:46.577 --> 00:05:48.018
And you have to control the duration.

57
00:05:48.098 --> 00:05:49.099
Otherwise, it's going to kill...

58
00:05:49.559 --> 00:05:51.560
or cause heat and things like that.

59
00:05:53.060 --> 00:05:55.621
Fundamentally, it's kind of a physics question, right?

60
00:05:57.601 --> 00:05:58.582
How do you actually do that?

61
00:05:58.882 --> 00:06:09.445
So for this segment, I wanted to go through five papers that Chu co-authored along her journey from the 2004 all the way till now.

62
00:06:09.505 --> 00:06:15.706
She is obviously a professor now at the University of Michigan, but she's actually also an alumni.

63
00:06:15.746 --> 00:06:17.467
She did her master's and her PhD there.

64
00:06:17.487 --> 00:06:17.607
Yeah.

65
00:06:18.067 --> 00:06:22.250
And the story starts when she was a graduate student at the University of Michigan.

66
00:06:22.710 --> 00:06:29.015
Obviously, the University of Michigan hired her because they're like, yeah, she's going to win the Golden Goose Awards one day.

67
00:06:29.975 --> 00:06:39.022
So it's an early experiment where you have to control cavitation and then try to figure out how it's going to actually work.

68
00:06:39.322 --> 00:06:39.502
Right.

69
00:06:39.842 --> 00:06:42.444
So the first paper we're going to look at is in 2004.

70
00:06:42.464 --> 00:06:46.647
This is the original paper that kind of puts her on the map.

71
00:06:47.751 --> 00:06:52.354
The original medical problem was hypoplastic left heart syndrome.

72
00:06:52.614 --> 00:06:59.920
You alluded to this with the idea that some newborns, the left side of the heart is severely underdeveloped.

73
00:07:00.040 --> 00:07:08.125
And if the partition, you can see over there on the left is a normal heart, and on the right-hand side is a heart with hypoplastic left heart syndrome.

74
00:07:08.426 --> 00:07:13.950
The left ventricle and the left atrium are much smaller than the right-hand side, right?

75
00:07:14.350 --> 00:07:14.510
And...

76
00:07:16.259 --> 00:07:24.825
The way the heart works is you need both of these ventricles to pump blood in and out of the lungs, oxygenate it, and then put it back, right?

77
00:07:25.045 --> 00:07:29.329
There's one pump for the entire body, and that's the heart.

78
00:07:29.749 --> 00:07:35.073
And if the pump has like two motors you can think about and one of them is missing, it's not very good.

79
00:07:35.553 --> 00:07:44.920
On top of that, the connection valve between the atrium and the ventricle, oftentimes there's no hole.

80
00:07:45.973 --> 00:07:47.153
Like they're not even connected.

81
00:07:47.233 --> 00:07:52.035
So it's very, very difficult to even have a normally functioning heart, right?

82
00:07:52.075 --> 00:08:01.318
And that's why hypoplastic left heart syndrome, it's a defect of where the structures on the left-hand side are hypoplastic, meaning very small or not even formed at all.

83
00:08:02.718 --> 00:08:08.280
Okay, so what we'd like to do is create a little hole in that valve.

84
00:08:08.858 --> 00:08:11.559
so that it becomes a valve and it can open and close.

85
00:08:11.940 --> 00:08:20.344
Okay, so Zhu's group asked whether focused ultrasound might make an opening without putting a cutting instrument through the tissue.

86
00:08:20.804 --> 00:08:21.545
Which makes sense.

87
00:08:21.605 --> 00:08:34.392
It's basically, can we change the structure such that the functionality we're looking for is a possibility without too much second and third order consequences or damage to the other surrounding?

88
00:08:34.412 --> 00:08:34.972
Yeah.

89
00:08:35.132 --> 00:08:35.473
areas.

90
00:08:35.513 --> 00:08:36.135
Yeah, that's right.

91
00:08:36.215 --> 00:08:40.868
And in 2004, they have this paper that comes out, Controlled Ultrasound Tissue Erosion.

92
00:08:41.349 --> 00:08:41.830
It's in the IEEE

93
00:08:43.852 --> 00:08:48.436
conference papers and the paper does not establish a clinical treatment.

94
00:08:48.576 --> 00:08:48.856
Okay.

95
00:08:48.937 --> 00:08:51.419
It's just doing experiments on a model organism.

96
00:08:51.459 --> 00:08:55.102
In this case, canines before the bubbles, let's just think about sound waves.

97
00:08:55.142 --> 00:08:55.402
Okay.

98
00:08:55.422 --> 00:08:57.323
So ultrasound is a sound wave.

99
00:08:57.864 --> 00:09:00.346
It's a sound wave as an oscillation in pressure.

100
00:09:01.146 --> 00:09:10.634
And at, at a point in tissue, you can focus all the pressure and the pressure alternates between like some ambient value and then above and below that ambient value.

101
00:09:11.335 --> 00:09:11.535
Um, um,

102
00:09:13.264 --> 00:09:17.325
You might have heard about ultrasound therapy for cancer, for example.

103
00:09:17.866 --> 00:09:26.028
So in this case, what ends up happening is you focus the ultrasound using a transducer into a tumor.

104
00:09:26.988 --> 00:09:30.449
And those pressure amplitudes are there away from the region.

105
00:09:30.469 --> 00:09:31.030
They're not there.

106
00:09:31.430 --> 00:09:35.911
This is called acoustic focusing because you're taking sound and you're focusing kind of like a lens.

107
00:09:36.631 --> 00:09:41.973
And in this conventional focused ultrasound treatment, the usual picture is heat.

108
00:09:42.782 --> 00:09:43.202
in the middle.

109
00:09:43.562 --> 00:09:47.404
You just want to heat up that tumor and like burn it off, literally, right?

110
00:09:47.965 --> 00:09:54.408
The tissue absorbs that acoustic energy, the temperature rises, and then the sufficiently hot cells die off.

111
00:09:54.888 --> 00:09:57.410
So perhaps you can kill the tumor that way, right?

112
00:09:58.470 --> 00:10:00.392
Zhu's group was after another outcome.

113
00:10:00.752 --> 00:10:06.656
They wanted the tissue to disappear from a very sharply defined region, like an actual opening.

114
00:10:07.097 --> 00:10:10.719
You don't want to merely patch the tissue that's like destined to die later.

115
00:10:10.739 --> 00:10:12.000
You want to literally open it up.

116
00:10:12.561 --> 00:10:17.024
You want to destroy the mechanically rather than cooking it in some sense.

117
00:10:17.625 --> 00:10:17.985
Right.

118
00:10:18.045 --> 00:10:23.729
And you want it to be precise in 3D space because you're trying to create this opening.

119
00:10:23.829 --> 00:10:24.150
Exactly.

120
00:10:24.170 --> 00:10:25.751
It's not just some generalized region.

121
00:10:25.951 --> 00:10:26.391
Exactly.

122
00:10:26.451 --> 00:10:33.996
And so their experience, they did a bunch of experiments where they varied pulse duration, how often the pulses are repeated.

123
00:10:35.157 --> 00:10:41.842
Very short, intense pulses, they found, delivered at suitable repetition rates, eroded the tissue very effectively.

124
00:10:42.622 --> 00:10:44.704
And you could see this effect happening.

125
00:10:44.724 --> 00:10:47.886
You could see in the model organisms...

126
00:10:48.226 --> 00:10:51.350
you can see they're creating holes in the atrial septum.

127
00:10:51.791 --> 00:10:57.058
And then on the right-hand side, they use Doppler ultrasound to observe flow through it.

128
00:10:57.118 --> 00:11:00.542
For anyone who has been lucky enough to have a child

129
00:11:02.955 --> 00:11:05.017
you recognize the Doppler ultrasound.

130
00:11:05.057 --> 00:11:09.901
That's like when they, you look at like the, you literally use the Doppler effect.

131
00:11:10.341 --> 00:11:16.166
This is something that I saw when I was in the hospital, like looking at the, you know, the fetus that ultimately became my child.

132
00:11:17.427 --> 00:11:26.515
But you can look at, you can visualize blood flow by using the Doppler effect to see like stuff that's moving out and stuff that's moving in.

133
00:11:26.575 --> 00:11:29.097
Like there's literally, you can observe red and blue shifts.

134
00:11:29.898 --> 00:11:35.623
And the color coding is red and blue for like stuff moving away from you and stuff moving closer to you.

135
00:11:35.663 --> 00:11:37.484
And I was looking at that going, that's pretty cool.

136
00:11:37.504 --> 00:11:38.265
That's pretty cool.

137
00:11:38.305 --> 00:11:39.146
That's pretty cool, right?

138
00:11:39.226 --> 00:11:44.311
Usually you learn about that stuff in astrophysics when it's like redshift, blueshift, stuff moving away.

139
00:11:44.371 --> 00:11:50.856
But here they use the same colors to show like movement of fluid.

140
00:11:51.337 --> 00:11:52.538
And so just briefly as we...

141
00:11:53.198 --> 00:11:57.139
particularly for some of our audio listeners, we're looking at two sets of images.

142
00:11:57.159 --> 00:12:03.481
The one on the left is showing basically a before and after of this approach of the actual physical tissue.

143
00:12:03.521 --> 00:12:03.761
Yeah.

144
00:12:03.801 --> 00:12:03.981
Yeah.

145
00:12:04.081 --> 00:12:11.023
I think these are both after actually both after where it's now you're creating these cavities, like literal holes in the tissue.

146
00:12:11.363 --> 00:12:18.845
And on the right hand side, you're using the Doppler effect to show that fluid is moving through that tissue, which is ultimately the functionality that you're trying to restore.

147
00:12:18.865 --> 00:12:19.025
Exactly.

148
00:12:19.065 --> 00:12:19.445
Exactly.

149
00:12:20.166 --> 00:12:23.073
And the way they do this is with cavitation.

150
00:12:23.855 --> 00:12:29.248
So cavitation is the idea of using sound waves to create a bubble and then...

151
00:12:30.367 --> 00:12:34.768
precisely maneuvering that bubble, the size of it, and so on and so forth.

152
00:12:34.848 --> 00:12:41.290
And it's an intuitive distinction between cooking versus mechanical disintegration, which we talked about earlier, right?

153
00:12:41.330 --> 00:12:46.732
Cooking is what we do with the traditional ultrasound therapies, when we want to just, like, heat stuff up.

154
00:12:47.112 --> 00:12:54.194
Here, we're using a bubble to, like, expand and contract rapidly, and that's going to mess with all of the tissue around it.

155
00:12:54.414 --> 00:12:57.075
And it's incredible that we can do this with sound waves, right?

156
00:12:57.095 --> 00:12:57.235
Yeah.

157
00:12:59.092 --> 00:13:01.994
The material is still there, but the proteins in the structure have changed.

158
00:13:02.354 --> 00:13:08.839
And like with this histotripsy, you can use that pressure to start this cavitation.

159
00:13:09.219 --> 00:13:11.921
The bubbles expand and contract over these short pulses.

160
00:13:11.942 --> 00:13:12.442
This is incredible.

161
00:13:12.462 --> 00:13:14.283
And they create mechanical strain.

162
00:13:15.864 --> 00:13:16.765
It's very, very cool.

163
00:13:17.105 --> 00:13:18.607
It's such an interesting, you know...

164
00:13:19.627 --> 00:13:28.089
even conceptually as an idea, to think to do so as a sort of clinical approach.

165
00:13:29.529 --> 00:13:29.989
Exactly.

166
00:13:30.430 --> 00:13:32.930
So you've been able to create a bubble, right?

167
00:13:33.450 --> 00:13:38.691
The question now becomes, how do you control it really, really well?

168
00:13:39.332 --> 00:13:39.532
Right.

169
00:13:39.572 --> 00:13:41.012
So we were able to...

170
00:13:43.016 --> 00:13:44.859
instantiate this cavitation process.

171
00:13:45.119 --> 00:13:45.299
Yeah.

172
00:13:45.640 --> 00:13:47.041
Now I need to localize this thing.

173
00:13:47.081 --> 00:13:48.143
Yeah, now we need controllability.

174
00:13:48.163 --> 00:13:48.363
Right?

175
00:13:48.583 --> 00:13:52.569
We need really precise controllability for this to be a clinical application.

176
00:13:52.849 --> 00:13:54.471
So that's where we get into the next paper.

177
00:13:54.852 --> 00:14:01.040
This is a paper, again, in IEEE, Explore Histotripsy Beyond the Intrinsic Cavitation Threshold.

178
00:14:01.930 --> 00:14:05.012
using very short ultrasound pulses.

179
00:14:05.052 --> 00:14:06.813
So they call it microtrypsy.

180
00:14:06.833 --> 00:14:08.374
Microtrypsy.

181
00:14:09.135 --> 00:14:09.935
Here's the idea.

182
00:14:10.456 --> 00:14:14.138
So, like, you're using sound, right?

183
00:14:16.940 --> 00:14:19.982
And sound has a particular wavelength.

184
00:14:21.568 --> 00:14:23.650
Even this ultrasound is going to have a particular wavelength.

185
00:14:24.070 --> 00:14:28.753
But my target region could be much smaller than this wavelength.

186
00:14:29.313 --> 00:14:32.775
It's a question about resolution at the end of the day.

187
00:14:33.116 --> 00:14:40.840
Let's say I've got a focused region and I want to put a bunch of pressure amplitude using the sound into this area.

188
00:14:42.001 --> 00:14:44.062
It's largest near the center and then it falls off.

189
00:14:44.583 --> 00:14:44.723
Right?

190
00:14:44.743 --> 00:14:45.023
Right.

191
00:14:45.123 --> 00:14:49.889
So there's going to be a large area where I could produce bubbles and it becomes a stochastic process.

192
00:14:50.310 --> 00:14:55.296
So how do I control where the bubbles happen and where they don't?

193
00:14:56.137 --> 00:15:00.523
I need a very nice turning on and turning off.

194
00:15:01.962 --> 00:15:26.242
mechanism that's also specific again to an xyz position in 3d space yeah yeah but but here's the idea the only thing i can focus really is the sound waves right and the sound waves have an inherent blurriness right so so how do i how do i actually control where the bubbles are happening well it turns out cavitation naturally is a non-linear switch

195
00:15:27.464 --> 00:15:27.724
Okay.

196
00:15:28.204 --> 00:15:30.305
And that's what this paper is about that.

197
00:15:30.465 --> 00:15:33.406
So on the X, here's, we've got a plot from that paper.

198
00:15:33.806 --> 00:15:35.087
And I thought this was very, very cool.

199
00:15:35.527 --> 00:15:37.728
So on the X axis is your pressure.

200
00:15:38.688 --> 00:15:38.928
Okay.

201
00:15:39.168 --> 00:15:41.489
This is the pressure that you're creating using your ultrasound.

202
00:15:41.929 --> 00:15:44.870
And on the Y axis is the cavitation probability.

203
00:15:45.738 --> 00:15:48.201
Notice it is not a linear function.

204
00:15:48.261 --> 00:15:52.545
It's not like at 10 megapascals, there's like 20% probability.

205
00:15:52.685 --> 00:15:53.887
At 20, there's 40.

206
00:15:54.287 --> 00:15:55.488
At 30, there's 60.

207
00:15:55.508 --> 00:15:57.190
And it just linearly increases.

208
00:15:57.270 --> 00:16:01.274
No, there is a regime where it switches on and off.

209
00:16:01.755 --> 00:16:01.895
Mm-hmm.

210
00:16:02.239 --> 00:16:02.539
Okay?

211
00:16:03.079 --> 00:16:09.341
So what you can do, you can't control the spatial extent of your sound waves.

212
00:16:09.941 --> 00:16:11.822
But you can control how loud they are.

213
00:16:13.082 --> 00:16:13.362
Okay?

214
00:16:13.382 --> 00:16:13.542
Yeah.

215
00:16:13.823 --> 00:16:17.304
And if you can control how loud they are, imagine a wave, right, that's massive.

216
00:16:17.804 --> 00:16:22.425
Even, like, take the extreme case of a wave that has a crest that is a meter big.

217
00:16:23.305 --> 00:16:28.607
But if I can control the amplitude of this wave very precisely, how loud this wave is,

218
00:16:30.453 --> 00:16:36.351
The spots, even though the wavelength of my sound wave is like a meter long, the spots where...

219
00:16:37.939 --> 00:16:41.462
the pressure goes above a certain threshold can be as small as I want it to be.

220
00:16:41.902 --> 00:16:44.725
So this is a degree of freedom where we have a level of precision.

221
00:16:45.245 --> 00:16:48.248
And you basically use that as part of your first point of control.

222
00:16:48.448 --> 00:16:52.492
Yeah, because you can't control the wavelength of the sound in some sense, right?

223
00:16:52.512 --> 00:16:56.735
Like you're limited by the technology of our time and by the physics of the thing.

224
00:16:56.775 --> 00:16:59.198
I mean, if you make the sound wave too short, then the bubble...

225
00:16:59.738 --> 00:17:01.219
is not going to happen anymore, right?

226
00:17:01.319 --> 00:17:03.720
And the focusing capability is not going to happen anymore.

227
00:17:03.740 --> 00:17:15.744
So there's level of precision where we can know that we can create this height of the wave, and that height of the wave is a factor in this cavitation process that is a good control to have.

228
00:17:15.784 --> 00:17:19.165
Exactly, and the height of the wave is the x-axis on this plot, right?

229
00:17:19.225 --> 00:17:26.008
And so suppose the threshold is exactly, I don't know, over here it seems like to be 25%.

230
00:17:26.909 --> 00:17:27.970
mega pascals, right?

231
00:17:28.551 --> 00:17:37.126
So if the entire amplitude of the wave is only, let's say, 26 or 27, there's going to be a very tiny region that goes above that 26, right?

232
00:17:37.146 --> 00:17:37.267
27.

233
00:17:37.287 --> 00:17:38.088
Yep.

234
00:17:39.439 --> 00:17:47.528
If, on the other hand, I want to target a very large area, I make the sound super loud, like 40, and then most of the sound wave is going to go above that threshold.

235
00:17:47.948 --> 00:17:48.289
You see?

236
00:17:48.489 --> 00:17:52.033
So I'm controlling the space based on the loudness of the ultrasound.

237
00:17:52.293 --> 00:17:56.958
So if you want to make the speaker go to 11, you're going to take care of everything.

238
00:17:57.038 --> 00:17:57.118
Yes.

239
00:17:57.138 --> 00:18:00.222
But if you only make the speaker go to 4, it's going to be...

240
00:18:01.103 --> 00:18:02.044
It's a smaller area.

241
00:18:02.084 --> 00:18:02.845
It's a smaller area.

242
00:18:02.885 --> 00:18:06.109
And they show this with the next figure that I have here there.

243
00:18:06.169 --> 00:18:13.637
They've got, they tested this idea with red blood cell phantoms where you can clear the clearing can make the damage visible.

244
00:18:14.338 --> 00:18:17.302
And this is again in canine tissues at lower levels.

245
00:18:18.479 --> 00:18:19.961
So that's on the left-hand side.

246
00:18:20.521 --> 00:18:24.224
At lower pressures, the size of the cavitation is very small.

247
00:18:24.985 --> 00:18:27.267
But all I'm doing now is increasing the pressure.

248
00:18:27.908 --> 00:18:34.694
And as I increase the pressure and I'm focusing on this region, I can make the size of my cavitation much larger.

249
00:18:34.754 --> 00:18:37.236
Where the bubbles are happening are now much larger.

250
00:18:37.516 --> 00:18:40.459
And that is actually how I control localization.

251
00:18:40.479 --> 00:18:40.919
Yeah.

252
00:18:41.657 --> 00:18:42.578
So this is interesting.

253
00:18:42.678 --> 00:18:50.824
So I think part of what we're saying is the knob we have that we can twist in this context is do we want small bubble or big bubble?

254
00:18:51.525 --> 00:18:57.009
And depending on the use case, you can have multiple small bubble, multiple big bubble.

255
00:18:57.150 --> 00:19:01.533
But the point is that there's a dial for the size of the cavitation.

256
00:19:01.553 --> 00:19:03.835
I'm using bubble as layman's term.

257
00:19:03.855 --> 00:19:04.295
No, but it is.

258
00:19:04.355 --> 00:19:04.856
It's a bubble.

259
00:19:04.896 --> 00:19:05.076
Yeah.

260
00:19:05.436 --> 00:19:20.833
And we have very – the ability to have sort of fine control over that size is an important factor in, again, having functional control over whatever outcome we want to implement this cavitation towards.

261
00:19:20.973 --> 00:19:21.434
Exactly.

262
00:19:21.494 --> 00:19:24.777
It's like – it reminds me of like the mountain Kilimanjaro.

263
00:19:24.957 --> 00:19:25.178
Mm-hmm.

264
00:19:25.498 --> 00:19:34.206
Kilimanjaro is a massive mountain, the size of states in the United States, several European countries, is a single mountain.

265
00:19:34.626 --> 00:19:38.950
But the amount of ice on that mountain is only at the summit.

266
00:19:39.470 --> 00:19:42.393
And so you're not breaking diffraction.

267
00:19:42.633 --> 00:19:47.417
There's a diffraction limit similar to the diffraction limits that we've talked about on this podcast with light.

268
00:19:47.497 --> 00:20:05.017
where it's like I can't localize the light beam to a certain small amount because of the wavelength of light where in this case similarly the wavelength of sound prevents me from localizing the sound in a certain spot but I can make it loud and I can control the loudness the height of Kilimanjaro for example

269
00:20:05.810 --> 00:20:13.295
And if only the tip is above that threshold, then I can make stuff small enough for my liking.

270
00:20:13.455 --> 00:20:14.536
I thought that was pretty cool.

271
00:20:14.556 --> 00:20:15.216
That's quite nice.

272
00:20:15.336 --> 00:20:15.957
That's quite nice.

273
00:20:16.157 --> 00:20:16.937
Yeah.

274
00:20:17.037 --> 00:20:19.939
So that was that second paper.

275
00:20:19.999 --> 00:20:20.620
Second paper.

276
00:20:21.160 --> 00:20:24.362
Now, our attention has been very local, right?

277
00:20:24.382 --> 00:20:25.383
Like, where are the bubbles?

278
00:20:25.603 --> 00:20:28.164
Which cells are being broken up?

279
00:20:29.245 --> 00:20:33.568
Now let's talk about what happens when we try to target tumors.

280
00:20:33.588 --> 00:20:33.728
Right.

281
00:20:34.430 --> 00:20:34.650
Yeah.

282
00:20:34.790 --> 00:20:35.090
Okay.

283
00:20:35.511 --> 00:20:35.711
Right.

284
00:20:36.111 --> 00:20:36.571
What happens?

285
00:20:36.811 --> 00:20:38.372
What happens to the immune response?

286
00:20:38.472 --> 00:20:41.474
Because this stuff doesn't happen in a vacuum in the body.

287
00:20:42.074 --> 00:20:43.375
You start messing with stuff.

288
00:20:43.815 --> 00:20:45.416
There's the police are going to come.

289
00:20:45.456 --> 00:20:48.037
There's a lot going on and be like, what's going on?

290
00:20:48.277 --> 00:20:51.479
As we talk about a lot, biology is not so simple.

291
00:20:51.559 --> 00:20:51.759
Yeah.

292
00:20:52.079 --> 00:20:54.540
So we're going to, we're going to jump start all the way to 2020.

293
00:20:54.560 --> 00:20:54.800
2020.

294
00:20:55.901 --> 00:21:07.747
This is a paper called Non-thermal Hyptotripsy Tumor Ablation Promotes Abscopal Immune Responses That Enhance Cancer Immunotherapy.

295
00:21:08.347 --> 00:21:19.272
So it turns out that using these cavitation bubbles to take care of tumors actually overpowers the immune response that is innate within cancer.

296
00:21:21.524 --> 00:21:34.258
I think what's interesting too here, just maybe taking a step back, is immunotherapies are one tool in the toolbox to combat cancer and tumors.

297
00:21:35.987 --> 00:21:43.017
What we're saying here is that by using this methodology of cavitation, we're actually increasing the efficacy of our hammer.

298
00:21:43.939 --> 00:21:44.099
Yep.

299
00:21:44.319 --> 00:21:45.781
That we're using as a separate tool.

300
00:21:46.282 --> 00:21:49.647
And so there's this increased correlation effect.

301
00:21:49.747 --> 00:21:50.688
Yeah.

302
00:21:51.069 --> 00:21:51.690
That it…

303
00:21:53.273 --> 00:22:06.117
It has a direct functional output, but it also has this other benefit to other clinical therapies that we're using to solve, in particular, this issue around cancer.

304
00:22:06.257 --> 00:22:07.457
Exactly.

305
00:22:07.697 --> 00:22:09.098
And I want to talk about why that is.

306
00:22:09.118 --> 00:22:09.438
That's so good.

307
00:22:09.458 --> 00:22:12.579
Because I was thinking, like, you know, we've got the ultrasound therapies as well.

308
00:22:12.599 --> 00:22:13.279
I was talking about, right?

309
00:22:13.299 --> 00:22:13.419
Yeah.

310
00:22:13.439 --> 00:22:16.000
Where you just heat up the tumor and then you burn it off.

311
00:22:16.320 --> 00:22:16.440
Yeah.

312
00:22:16.460 --> 00:22:16.560
Yeah.

313
00:22:17.360 --> 00:22:19.001
How come this is somehow better?

314
00:22:19.422 --> 00:22:22.724
Because we have two things that both affect tumor structures.

315
00:22:22.764 --> 00:22:24.084
And they both use sound, ultrasound.

316
00:22:24.105 --> 00:22:26.086
And they both use sound, but why is it that this...

317
00:22:26.406 --> 00:22:27.607
This particular one is better.

318
00:22:27.727 --> 00:22:27.967
Okay.

319
00:22:28.167 --> 00:22:33.470
Well, it's that cooking versus disintegrating analogy that I was talking about.

320
00:22:34.251 --> 00:22:38.393
With the traditional ultrasound therapies, you heat up a tumor, right?

321
00:22:38.873 --> 00:22:43.516
And when you heat that up, that's going to burn off the tumor cells, the cancerous cells.

322
00:22:47.415 --> 00:22:52.058
You're going to break cells apart, but the components of those cells are still around.

323
00:22:52.218 --> 00:22:52.899
Yeah.

324
00:22:52.919 --> 00:22:55.760
And they're going to be available for immune recognition.

325
00:22:57.201 --> 00:22:57.982
You see what I'm saying?

326
00:22:58.242 --> 00:23:00.043
We're not deleting evidence from the crime scene.

327
00:23:00.063 --> 00:23:01.084
Yeah, we're not melting it.

328
00:23:01.144 --> 00:23:02.685
We're just like pulverizing it.

329
00:23:03.045 --> 00:23:07.508
But the immune cells can still, the police can still come and look at the evidence and be like...

330
00:23:09.093 --> 00:23:10.654
I probably should have been paying attention.

331
00:23:10.774 --> 00:23:11.595
Yeah, yeah, yeah.

332
00:23:11.755 --> 00:23:12.655
Right?

333
00:23:13.316 --> 00:23:13.896
You know what I mean?

334
00:23:14.136 --> 00:23:20.400
And so some of the components of those tumor cells that are broken down are tumor antigens.

335
00:23:20.880 --> 00:23:25.143
They're molecular features that an immune response is going to learn to target.

336
00:23:25.783 --> 00:23:28.905
On the other hand, if you just burn it off, those are now denatured proteins.

337
00:23:29.486 --> 00:23:32.748
The immune cells don't really know what they're looking at.

338
00:23:32.828 --> 00:23:36.470
It's the melted steel versus like intact stuff.

339
00:23:36.790 --> 00:23:41.833
Would an analogy for this be we have undercover immune agents embedded everywhere.

340
00:23:42.493 --> 00:23:49.237
When we burn them off, the undercover immune agents, when the immune response comes to the crime scene, they don't see them.

341
00:23:49.417 --> 00:23:51.838
And so they're like, our undercovers are gone.

342
00:23:52.238 --> 00:23:53.159
So we need to do something.

343
00:23:53.179 --> 00:23:53.319
Yeah.

344
00:23:54.122 --> 00:24:00.003
With cavitation, we're saying the undercover immune agents, they may not be at the spot they were assigned to, but they're in the vicinity.

345
00:24:00.483 --> 00:24:02.324
And there's like pieces of them.

346
00:24:02.444 --> 00:24:05.964
And you can still be, okay, okay, it's still here.

347
00:24:06.625 --> 00:24:11.266
And so I'm not going to overreact in my response.

348
00:24:11.326 --> 00:24:13.806
Yeah, but I'm at least going to learn what's going on here.

349
00:24:13.966 --> 00:24:14.586
Because I can pick it up.

350
00:24:14.606 --> 00:24:16.167
And I am going to heighten my response.

351
00:24:16.467 --> 00:24:16.647
I see.

352
00:24:16.667 --> 00:24:17.267
I'm going to like...

353
00:24:18.307 --> 00:24:42.445
be more vigilant that's a good distinction you know what i mean yeah yeah yeah no that's a good yeah exactly rather than being eviscerated yeah from the from the scene there's presence and you can then glean what your next action should be yeah because there's some presence because a lot of the immune system are detectives right they're not just like like dumb beat cops there's a lot of people that have like there's a lot of detectives that have memory that have

354
00:24:43.282 --> 00:24:51.328
the ability to take antigens, recognize them for the rest, so that if the tumor shows up elsewhere or something, and that's something that we're going to get to.

355
00:24:51.348 --> 00:24:53.710
We got some immune columbos on the scene.

356
00:24:53.730 --> 00:24:54.910
Yeah, yeah, yeah, exactly.

357
00:24:55.191 --> 00:25:02.616
And so this histotripsy, it's associated actually with the infiltration by innate and adaptive immune cells.

358
00:25:03.133 --> 00:25:07.598
in the treated tumor, including tumor-reactive T-cell responses.

359
00:25:07.618 --> 00:25:11.021
So in the control, you have just normal ablation.

360
00:25:11.802 --> 00:25:20.431
And when you use histotripsy, the concentration of these T-cell receptors, the CD8 plus T-cells,

361
00:25:22.023 --> 00:25:22.684
they go up.

362
00:25:23.885 --> 00:25:26.327
This is a CD4 T cells.

363
00:25:26.667 --> 00:25:34.214
Like all of these guys that are like geared to learn about what the tumor is and make sure that that tumor does not show up elsewhere.

364
00:25:34.234 --> 00:25:37.657
It's going up.

365
00:25:38.377 --> 00:25:43.101
It's so interesting because I don't know that this outcome is obvious.

366
00:25:43.221 --> 00:25:43.542
Mm-mm.

367
00:25:44.652 --> 00:26:04.448
you know but it's not i mean now that i think about it sure right but like in 2000 when when you know like fiber however many 10 10 years ago when like this cavitation was being used not obvious at all i think and it's really it it's such a nice also because again we always talk about how

368
00:26:05.773 --> 00:26:13.539
there's a building block process to get to ultimately something that is viewed as practical or functional for humanity.

369
00:26:13.579 --> 00:26:16.101
And it's like, hey, look, we have this new functionality.

370
00:26:16.761 --> 00:26:20.184
Just because you have functionality doesn't mean it has practical application.

371
00:26:20.204 --> 00:26:23.727
Yeah, and this has more practical applications than what you even started out with.

372
00:26:23.787 --> 00:26:26.569
I mean, you started out with just like, why is the sound bad?

373
00:26:26.829 --> 00:26:31.353
Into, okay, let's like try to burn a hole in the left ventricle.

374
00:26:31.373 --> 00:26:31.673
Mm-hmm.

375
00:26:31.973 --> 00:26:47.965
to wait this can supercharge cancer cancer therapies right right right that are already existing and what what folks who are dealing with this have to do yeah to stand a fighting chance and you're basically saying oh yeah oh by the way yeah

376
00:26:49.125 --> 00:26:53.147
it's going to make the thing you're already doing, which is, it's tough on the, it's complicated.

377
00:26:53.168 --> 00:26:58.990
There's a lot, but it's, it's going to make it more effective that I'm, I'm just, I did not expect that outcome.

378
00:26:59.131 --> 00:26:59.351
Yeah.

379
00:26:59.551 --> 00:27:00.551
I didn't expect this one either.

380
00:27:00.771 --> 00:27:07.495
And just one more figure from this paper is, it shows that the effects on the tumors outside of the treated site.

381
00:27:07.955 --> 00:27:08.175
Okay.

382
00:27:08.375 --> 00:27:13.478
So you're treating with cavitation with this histotripsy in a certain spot.

383
00:27:13.978 --> 00:27:17.140
And now you look at pulmonary metastases.

384
00:27:18.160 --> 00:27:40.076
elsewhere okay so so lung cancer happening maybe in the other lung or like there's this cancer happening here did it travel to the lung right um how is it moving out because the immune response has been supercharged right the the ones that are treated with the histotripsy versus the control

385
00:27:41.802 --> 00:27:42.722
Night and day difference.

386
00:27:42.822 --> 00:27:50.124
The ones that are treated with histotripsy have a much higher suppression of these pulmonary metastases.

387
00:27:52.085 --> 00:27:58.847
The immune response sets up a tighter perimeter, which then prevents the spread more than if it had not happened at all.

388
00:27:59.047 --> 00:27:59.267
Yeah.

389
00:27:59.307 --> 00:28:00.927
I mean, this is really, really cool.

390
00:28:00.987 --> 00:28:03.468
I thought this was the most interesting conceptual turn.

391
00:28:03.888 --> 00:28:07.149
And then like a bubble cloud has become a biological instrument.

392
00:28:07.169 --> 00:28:07.689
Yeah.

393
00:28:07.749 --> 00:28:08.309
Right.

394
00:28:08.329 --> 00:28:10.770
In the first paper, it was just like, okay, we want to remove tissue.

395
00:28:10.790 --> 00:28:10.890
Yeah.

396
00:28:11.610 --> 00:28:20.996
Here, the same kind of mechanical intervention can change the information available to the immune cells throughout the organism in the entire body.

397
00:28:21.317 --> 00:28:27.841
So this is like a definition of like a precise surgical tool that is doing way more than what we paid for.

398
00:28:28.581 --> 00:28:32.764
Way more than just the initial precise surgical outcome that you were looking for.

399
00:28:32.884 --> 00:28:33.285
That's right.

400
00:28:33.985 --> 00:28:41.091
So now we get to our fourth paper in 2021, transcranial magnetic resonance guided histotripsy for brain surgery.

401
00:28:41.111 --> 00:28:42.592
We are not just done with cancer.

402
00:28:42.612 --> 00:28:43.833
We're going to go to brain surgery.

403
00:28:44.194 --> 00:28:47.116
This is a preclinical investigation by Zhu and her colleagues.

404
00:28:47.156 --> 00:28:51.640
They tested whether they could make controlled lesions behind pig brains

405
00:28:52.718 --> 00:28:57.499
inside pig brains while transmitting through an excised human skull.

406
00:28:57.539 --> 00:29:01.140
This is a human skull that has been donated for medical research.

407
00:29:02.000 --> 00:29:07.201
So this is a preclinical experiment and the geometry is worth explaining.

408
00:29:07.241 --> 00:29:11.962
So you've got the pig skull which is the model organism and there's a human skull around it.

409
00:29:12.022 --> 00:29:19.103
So we want to test whether this ultrasound can go through the human skull and actually do brain surgery on neural tissue inside.

410
00:29:19.963 --> 00:29:23.284
The skull is notoriously a terrible interface.

411
00:29:23.704 --> 00:29:23.964
Okay.

412
00:29:24.364 --> 00:29:26.825
I dealt with this as a PhD student, right?

413
00:29:26.845 --> 00:29:31.206
Like if you do, if you do just ECOG, ECOG data is terrible.

414
00:29:31.486 --> 00:29:32.066
Okay.

415
00:29:32.806 --> 00:29:33.407
Look, I'm sorry.

416
00:29:33.847 --> 00:29:34.687
It's not terrible.

417
00:29:34.907 --> 00:29:35.647
It's not terrible.

418
00:29:35.747 --> 00:29:44.489
But like I'm an e-fiz guy, which means like it's invasive, but I had to do it on rodents, obviously, because I wanted high fidelity, 40 kilohertz data.

419
00:29:44.810 --> 00:29:45.370
I want 40,000.

420
00:29:46.650 --> 00:29:50.851
data points per second so that I can localize single neurons and things like that.

421
00:29:50.951 --> 00:29:51.211
Right.

422
00:29:51.271 --> 00:29:53.352
You wanted 4k, you're getting 480p.

423
00:29:53.472 --> 00:29:53.772
Yeah.

424
00:29:54.072 --> 00:30:03.734
Like, no, it's like a 20 and really like ECOG data, like the stuff that you see of like the, you know, patients with electrodes on their skull.

425
00:30:04.795 --> 00:30:12.917
Even if you had, even, even if you put that ECOG wires on a data acquisition system that was taking data at 40,000,

426
00:30:14.130 --> 00:30:21.949
hertz, like 40,000 data points a second, it would be a total waste because all frequencies higher than like 40 hertz are

427
00:30:23.600 --> 00:30:25.141
Don't pass through the brain.

428
00:30:25.221 --> 00:30:28.201
So you would just see giant waves like this, right?

429
00:30:28.502 --> 00:30:35.204
Any blip that is like a millisecond long, which is about the size of an action potential, is not going to make it through the skull.

430
00:30:35.504 --> 00:30:39.125
So there's no point in like taking data at that fast of a rate.

431
00:30:39.605 --> 00:30:48.008
And I think part of the – so what I'm understanding, part of the point here is ultimately if we want to take this from the experimental environment to being functional for humans –

432
00:30:49.327 --> 00:30:51.188
Anything that's passing through a skull matters.

433
00:30:51.829 --> 00:30:57.592
And the human skull specifically is the barrier that we need to make sure that this actually gets through.

434
00:30:57.852 --> 00:30:59.613
I mean, there's a very good reason, right?

435
00:31:00.293 --> 00:31:04.796
The brain is a nice thing that we have evolved over the last 10 million years.

436
00:31:04.976 --> 00:31:08.158
And the skull has been evolved to take care of that brain.

437
00:31:08.298 --> 00:31:08.678
Okay.

438
00:31:08.758 --> 00:31:11.280
So, like, I'm not saying the skull is a bad thing.

439
00:31:11.300 --> 00:31:14.742
I'm just saying it's a little bit annoying for basic science research.

440
00:31:15.322 --> 00:31:38.422
and for basic medical procedures so the skull is a difficult electronic device as i just told you but it's also difficult acoustically for the same reason so like ultrasound is there's going to be impedance from the skull and so and so there's going to be problems right it attenuates sound and the thickness and the internal structure vary across the beam like there's skulls can be thick here and and thinner on other parts

441
00:31:38.711 --> 00:31:42.254
So different paths through the bone delay the wave by different amounts.

442
00:31:42.274 --> 00:31:47.038
You're going to get like weird diffraction signatures when the ultrasound goes through the brain.

443
00:31:47.078 --> 00:31:51.341
So all of that has to be dealt with.

444
00:31:51.441 --> 00:31:56.525
These are engineering problems as much as biological ones if you want to use ultrasound.

445
00:31:56.785 --> 00:31:59.067
And precision matters as we've discussed.

446
00:31:59.327 --> 00:31:59.767
Exactly.

447
00:32:00.208 --> 00:32:03.170
So what these guys did was they used 128 element 700 kilohertz array.

448
00:32:03.190 --> 00:32:03.330
Okay.

449
00:32:03.350 --> 00:32:03.630
Okay.

450
00:32:07.073 --> 00:32:09.475
Compatible with a three Tesla MRI scanner.

451
00:32:09.535 --> 00:32:14.299
So they've got an MRI scanner at the same time that they're administering ultrasound.

452
00:32:14.439 --> 00:32:14.699
Why?

453
00:32:15.019 --> 00:32:18.622
Because the ultrasound, they're trying to mess with the brain that's on the inside.

454
00:32:18.982 --> 00:32:20.744
The MRI scanner is their read.

455
00:32:21.504 --> 00:32:26.568
They want to see what it was like before, and then they mess with it with the ultrasound, what it's like after.

456
00:32:26.928 --> 00:32:28.790
Can I localize my pulses, right?

457
00:32:29.590 --> 00:32:30.351
Turns out they can.

458
00:32:30.371 --> 00:32:30.471
Okay.

459
00:32:31.195 --> 00:32:35.859
They had successful treatment in eight pig brains, and the MRI treatment regions show that.

460
00:32:36.460 --> 00:32:39.943
They actually also did histology, which is the real ground truth, I think.

461
00:32:40.183 --> 00:32:46.908
Histology meaning you take out the brain, you put it into slices, you stain it, and then you can really see what's happening in the tissue, right?

462
00:32:46.948 --> 00:32:48.950
And this is a slice of that brain.

463
00:32:49.370 --> 00:32:51.812
And there's clear demarcation, okay?

464
00:32:51.952 --> 00:32:55.956
The D is the disrupted debris from my target.

465
00:32:56.516 --> 00:32:57.157
Oh my goodness.

466
00:32:57.177 --> 00:32:59.198
The eye is the intact neurons.

467
00:32:59.298 --> 00:33:02.320
And you can see just how sharp the interface is.

468
00:33:02.400 --> 00:33:02.801
Yeah.

469
00:33:03.141 --> 00:33:03.341
Right?

470
00:33:03.661 --> 00:33:10.746
Between the part of the brain that you wanted to ablate and sort of disrupt and the part that you wanted to keep intact.

471
00:33:11.126 --> 00:33:12.427
The scale bar there is 100 microns.

472
00:33:13.668 --> 00:33:14.288
This is unbelievable.

473
00:33:14.308 --> 00:33:14.728
Incredible.

474
00:33:14.968 --> 00:33:19.909
I mean, the thickness of that black line there, I would say tops is 10 microns.

475
00:33:19.949 --> 00:33:22.150
I'd even say maybe one to five.

476
00:33:22.730 --> 00:33:24.990
That's the size of large neurons.

477
00:33:25.450 --> 00:33:27.131
Okay, maybe groups of small neurons.

478
00:33:27.631 --> 00:33:41.474
And so we're now getting conformational data that are the level of precision when going through a human skull of this ultrasound approach has micron level...

479
00:33:42.394 --> 00:33:43.195
Thick.

480
00:33:43.635 --> 00:33:44.115
Precision.

481
00:33:44.175 --> 00:33:44.716
Precision.

482
00:33:45.276 --> 00:33:46.517
That's right.

483
00:33:46.617 --> 00:33:47.698
And we built this, right?

484
00:33:47.718 --> 00:33:49.479
Because in the beginning, we were just doing bubbles.

485
00:33:49.799 --> 00:33:54.603
And then we built, okay, how do we make sure that the precision is there?

486
00:33:54.643 --> 00:33:57.105
Well, we can actually tune the amplitude of the pressure wave.

487
00:33:57.485 --> 00:33:59.186
And now we can actually do it within the skull.

488
00:33:59.226 --> 00:34:04.110
We can deal with all of the diffraction and all of the weird wave stuff that's happening on the skull.

489
00:34:04.130 --> 00:34:04.350
Yeah.

490
00:34:04.590 --> 00:34:10.112
And it seems like the skull is a region because obviously we've targeted different parts of the body over the course of this discussion.

491
00:34:10.412 --> 00:34:15.094
The heart, we're going to get to other places, but the brain now is another.

492
00:34:15.774 --> 00:34:21.556
And the TSA of the brain is pretty tough to get through.

493
00:34:21.636 --> 00:34:22.857
Yes, exactly.

494
00:34:25.258 --> 00:34:29.262
One other thing I want to just touch on here before we move on to our fifth paper.

495
00:34:30.023 --> 00:34:34.927
Why use histotripsy rather than just heating for this job?

496
00:34:35.688 --> 00:34:38.511
Well, it's in the name itself.

497
00:34:38.551 --> 00:34:39.472
You don't want to heat the brain.

498
00:34:40.193 --> 00:34:41.794
The brain is incredibly conductive.

499
00:34:42.515 --> 00:34:43.576
There's fluid everywhere.

500
00:34:43.676 --> 00:34:47.380
And so if you heat part of the brain, that heat is going to travel.

501
00:34:48.099 --> 00:34:49.682
Thanks to the Navier Stokes sequence.

502
00:34:49.702 --> 00:34:51.164
Which we talked about.

503
00:34:51.204 --> 00:34:53.488
Tune into our last episode for that one.

504
00:34:53.508 --> 00:34:59.438
There's going to be a lot more Navier Stokes when we talk about Sidney and his physics stuff.

505
00:34:59.758 --> 00:35:00.279
But for now...

506
00:35:01.180 --> 00:35:04.922
Because fluid can sort of transport energy very well, that's a really bad thing.

507
00:35:05.282 --> 00:35:09.444
On the other hand, histotripsy, the pulses can have a very low duty cycle.

508
00:35:10.144 --> 00:35:13.166
A duty cycle is, imagine you've got a sine wave, right?

509
00:35:13.506 --> 00:35:15.807
A sine wave is exactly half duty cycle.

510
00:35:15.847 --> 00:35:16.967
It's one up, one down.

511
00:35:17.548 --> 00:35:18.228
Just up and down.

512
00:35:19.048 --> 00:35:19.869
But you can imagine...

513
00:35:22.622 --> 00:35:24.163
that's a really small duty cycle.

514
00:35:24.483 --> 00:35:29.606
Because the amount of time that something is on is very short compared to the amount of time something's off.

515
00:35:29.907 --> 00:35:31.408
The frequency can still be the same.

516
00:35:32.494 --> 00:35:39.519
But the part of the wave that is above threshold and the part of the wave that is below threshold, that ratio is called the duty cycle.

517
00:35:39.739 --> 00:35:48.786
And for these cavitation experiments, the duty cycle can be very low, which means that you're only actively sending this high-pressure ultrasound not that long.

518
00:35:49.267 --> 00:35:49.507
Right?

519
00:35:49.627 --> 00:35:49.887
Yeah.

520
00:35:50.067 --> 00:35:52.449
And so you're not heating up the skull that much.

521
00:35:53.089 --> 00:35:55.091
It circumnavigates the skull heating problem.

522
00:35:57.597 --> 00:36:04.900
And you can really target that mechanical disruption that you're talking about without really heating up the brain tissue in the first place.

523
00:36:04.960 --> 00:36:09.482
The idea is a longer duty cycle would mean we're doing too much, causing too much chaos.

524
00:36:09.522 --> 00:36:12.243
Chaos creates heat, creates more destruction than intended.

525
00:36:12.583 --> 00:36:19.506
So the fact that we do have a short duty cycle is actually a benefit to the functionality that we're trying to accomplish.

526
00:36:19.886 --> 00:36:20.666
Yeah, exactly.

527
00:36:21.767 --> 00:36:22.227
And so...

528
00:36:23.378 --> 00:36:32.504
There's still something that is lacking here, which is fully like now we're trying to do it on straight up human beings.

529
00:36:32.684 --> 00:36:32.985
Yeah.

530
00:36:33.245 --> 00:36:33.525
Okay?

531
00:36:33.865 --> 00:36:34.005
Yeah.

532
00:36:34.085 --> 00:36:38.088
Like we've proven it in all these model organisms.

533
00:36:38.528 --> 00:36:43.552
We've proven it with sort of a half model organism and a donated human skull.

534
00:36:43.972 --> 00:36:46.674
But now can we actually do this very precisely?

535
00:36:46.974 --> 00:36:49.336
Can we compensate for the variable acoustic medium?

536
00:36:49.676 --> 00:36:52.318
The brain, you know, once you get past the skull, the brain is kind of,

537
00:36:53.467 --> 00:36:53.947
Same, same.

538
00:36:54.687 --> 00:36:57.188
But other parts of the body, there's organs everywhere.

539
00:36:57.308 --> 00:36:57.508
Yeah.

540
00:36:57.828 --> 00:36:58.008
Right?

541
00:36:58.088 --> 00:36:59.808
So there's going to be variable acoustic medium.

542
00:37:00.088 --> 00:37:04.309
Can you maintain a useful pressure focus and generate bubbles exactly where you intend?

543
00:37:04.329 --> 00:37:05.409
That's actually a really good point.

544
00:37:06.069 --> 00:37:09.050
All these environments have very different variables you're dealing with.

545
00:37:09.130 --> 00:37:09.370
Yeah.

546
00:37:09.510 --> 00:37:09.730
Yeah.

547
00:37:10.410 --> 00:37:14.251
And that's why the last paper that I'm going to focus on is the one in the liver.

548
00:37:14.871 --> 00:37:16.031
This is in 2024.

549
00:37:16.491 --> 00:37:17.491
It's a clinical paper.

550
00:37:17.811 --> 00:37:22.572
So this is, you know, now we're getting into like human clinical trials.

551
00:37:23.878 --> 00:37:33.124
The HOPE4Liver Single Arm Pivotal Trial for Histotripsy of Primary and Metastatic Liver Tumors.

552
00:37:33.544 --> 00:37:34.365
She was a co-author.

553
00:37:34.385 --> 00:37:36.006
This is a multi-institution study.

554
00:37:36.626 --> 00:37:38.707
And the engineering now has changed form.

555
00:37:39.668 --> 00:37:43.030
We started from just making holes to now we're doing...

556
00:37:43.050 --> 00:37:44.491
There's a clinical transducer.

557
00:37:44.571 --> 00:37:47.093
There's an ultrasound imaging probe for targeting.

558
00:37:47.113 --> 00:37:47.433
Yeah.

559
00:37:47.793 --> 00:37:50.156
There's a software to describe the treatment volume.

560
00:37:50.736 --> 00:37:54.400
There's a robotic arm that moves the focus through that volume.

561
00:37:54.680 --> 00:37:55.221
This is crazy.

562
00:37:55.241 --> 00:38:02.028
They've got like the thing that is doing the ultrasound is like moving around in a robotic arm to like target stuff.

563
00:38:02.128 --> 00:38:08.655
And each tiny focal treatment has to be assembled into a three-dimensional area that covers a tumor that is in the liver.

564
00:38:10.248 --> 00:38:14.629
So we're now getting to let's make this work for people.

565
00:38:14.769 --> 00:38:15.189
Exactly.

566
00:38:15.229 --> 00:38:15.429
Yeah.

567
00:38:15.909 --> 00:38:18.050
So there's like margins that you have to plan.

568
00:38:18.070 --> 00:38:20.050
There's nearby structures that you got to avoid.

569
00:38:20.350 --> 00:38:22.950
This is now an intact like human being.

570
00:38:23.110 --> 00:38:23.371
Yeah.

571
00:38:23.551 --> 00:38:23.771
Okay.

572
00:38:23.791 --> 00:38:24.771
This is a real deal.

573
00:38:25.211 --> 00:38:36.193
Like and the technical success, the definition of technical success is whether the resulting treatment zone completely covered the intended tumor volume.

574
00:38:37.857 --> 00:38:42.198
on the contrast-enhanced CT or the MRI within 36 hours.

575
00:38:42.258 --> 00:38:48.359
Like, I do the treatment, and within 36 hours, when I take the MRI, do I see the effect, right?

576
00:38:49.359 --> 00:38:50.039
Well, let's see.

577
00:38:50.139 --> 00:38:51.239
Let's see if we do.

578
00:38:51.799 --> 00:38:52.439
In fact, we do.

579
00:38:52.860 --> 00:38:55.080
So this is figure four from their paper.

580
00:38:55.740 --> 00:39:02.601
It shows a 1.3-centimeter metastatic colorectal cancer in segment three of the liver.

581
00:39:02.621 --> 00:39:04.441
That's the white arrow on the top left.

582
00:39:04.942 --> 00:39:06.322
There's a little tiny tumor there.

583
00:39:07.042 --> 00:39:07.302
Okay?

584
00:39:08.638 --> 00:39:09.659
These are MRI scans.

585
00:39:10.220 --> 00:39:13.743
So in B, this is 36 hours post-procedure.

586
00:39:14.624 --> 00:39:19.168
The treatment zone is in the solid arrow encompassing the side of the tumor, and there's a margin.

587
00:39:20.068 --> 00:39:25.613
Crucially, the smaller arrow there, that's a blood vessel that traverses the treatment zone.

588
00:39:26.334 --> 00:39:27.475
It hasn't been passed.

589
00:39:29.131 --> 00:39:29.691
Yeah.

590
00:39:29.891 --> 00:39:30.632
Yeah.

591
00:39:30.912 --> 00:39:31.112
Okay.

592
00:39:31.132 --> 00:39:32.313
The traffic is still flowing.

593
00:39:32.413 --> 00:39:33.734
Yeah.

594
00:39:33.794 --> 00:39:36.555
We just took apart the riffraff around it.

595
00:39:36.715 --> 00:39:39.477
But the traffic of the blood is still flowing through that.

596
00:39:40.057 --> 00:39:42.778
And then in part C, this is 30 days post-procedure.

597
00:39:43.078 --> 00:39:49.082
It shows involution of the treatment zone, meaning the giant hole that you made has gotten smaller.

598
00:39:49.482 --> 00:39:53.044
There is recovery happening within the tissue.

599
00:39:53.124 --> 00:39:53.244
Yeah.

600
00:39:53.264 --> 00:39:53.444
Yeah.

601
00:39:54.245 --> 00:40:00.668
Yeah, the region we've targeted is now repairing, and we still have our traffic flow.

602
00:40:00.828 --> 00:40:01.809
Yes, exactly.

603
00:40:03.769 --> 00:40:05.050
It's an incredible story.

604
00:40:05.630 --> 00:40:06.031
Yeah.

605
00:40:06.591 --> 00:40:10.893
I mean, it's incredible for, you know, obviously so many people in every, you know,

606
00:40:11.513 --> 00:40:23.441
Many listeners are less than two or three degrees of separation away from someone who deals with or has suffered from cancer actively or in the past.

607
00:40:23.681 --> 00:40:23.901
Yeah.

608
00:40:24.241 --> 00:40:26.863
And obviously this has been an area that has had.

609
00:40:28.718 --> 00:40:34.983
incredible amounts of human scientific research as well as financial backing to try to find ways to deal with.

610
00:40:35.423 --> 00:40:51.654
There's so many interesting aspects of this area of research that, again, the construction, the building of, oh, hey, this methodology has functional benefit.

611
00:40:52.334 --> 00:40:52.735
Oh, hey,

612
00:40:58.352 --> 00:41:03.617
Oh, hey, we can actually do it in human-like contexts.

613
00:41:04.337 --> 00:41:19.811
Oh, and now when we get to the point where we have to do this in a very precise way for clinical outcomes, when we look to see the results, we are seeing that what we thought was maybe possible at the beginning of this journey is proving to be true.

614
00:41:20.011 --> 00:41:20.211
Yeah.

615
00:41:21.583 --> 00:41:22.163
It's incredible.

616
00:41:22.463 --> 00:41:24.624
And I mean, I don't want to oversell this thing.

617
00:41:25.104 --> 00:41:32.685
I do want to, I do want to say that, you know, cancer is an extremely multifaceted disease, ailment.

618
00:41:34.085 --> 00:41:36.926
You know, 95%, they had 95% success.

619
00:41:37.306 --> 00:41:42.827
Three out of the 44 patients actually did have major complications related to the surgery.

620
00:41:42.867 --> 00:41:46.568
So something like 7% had major, major complications.

621
00:41:48.408 --> 00:41:48.648
And yeah,

622
00:41:50.982 --> 00:41:57.144
One of the limitations of this is that you need to be able to image this tumor in the first place very, very well.

623
00:41:57.524 --> 00:42:03.567
There's some tumors where it's like impossible to image because based on their placement and where they are and so on and so forth, right?

624
00:42:03.867 --> 00:42:07.888
So, but in its limited capacity, it could be doing wonders, right?

625
00:42:07.908 --> 00:42:09.409
It could be the next step.

626
00:42:09.509 --> 00:42:14.691
And the treatment for cancer in the future is never going to be a one- Right.

627
00:42:16.327 --> 00:42:18.010
one thing rules them all, right?

628
00:42:18.131 --> 00:42:23.340
Unless we finally, like, everyone gets genetic engineering somehow, right?

629
00:42:23.500 --> 00:42:26.385
Like, and we all live in Gattaca, and it's, like, mandatory.

630
00:42:27.187 --> 00:42:28.629
Other than that scenario, like...

631
00:42:29.548 --> 00:42:32.009
Dealing with cancer is going to be a multifaceted.

632
00:42:32.349 --> 00:42:33.970
There's going to be immunotherapy drugs.

633
00:42:33.990 --> 00:42:38.552
There's going to be like chemo is always going to be around, I think.

634
00:42:39.012 --> 00:42:42.454
But more and more tools are necessary, right?

635
00:42:42.494 --> 00:42:44.415
And more and more tricks that we have.

636
00:42:44.835 --> 00:42:46.676
The better tricks we have, the better it is.

637
00:42:47.176 --> 00:42:47.836
In 2023, histosonics...

638
00:42:50.616 --> 00:42:55.459
received US FDA de novo clearance for destruction of liver tumors based on this study.

639
00:42:55.699 --> 00:43:00.903
Because even that 7%, that was like lower than the threshold that is required for the FDA to approve something.

640
00:43:01.663 --> 00:43:09.328
And this marked a clinical application of this technology developed through CHU's collaboration with these engineers, physicians, and many other researchers.

641
00:43:09.948 --> 00:43:14.191
This 2024 trial helps explain the evidence behind that step, right?

642
00:43:14.291 --> 00:43:17.193
And this explains why now the FDA is approving this.

643
00:43:18.627 --> 00:43:27.893
In 2025, Histosonics announced a management-led transaction, valuing the company at approximately $2.25 billion.

644
00:43:27.953 --> 00:43:29.153
That's billion with a B.

645
00:43:29.173 --> 00:43:30.454
With a B. Yeah.

646
00:43:30.554 --> 00:43:31.855
And that's the company evaluation.

647
00:43:32.135 --> 00:43:36.197
I don't think it's a measure of any scientist personally and how much they earned.

648
00:43:36.298 --> 00:43:40.800
I don't think Zhu earned a giant significant portion of that.

649
00:43:42.221 --> 00:43:44.462
Which is problematic, but that's for another episode.

650
00:43:44.482 --> 00:43:45.423
Yeah, exactly.

651
00:43:45.763 --> 00:43:53.887
But it's one sign that, you know, this approach that can begin as a difficult laboratory question, it now sits as a real medical and commercial ecosystem.

652
00:43:55.148 --> 00:43:57.149
There are so many interesting aspects of this story.

653
00:43:57.249 --> 00:44:10.436
And again, part of what the Golden Goose Awards tries to do, as you've seen in our first story here through the progression, is when we look at and we've talked about federal funding for basic research on this podcast quite a bit.

654
00:44:12.358 --> 00:44:14.539
There is a difference between basic research and applied research.

655
00:44:15.440 --> 00:44:22.425
We're trying to increase our understanding because from that, we are so brilliant as human beings, so creative.

656
00:44:22.945 --> 00:44:30.391
We have an ingenuity that when we find these unlocks, as we've talked about, we will find the practical path.

657
00:44:31.192 --> 00:44:31.352
Yeah.

658
00:44:32.102 --> 00:44:41.847
This has led to for in terms of folks focus on the economic outcomes of this investment of taxpayer dollars towards some of this stuff.

659
00:44:42.387 --> 00:44:47.110
You know, there's a $2.25 billion acquisition that's happened in the last 12 months.

660
00:44:47.970 --> 00:44:58.475
There are huge real world human benefits from the clinical application of this, which is why it's been valued so high.

661
00:44:59.596 --> 00:44:59.736
And.

662
00:45:00.840 --> 00:45:01.944
It's a platform.

663
00:45:03.144 --> 00:45:06.726
not necessarily just a single functional outcome.

664
00:45:06.906 --> 00:45:16.411
The liver cancer treatment pathway happens to be the pathway that we've gotten to the place where there's commercial and practical applications.

665
00:45:16.851 --> 00:45:25.215
Are there potential practical applications in other areas of cancer and applications for immunotherapy yet to be seen?

666
00:45:26.236 --> 00:45:30.418
But as we kind of discussed, the breadth and scope of potential

667
00:45:31.937 --> 00:45:33.438
It's quite wide.

668
00:45:34.559 --> 00:45:45.287
And it is a perfect starting story for understanding what are the Golden Goose Awards and what are they trying to highlight.

669
00:45:45.607 --> 00:45:56.054
It's this idea that we should not lose our curiosity as we look at these things because applied research and development in the United States –

670
00:45:58.109 --> 00:46:00.830
dwarfs by orders of magnitude basic research.

671
00:46:00.930 --> 00:46:01.151
Yeah.

672
00:46:01.631 --> 00:46:16.058
And had we not had funding for a project like this, as well as the rest of our stories, there are real world benefits that we would have left on the table that were not necessarily obvious in this case because the roommates were complaining about

673
00:46:16.638 --> 00:46:41.482
noise as a former dj i know what that feels like too but i'm not saving people's lives in the way that uh this story really exemplifies it's just it's really incredible and the science is also super interesting on this one um it's super super interesting yeah i will we'll maybe follow up on this in a future show but yeah and i just had an idea for a comment yeah if if you guys are

674
00:46:46.477 --> 00:47:11.749
about the stuff that your lab mates complain about because i've got several yeah just just put down if if in case you're like a phd student or a postdoc in your lab and just the complaints that lab mates do you know because who knows who knows you may both save people's lives and be the recipient of a future golden goose award yeah um

675
00:47:12.438 --> 00:47:13.819
An incredible first story.

676
00:47:13.879 --> 00:47:21.925
Hopefully for our listeners who may not be familiar, you kind of get an understanding now of what these awards really are about.

677
00:47:22.465 --> 00:47:27.409
And before we get into our next story, we're going to take this opportunity to do some quick housekeeping.

678
00:47:28.030 --> 00:47:33.694
For many of you who may be first-time listeners, welcome to the best science show on the planet.

679
00:47:34.574 --> 00:47:39.678
For those of you who are returning, thank you for joining FFP Nation and continuing to support the show.

680
00:47:42.160 --> 00:47:46.701
Any platform you can think of, our videos are both on YouTube and Spotify currently.

681
00:47:47.081 --> 00:47:54.743
For audio listeners who may be listening on Apple, we are waiting for Apple to turn on video, which allegedly is happening before the end of the year.

682
00:47:55.163 --> 00:48:01.605
If you're interested in some of the research papers that we talk about on the show, you can check out our website, ffppod.com.

683
00:48:02.125 --> 00:48:07.212
Every episode page always includes the research papers we discuss in each episode.

684
00:48:07.612 --> 00:48:16.463
There's a lot of great other features on the site, including particularly relevant to this episode, the U.S. Funding Tracker for Research and Development.

685
00:48:16.744 --> 00:48:21.029
And so you can really dig into several different components about research.

686
00:48:22.282 --> 00:48:28.408
decades worth of funding as it relates to R&amp;D across dozens of agencies.

687
00:48:28.428 --> 00:48:30.811
You can see how it's increased or decreased over time.

688
00:48:31.191 --> 00:48:38.758
You can see how the current funding cycle is shaping up for fiscal year 2027 with explainers, graphs, charts.

689
00:48:39.239 --> 00:48:43.523
It's actually a really great way for you to ground yourself in how much are we spending on how much we're spending.

690
00:48:43.863 --> 00:48:45.586
in these areas and where is it going to?

691
00:48:45.987 --> 00:48:51.975
Because interestingly, as we've talked about recently, many people think of the National Science Foundation as like all this money is going there.

692
00:48:52.516 --> 00:48:58.245
And from a research and development perspective, they're actually a very small slice as compared to some other agencies.

693
00:48:58.685 --> 00:49:03.107
I would have you guess which those other agencies may or may not be.

694
00:49:03.467 --> 00:49:13.031
But we really appreciate you all who are listening and joining us, especially for this super special episode in what is going to be an incredible couple of weeks.

695
00:49:13.531 --> 00:49:19.674
We are going to jump in to our second research story for the Golden Goose Awards.

696
00:49:20.454 --> 00:49:28.742
Our next story is tracking holiday lights from space, recognizing NASA's black marble team.

697
00:49:28.782 --> 00:49:36.088
In 2012, researchers studying satellite observations of Cairo noticed an unexpected jump in nighttime brightness.

698
00:49:36.389 --> 00:49:41.533
They traced it to Ramadan when meals, shopping, and social gatherings shifted into the night.

699
00:49:41.593 --> 00:49:46.177
And they also found holiday brightening around Christmas in U.S. cities.

700
00:49:46.918 --> 00:49:51.143
Patterns of human activity were showing up in satellite data.

701
00:49:51.904 --> 00:50:02.458
With NASA's funding support, which is one of the most popular federal agencies currently in the U.S. government, the team's broader work produced Black Marble, a system that turns NASA

702
00:50:02.718 --> 00:50:08.906
nighttime satellite observations into consistent measurements of lighting on the ground.

703
00:50:09.046 --> 00:50:16.374
And the same ability to measure those changes also helps to know, for example, when the lights go out.

704
00:50:16.715 --> 00:50:16.815
And

705
00:50:19.678 --> 00:50:24.923
is comparing Fort Myers before and after Hurricane Ian in 2022.

706
00:50:25.344 --> 00:50:30.569
Areas that had been brightly lit became much darker as the storm knocked out power.

707
00:50:30.609 --> 00:50:33.352
And this kind of makes intuitive sense conceptually.

708
00:50:33.953 --> 00:50:39.438
By following these changes over time, researchers can map outages and track recovery and,

709
00:50:39.498 --> 00:50:42.320
helping identify communities that may need support.

710
00:50:42.440 --> 00:50:49.967
Obviously, if the lights come back on, you don't have to send first responders into that area necessarily because we've now restored power.

711
00:50:50.007 --> 00:51:01.276
But how do you tell whether a place has actually lost power or whether clouds or moonlight or other atmospheric effects are changing what the satellite sees?

712
00:51:01.336 --> 00:51:06.120
This is the setup for our next Golden Goose Award winner.

713
00:51:06.596 --> 00:51:06.996
That's right.

714
00:51:07.036 --> 00:51:16.978
So in September 2017, if we take another specific example, Hurricane Maria crossed Puerto Rico, and the electrical grid failed across much of the island.

715
00:51:17.038 --> 00:51:18.579
And you can see a before and after.

716
00:51:18.739 --> 00:51:21.279
It's almost like the whole island goes dark.

717
00:51:21.319 --> 00:51:23.860
And from orbit, the change is immediate, right?

718
00:51:23.900 --> 00:51:28.201
Places that admitted light night after night suddenly go dark.

719
00:51:29.001 --> 00:51:31.702
Now, scientifically, there's an interesting question about,

720
00:51:32.452 --> 00:51:37.716
And it's not simply whether the island went dark, it's how the power returned, right?

721
00:51:38.196 --> 00:51:43.520
Street after street, month by month, and whether some communities were left waiting longer than others.

722
00:51:44.000 --> 00:51:50.865
Now, that question can't be answered simply by taking a before and after and after and after photo.

723
00:51:51.546 --> 00:51:57.810
Because there's a lot that happens between a satellite getting light from an area and...

724
00:51:58.503 --> 00:52:03.086
and that light being attributed to human activity.

725
00:52:03.646 --> 00:52:03.886
Okay?

726
00:52:04.167 --> 00:52:18.515
So Miguel Ramon, Eleanor Stokes, Xu Shen Wang, Virginia Kolb, and Ranjay Shrestha and their collaborators, they've been working on a way to measure artificial light from space with enough care to answer questions like that.

727
00:52:19.156 --> 00:52:20.897
How much of this is because of humans?

728
00:52:21.797 --> 00:52:23.939
And this is NASA's Black Marble.

729
00:52:24.999 --> 00:52:25.740
It begins with a...

730
00:52:26.464 --> 00:52:29.006
remarkably sensitive satellite camera.

731
00:52:29.746 --> 00:52:49.518
And the real work is actually taking the measurements from the satellite camera and turning it into how much of this is human activity, how much of this is a different moon phase, how much of this is a cloud, a different viewing angle, snow on the ground, all of these other effects, taking all of that out and focusing directly on human activity.

732
00:52:49.962 --> 00:52:57.555
So in the same tradition that we talked about the previous one, we're going to do five papers that are connected to this group.

733
00:52:59.164 --> 00:53:09.847
It's an incredible, incredible set of papers, and we'll keep the distinction in mind throughout, that the satellite measures light, but what we are interested in is human activity, right?

734
00:53:09.887 --> 00:53:16.048
There's a longer history here because military weather satellites have revealed city lights decades before black marble.

735
00:53:16.688 --> 00:53:25.751
There's a joint NASA and NOAA Suomi NPP satellite that had the VIIRS instrument.

736
00:53:25.791 --> 00:53:26.631
This was launched in 2011.

737
00:53:28.214 --> 00:53:30.156
And this has a day and night band.

738
00:53:30.216 --> 00:53:34.201
Its sensitivity made a new class of nighttime observations possible.

739
00:53:34.461 --> 00:53:40.728
And in 2012, we got this striking image of Earth at night.

740
00:53:40.748 --> 00:53:44.292
I remember, actually, this was one of the first big images of Earth at night.

741
00:53:44.372 --> 00:53:49.037
And you could see some very stark differences in.

742
00:53:50.198 --> 00:53:51.218
based on geography.

743
00:53:51.298 --> 00:53:57.020
One of the big ones that was very famous was that if you looked at North Korea versus South Korea, it was massive.

744
00:53:57.040 --> 00:54:02.042
Like South Korea has lit up so much city lights.

745
00:54:02.742 --> 00:54:11.005
And then there's a band which is the demilitarized zone, which is an extremely bright ribbon in between South Korea and North Korea.

746
00:54:11.045 --> 00:54:15.507
And then in North Korea, there's just really one kind of tiny pocket compared to South Korea.

747
00:54:15.787 --> 00:54:21.937
And that tiny pocket is Pyongyang, which is so much smaller than Seoul, which is just across the border.

748
00:54:21.957 --> 00:54:24.922
And you could just see the stark difference between.

749
00:54:25.083 --> 00:54:31.667
between the lighting that tells you so much about the economic activity and about the people and the regimes there, right?

750
00:54:32.928 --> 00:54:35.850
So that was an early, wonderfully specific observation.

751
00:54:36.630 --> 00:54:45.196
And Ramon and Stokes studied how this nighttime lighting changed during Christmas and New Year's in the United States and during Ramadan, as you were saying.

752
00:54:45.776 --> 00:54:52.080
And they showed that lights during Ramadan and Eid, the green in this plot...

753
00:54:52.827 --> 00:54:55.953
is an upscaling of nighttime lights.

754
00:54:56.094 --> 00:55:04.790
And you can see that in Amman, Georgia, and in Cairo, there's a giant upscaling of lights that happen during the nighttime, during Ramadan, and during Eid.

755
00:55:06.140 --> 00:55:11.323
Which shows so much about the people who live there and their ways of life and things like that.

756
00:55:11.383 --> 00:55:16.746
So this already tells you that there's variations between places that these satellite observations might capture.

757
00:55:17.107 --> 00:55:22.390
And these satellite observations can tell us about human activity, the locations of cities.

758
00:55:22.850 --> 00:55:27.052
And it's an open scientific question how far that idea can be taken.

759
00:55:27.113 --> 00:55:27.433
Yeah.

760
00:55:27.593 --> 00:55:27.773
Right?

761
00:55:29.137 --> 00:55:33.101
And so that's where we're going to start with paper number one in 2018.

762
00:55:33.301 --> 00:55:34.222
This is Ramon.

763
00:55:34.982 --> 00:55:35.863
He's the first author.

764
00:55:37.204 --> 00:55:37.805
Roman, sorry.

765
00:55:38.105 --> 00:55:39.126
Roman is the first author.

766
00:55:39.186 --> 00:55:47.434
He was called NASA's Black Marble Nighttime Lights product suite.

767
00:55:47.874 --> 00:55:53.880
It's just a product suite that showcases all of the capabilities that this team has created.

768
00:55:55.070 --> 00:56:00.273
Ultimately, especially as a public agency, right, being backed by NASA, they're always trying to create things.

769
00:56:00.673 --> 00:56:12.280
Not necessarily open source might not be the right word, but creating platforms and tools that can be utilized by whoever for endeavors that are fruitful for their function.

770
00:56:12.420 --> 00:56:12.941
Yeah, exactly.

771
00:56:12.961 --> 00:56:14.502
And this is kind of like a methods paper.

772
00:56:15.662 --> 00:56:21.823
is what we would call it in the business, because this is something that provides a technical foundation for everything else in this episode.

773
00:56:22.183 --> 00:56:24.904
So let's start with what exactly a satellite sees.

774
00:56:25.244 --> 00:56:28.684
Okay, so it's actually a lot more complicated than you would think.

775
00:56:28.764 --> 00:56:31.325
Imagine a patch of ground that's hundreds of meters across.

776
00:56:32.125 --> 00:56:37.926
The day-night band in that satellite collects light from the direction of that patch, right?

777
00:56:37.966 --> 00:56:40.046
Like my sensor is pointing over there.

778
00:56:40.426 --> 00:56:41.686
It's collecting light from there.

779
00:56:42.107 --> 00:56:44.547
Now what arrives in the instrument is,

780
00:56:44.973 --> 00:56:56.888
something called radiance, which roughly speaking is the optical power per unit area, per unit angle in that particular range of wavelengths, whatever range of wavelengths the satellite instrument is sensitive to.

781
00:56:57.340 --> 00:56:59.602
So it's not a count of illuminated houses.

782
00:57:00.203 --> 00:57:03.406
It's not an electricity bill for that patch.

783
00:57:03.866 --> 00:57:07.129
It includes light that's emitted upwards from artificial sources.

784
00:57:07.529 --> 00:57:09.731
That's light that's reflected from the moon.

785
00:57:10.052 --> 00:57:12.394
There's light that's reflected through the atmosphere.

786
00:57:12.454 --> 00:57:13.895
The atmosphere can absorb light.

787
00:57:14.215 --> 00:57:16.417
It can scatter light on its way out.

788
00:57:17.298 --> 00:57:21.082
And what the instrument sees depends on the angle relative to the ground as well.

789
00:57:21.797 --> 00:57:22.037
Right.

790
00:57:22.137 --> 00:57:22.398
Right.

791
00:57:22.438 --> 00:57:36.109
Like if it's looking straight down or if it's looking like sort of towards the curved side of the earth and the measured radiance equals the artificial light.

792
00:57:36.878 --> 00:57:42.700
That's getting attenuated by the atmosphere, plus the reflected moonlight, plus all of these other contributions.

793
00:57:43.200 --> 00:57:53.723
There's a variety of, again, as someone who's not done something in this area, in my head, I just think, oh, you just point it and everything is perfect.

794
00:57:53.743 --> 00:57:53.863
Yeah.

795
00:57:53.883 --> 00:57:54.063
Yeah.

796
00:57:54.623 --> 00:57:59.046
And I can just have very clear resolution on what I'm looking at.

797
00:58:00.347 --> 00:58:11.355
But I think as you've eloquently described, there are a number of variables between the light, between where the light comes from and it then ultimately being received on the sensor on the satellite.

798
00:58:11.455 --> 00:58:11.935
Exactly.

799
00:58:12.015 --> 00:58:18.000
And so that is where the meat of why these guys got the golden goose is coming from.

800
00:58:18.480 --> 00:58:18.960
It is this.

801
00:58:19.581 --> 00:58:21.322
It is the black marble algorithm.

802
00:58:21.702 --> 00:58:22.923
The job of the algorithm is,

803
00:58:23.243 --> 00:58:32.597
is to estimate and screen those other contributions using the satellite's geometry, like where it is in space and where it's pointing, and all of these additional measurements.

804
00:58:32.737 --> 00:58:34.740
Like on a cloudy night, the ground can be hidden,

805
00:58:35.336 --> 00:58:38.737
That observation should be flagged rather than treated as a real power outage.

806
00:58:39.537 --> 00:58:44.258
On a moonlit night, a field may appear brighter than before.

807
00:58:44.438 --> 00:58:46.359
Snow can amplify reflected light.

808
00:58:46.399 --> 00:58:46.979
Of course.

809
00:58:47.039 --> 00:58:48.900
Trees can hide light seasonally.

810
00:58:49.280 --> 00:58:50.780
That depends on the seasons, right?

811
00:58:50.800 --> 00:58:54.001
Because trees in the winter and in the deciduous forests are...

812
00:58:54.965 --> 00:58:56.725
are not going to hide light as much as...

813
00:58:56.826 --> 00:58:58.486
So that's not...

814
00:58:58.586 --> 00:59:02.247
If you were to just not take care of that, you'd be like, oh, winter is brighter than summer.

815
00:59:02.827 --> 00:59:04.928
No, that's just because the trees don't have any leaves.

816
00:59:05.028 --> 00:59:05.488
Yeah.

817
00:59:05.808 --> 00:59:06.008
Right?

818
00:59:06.048 --> 00:59:10.229
So there's all of these other factors that come into it.

819
00:59:10.289 --> 00:59:21.872
And the 2018 paper describes these corrections from the atmosphere, from the terrain, from the vegetation, from the snow, from the stray light, and all of the strategies to fill the gaps in that record.

820
00:59:21.992 --> 00:59:24.373
This is a combinatorial problem.

821
00:59:25.180 --> 00:59:25.760
Exactly.

822
00:59:26.880 --> 00:59:29.001
And that's the first major conceptual step.

823
00:59:29.961 --> 00:59:33.942
A nighttime photograph, for example, on the left-hand side, that's the raw data.

824
00:59:34.802 --> 00:59:37.663
And after it runs through the algorithm, we have the stuff on the left.

825
00:59:37.703 --> 00:59:38.903
This is in France.

826
00:59:38.963 --> 00:59:40.104
Stuff on the right.

827
00:59:40.184 --> 00:59:40.424
Yeah.

828
00:59:40.684 --> 00:59:40.884
Yeah.

829
00:59:41.004 --> 00:59:41.264
Sorry.

830
00:59:41.284 --> 00:59:43.544
The stuff on the right is after the algorithm.

831
00:59:43.584 --> 00:59:47.105
The stuff on the left is the raw data that we're getting from the satellite.

832
00:59:47.365 --> 00:59:49.206
And you can already see there's so many clouds.

833
00:59:49.626 --> 00:59:49.906
Yeah.

834
00:59:49.946 --> 00:59:52.426
The Pyrenees have this massive band.

835
00:59:52.526 --> 00:59:54.387
That's because of the snow there.

836
00:59:54.607 --> 00:59:54.807
Yeah.

837
00:59:55.307 --> 00:59:59.894
When you take all of that out, then you get the actual cities of France.

838
01:00:01.396 --> 01:00:02.558
I just can't even imagine.

839
01:00:02.598 --> 01:00:04.781
I mean, even just in the variables you talked through.

840
01:00:07.075 --> 01:00:11.958
Because it's not only a function of the different things happening, but it's also the different things happening over time.

841
01:00:12.538 --> 01:00:13.098
Yeah, yeah.

842
01:00:13.178 --> 01:00:16.740
It's like a, yeah, it's a time varying phenomenon that you have to take out.

843
01:00:16.900 --> 01:00:19.101
The noise is extremely dynamic.

844
01:00:19.281 --> 01:00:19.481
Right.

845
01:00:20.402 --> 01:00:27.385
The next one, we've got a photo of the north of the Scandinavian provinces.

846
01:00:28.005 --> 01:00:29.206
There's an aurora happening.

847
01:00:29.226 --> 01:00:29.506
Yeah.

848
01:00:30.427 --> 01:00:31.168
in the raw data.

849
01:00:31.708 --> 01:00:32.409
That's not real.

850
01:00:32.529 --> 01:00:38.796
You got to take that out, but you got to be careful because there's lights lurking underneath that you'd want to account for.

851
01:00:39.437 --> 01:00:41.039
And so taking out all of that,

852
01:00:41.668 --> 01:00:43.829
is the job of the black marble algorithm.

853
01:00:43.949 --> 01:00:44.709
Unbelievable.

854
01:00:45.129 --> 01:00:46.790
And the last one, this is for you, Lester.

855
01:00:47.170 --> 01:00:47.830
Oh, okay.

856
01:00:47.850 --> 01:00:50.531
This is Raleigh, North Carolina.

857
01:00:50.791 --> 01:00:51.312
Raleigh.

858
01:00:51.452 --> 01:00:52.592
Raleigh, sorry, yeah.

859
01:00:52.892 --> 01:00:54.793
I always mess that up.

860
01:00:54.833 --> 01:00:56.334
But this is the U.S. East Coast.

861
01:00:56.714 --> 01:01:04.517
The moon has a fraction of 83%, so that means like 83% of the moon is illuminated, and that's why the clouds are illuminated in that way.

862
01:01:04.537 --> 01:01:05.477
You've got to take that out.

863
01:01:06.424 --> 01:01:11.267
And now you can see Raleigh, Richmond, Virginia, Washington, D.C., Baltimore is up there.

864
01:01:11.367 --> 01:01:13.508
Northern Virginia is also.

865
01:01:13.908 --> 01:01:14.829
This is incredible.

866
01:01:15.249 --> 01:01:15.429
Right?

867
01:01:15.909 --> 01:01:20.872
But taking all of that data out, that is the job of the black marble algorithm.

868
01:01:21.332 --> 01:01:26.675
And that's what the 2018 paper was all about, was laying out the foundations of that black marble algorithm.

869
01:01:27.375 --> 01:01:43.640
We're taking a raw data set that's not clean for this specific use case of identifying artificial light driven by human activity and removing everything that is not that.

870
01:01:43.880 --> 01:01:44.320
Exactly.

871
01:01:44.640 --> 01:01:44.860
Yeah.

872
01:01:45.280 --> 01:01:50.362
And so in the next paper published in 2019, we use the Black Marble algorithm.

873
01:01:51.062 --> 01:01:53.804
to talk about what happened after Hurricane Maria.

874
01:01:54.564 --> 01:01:55.725
This is Roman and Stokes.

875
01:01:56.105 --> 01:01:57.385
They shared first author credit.

876
01:01:57.906 --> 01:02:02.148
All five of the people that are being awarded the Golden Goose are also on the authors list here.

877
01:02:02.508 --> 01:02:09.832
The paper asks how nighttime lights can be used to estimate the geography and duration of electricity loss after Hurricane Maria.

878
01:02:10.452 --> 01:02:13.894
And the first step is you just want to establish a baseline, right?

879
01:02:13.934 --> 01:02:18.417
You want to say this is what Puerto Rico was like before the hurricane came about.

880
01:02:19.818 --> 01:02:24.324
And then you can ask, okay, what is it looking like afterwards?

881
01:02:24.905 --> 01:02:30.792
Now, there is a subtle question here that has to do with resolution, which I thought was very cool.

882
01:02:30.853 --> 01:02:33.456
So look at the top left.

883
01:02:33.496 --> 01:02:36.360
The top left is actually the resolution of the satellite.

884
01:02:37.620 --> 01:02:37.940
Okay.

885
01:02:38.180 --> 01:02:43.142
And so for audio listeners, this looks like an 8-bit pixel video game.

886
01:02:43.462 --> 01:02:44.362
Of Puerto Rico.

887
01:02:44.382 --> 01:02:47.063
Or whatever the correct number is.

888
01:02:47.443 --> 01:02:49.264
But highly pixelated.

889
01:02:49.324 --> 01:02:50.284
Highly pixelated.

890
01:02:50.384 --> 01:02:51.644
And there's a reason for that, right?

891
01:02:51.664 --> 01:02:54.365
There's not a lot of light coming at night.

892
01:02:54.505 --> 01:02:54.845
Yeah.

893
01:02:55.106 --> 01:02:55.386
Okay?

894
01:02:56.166 --> 01:02:59.107
Which is a challenge in and of itself to capture it in the first place.

895
01:02:59.127 --> 01:02:59.647
Yeah, exactly.

896
01:02:59.687 --> 01:03:05.549
Like in the daytime, there's so much light coming in there that you can really resolve things

897
01:03:06.523 --> 01:03:07.383
very, very finely.

898
01:03:07.803 --> 01:03:10.264
At nighttime, there's not a lot of light coming in.

899
01:03:10.604 --> 01:03:14.445
And so naturally, your resolution is going to be not that great.

900
01:03:14.925 --> 01:03:15.945
How do you get over that?

901
01:03:16.225 --> 01:03:27.987
Well, you can combine it with Landsat and Sentinel-2 images, which are separate satellites that tell you about the distribution of houses and during the daytime.

902
01:03:28.348 --> 01:03:30.668
During the daytime, they have all of this resolution.

903
01:03:30.708 --> 01:03:35.029
If you combine these two, then you can ask things like, okay, you know,

904
01:03:36.403 --> 01:03:37.224
There's a pixel.

905
01:03:37.264 --> 01:03:43.934
There's a giant pixel, let's say from the nighttime, a 500 meter by 500 meter pixel that's covering a town.

906
01:03:43.954 --> 01:03:45.616
So this is half a kilometer wide.

907
01:03:46.538 --> 01:03:46.698
Um,

908
01:03:47.040 --> 01:03:49.763
But that town, that same pixel has a town.

909
01:03:50.143 --> 01:03:52.866
It's got a wooded hillside, and it's got part of a reservoir.

910
01:03:53.126 --> 01:04:01.294
The satellite might report the pixel that's lost most of its light, but that measurement alone is not going to pinpoint where exactly the light came from.

911
01:04:01.774 --> 01:04:11.084
If you combine it with the Landsat images and the Sentinel-2 images, then the finer daytime resolution is going to give these researchers a guide

912
01:04:11.840 --> 01:04:13.962
And you build that in into your algorithm.

913
01:04:13.982 --> 01:04:14.483
That's so good.

914
01:04:14.763 --> 01:04:14.963
Right?

915
01:04:15.323 --> 01:04:28.737
So in practice, what they're doing is interpolating, in some sense, the nighttime data onto a 30-meter grid, from a 500-meter grid to now 30 meters, which is, you know, the size of communities, big houses kind of thing.

916
01:04:29.237 --> 01:04:29.698
Mm-hmm.

917
01:04:31.021 --> 01:04:32.422
It's an incredible thing that they've done.

918
01:04:33.222 --> 01:04:35.582
So let's see how it worked, right?

919
01:04:36.563 --> 01:04:44.065
The paper actually reports an approximately 80% reduction in the island's total observed lights immediately after Maria.

920
01:04:44.085 --> 01:04:45.165
That's so crazy.

921
01:04:45.185 --> 01:04:45.365
Right?

922
01:04:45.425 --> 01:04:49.567
Because there was just so much devastation caused by this hurricane.

923
01:04:49.667 --> 01:04:54.228
And it was on the news that the recovery from that hurricane happened.

924
01:04:55.415 --> 01:05:00.621
especially in terms of the power needs of the people of Puerto Rico, it was not that great, right?

925
01:05:01.683 --> 01:05:07.269
What this actually shows is a resolution depth of which communities fared better than others.

926
01:05:07.730 --> 01:05:09.251
Oh, that's quite nice.

927
01:05:09.592 --> 01:05:12.135
So here we're looking at a color-coded graph

928
01:05:13.403 --> 01:05:20.170
Where the color represents the number of days without electricity post-Hurricane Maria.

929
01:05:20.450 --> 01:05:20.750
Okay.

930
01:05:20.950 --> 01:05:21.211
Okay?

931
01:05:21.511 --> 01:05:28.198
The brighter, yellower spots means that almost immediately, within 30 days, within a month, the power was back.

932
01:05:29.165 --> 01:05:32.487
The really dark spots, that's 150 days.

933
01:05:32.507 --> 01:05:34.748
That's like six months plus.

934
01:05:34.888 --> 01:05:35.088
Right?

935
01:05:35.548 --> 01:05:40.711
And you can see different parts of Puerto Rico have different amounts of power loss.

936
01:05:41.231 --> 01:05:43.272
On the order of months, we're talking.

937
01:05:43.792 --> 01:05:53.417
And even if you look at the Arecibo, which is the first row, third column, you'll find parts that the power returned immediately and then other parts where it didn't.

938
01:05:54.376 --> 01:06:01.281
You can then go into the data and say, what is the difference between the parts that returned early and the parts that returned late?

939
01:06:01.701 --> 01:06:06.464
And almost always it has to do with poorer rural communities.

940
01:06:06.684 --> 01:06:08.365
The urban communities got it first.

941
01:06:08.826 --> 01:06:11.207
The poor rural communities got it second.

942
01:06:11.527 --> 01:06:15.210
And so the poorer residents often bore long interruptions.

943
01:06:18.119 --> 01:06:19.681
I mean, it's a story as old as time.

944
01:06:20.081 --> 01:06:31.753
And what's interesting is you're getting this empirical data related to disaster recovery that makes it very black and white.

945
01:06:32.474 --> 01:06:34.916
Because it's also happening over time.

946
01:06:34.977 --> 01:06:35.077
Yeah.

947
01:06:35.097 --> 01:06:35.197
Yeah.

948
01:06:35.577 --> 01:06:41.980
And so you get this level of granularity where, like you talked about earlier, where it's like 150 days versus almost immediately.

949
01:06:42.580 --> 01:07:02.650
And I can just imagine in my mind, you know, if I'm thinking of, you know, something like FEMA in the U.S. in terms of responses, like that data, if I was someone who was trying to plan to make disaster recovery more improved in the future, just something as simple as that context.

950
01:07:02.850 --> 01:07:03.210
Mm-hmm.

951
01:07:03.790 --> 01:07:12.460
Um, you know, cause then I could then go as someone who's in the disaster recovery space and say, okay, well, in these areas, are our processes different?

952
01:07:12.801 --> 01:07:13.962
Did we access them later?

953
01:07:14.002 --> 01:07:14.703
Do we have enough?

954
01:07:16.557 --> 01:07:35.225
coverage from like a resources and human capital perspective um and in theory what you would want to be able to do i'm just extrapolating over time is look at because obviously these weather events continue to happen yeah um they might be happening more frequently

955
01:07:35.665 --> 01:07:36.666
Because of climate change?

956
01:07:36.806 --> 01:07:37.226
Correct.

957
01:07:37.787 --> 01:07:45.672
You could start to at least look at this as a way to make progressive progress because you can see how it happened in X year in the past.

958
01:07:46.012 --> 01:07:51.976
Look at Y year and then say, okay, have we increased – have some of these communities decreased in response time?

959
01:07:52.416 --> 01:07:55.218
And then you can then have real data to say we've done better.

960
01:07:55.518 --> 01:07:56.099
Yeah.

961
01:07:56.319 --> 01:07:58.100
Just as an example, that's where my mind goes.

962
01:07:58.120 --> 01:07:58.280
Yeah.

963
01:07:59.200 --> 01:07:59.540
Totally.

964
01:08:00.061 --> 01:08:08.307
I mean, it's, it's, I think it's a really cool way to look at empirical data from satellites and then like answer these societal questions.

965
01:08:08.647 --> 01:08:10.669
Like, do we, do we, is this okay?

966
01:08:11.129 --> 01:08:11.329
Yeah.

967
01:08:11.530 --> 01:08:12.210
Yeah.

968
01:08:12.530 --> 01:08:15.993
So we move on to our paper number three, which is 2022.

969
01:08:16.633 --> 01:08:19.796
This is Eleanor Stokes and Miguel Roman.

970
01:08:20.937 --> 01:08:22.278
This is in scientific reports.

971
01:08:24.020 --> 01:08:32.666
Here, rather than a hurricane cutting power, we've got governments that impose restrictions and people are altered because of the first months of COVID-19, the pandemic.

972
01:08:32.966 --> 01:08:40.712
And the paper examines hundreds of urban areas across the Middle East using this black marble observations over several years.

973
01:08:42.408 --> 01:08:46.092
And it's, it's, I think an incredible paper.

974
01:08:46.152 --> 01:08:50.716
The study actually found that nightlights often captured the onset of curfews and lockdowns.

975
01:08:50.756 --> 01:08:54.459
I was literally just thinking like, oh, when the curfew started, it probably lit up.

976
01:08:54.619 --> 01:08:54.860
Yeah.

977
01:08:55.580 --> 01:08:56.521
Or no.

978
01:08:56.921 --> 01:08:57.102
Yeah.

979
01:08:57.142 --> 01:08:57.422
Lit up.

980
01:08:57.582 --> 01:08:57.782
Yeah.

981
01:08:57.942 --> 01:08:58.122
Yeah.

982
01:08:58.203 --> 01:08:58.843
Because people were.

983
01:08:58.863 --> 01:08:59.263
Yeah.

984
01:08:59.304 --> 01:09:00.024
They were at home.

985
01:09:00.244 --> 01:09:01.105
Yeah.

986
01:09:01.125 --> 01:09:08.372
And the response varied within and between countries, but it was a little more nuanced because even when the curfew started, that means there's less traffic lights.

987
01:09:09.640 --> 01:09:14.142
So, you know, it's a bit more complicated than that.

988
01:09:14.162 --> 01:09:19.904
So in places that were affected by conflict or weak governance, the patterns could be very, very different.

989
01:09:20.464 --> 01:09:25.726
And the authors also considered Ramadan, which happened around the same time, which provided a stringent list.

990
01:09:25.746 --> 01:09:30.888
We need like a familiar annual change of social activity, which we've observed in other places.

991
01:09:31.594 --> 01:09:32.314
years, right?

992
01:09:32.334 --> 01:09:34.295
Because you've got this backlog of activity.

993
01:09:34.555 --> 01:09:38.117
So you know what a normal Ramadan looks like in these Muslim countries.

994
01:09:38.297 --> 01:09:43.339
And then you can say, okay, what is Ramadan plus COVID look like in these Muslim countries?

995
01:09:43.579 --> 01:09:53.563
I imagine that's going to be a valuable also the COVID data set will have similar baseline value for other questions later.

996
01:09:53.743 --> 01:09:54.643
Yeah.

997
01:09:56.044 --> 01:09:58.745
Oh, that's... Yeah, I think it's a really cool way to do like

998
01:09:59.904 --> 01:10:01.425
Data science in general, right?

999
01:10:01.685 --> 01:10:06.948
Because there's a natural oscillation and a natural rhythm to human life in these areas.

1000
01:10:07.308 --> 01:10:09.730
And you have a record of that from the previous data.

1001
01:10:10.030 --> 01:10:13.092
And now all of a sudden you've introduced the COVID-19 pandemic.

1002
01:10:13.452 --> 01:10:14.773
How does that affect, right?

1003
01:10:17.114 --> 01:10:20.637
It's a pretty incredible paper that came out.

1004
01:10:20.917 --> 01:10:22.478
So now let's go to paper number four.

1005
01:10:23.378 --> 01:10:25.600
This is the fourth paper published in 2025. Who has electricity?

1006
01:10:30.493 --> 01:10:30.653
Right.

1007
01:10:30.753 --> 01:10:31.273
Is the idea.

1008
01:10:31.353 --> 01:10:34.835
It asks the question that sounds pretty straightforward until you look at it closely.

1009
01:10:34.875 --> 01:10:41.177
Can a decade of nighttime imagery tell us where people have access to electricity?

1010
01:10:41.997 --> 01:10:43.878
I mean, consider like two dark locations.

1011
01:10:43.898 --> 01:10:49.820
There's one that's an empty tract of land and another one with a settlement of thousands of people and very few outdoor lights.

1012
01:10:50.480 --> 01:10:53.882
A darkness map cannot alone tell you which is which.

1013
01:10:54.262 --> 01:10:55.802
So these researchers...

1014
01:10:57.583 --> 01:11:05.306
combined the Black Marble time series with the settlement mask and the population estimates that they have from other data sources.

1015
01:11:05.846 --> 01:11:13.690
And they can classify lighting evidence over time with a scale of one kilometer, which is the stuff that you get from the Black Marble.

1016
01:11:13.750 --> 01:11:20.532
And then they use the population layer to estimate how many people live in those areas that are assigned to all these different access categories.

1017
01:11:20.592 --> 01:11:26.655
And the logic is that you first determine whether people plausibly live there using population data,

1018
01:11:26.915 --> 01:11:28.316
that you have from all these other sources.

1019
01:11:28.977 --> 01:11:34.161
And then you examine whether the night signal is consistent with electric lighting for that number of people.

1020
01:11:35.022 --> 01:11:40.466
And then you estimate the population affected and compare broad totals against these surveys.

1021
01:11:40.566 --> 01:11:45.030
And that's kind of the theme across the Black Marble papers.

1022
01:11:45.530 --> 01:11:51.755
It's like, how do I combine disparate data sources about humans on this planet, right?

1023
01:11:51.795 --> 01:11:54.597
You have the data source from this satellite

1024
01:11:55.398 --> 01:11:58.121
about nighttime lights, but that can't be enough.

1025
01:11:58.681 --> 01:11:58.882
Right?

1026
01:11:59.302 --> 01:12:00.383
And that's what the story is.

1027
01:12:00.443 --> 01:12:00.824
Yeah.

1028
01:12:00.844 --> 01:12:05.629
The story is one of data aggregation and data analysis.

1029
01:12:06.635 --> 01:12:07.216
It's pretty cool.

1030
01:12:07.997 --> 01:12:12.142
It's also such an interesting set of questions that were not necessarily intuitive to me.

1031
01:12:12.903 --> 01:12:16.987
Like if you were to give me a million years to think about ideas to think about.

1032
01:12:18.229 --> 01:12:24.416
But it makes – it's clever and makes obvious sense in retrospect now.

1033
01:12:24.476 --> 01:12:24.696
Yeah.

1034
01:12:24.836 --> 01:12:25.077
Yeah.

1035
01:12:25.497 --> 01:12:38.263
that, you know, oh, this satellite that we're using to do nighttime surveys, something arose out of that, which is like, oh, actually, we could look at human activity because of light.

1036
01:12:38.503 --> 01:12:39.503
Yeah.

1037
01:12:39.803 --> 01:12:49.848
And one of the cool things they did in the next figure is they looked at what the countries reported their access was,

1038
01:12:50.443 --> 01:12:51.425
For their population.

1039
01:12:52.086 --> 01:13:00.158
And what they estimate was the level of access.

1040
01:13:00.538 --> 01:13:04.985
Now, one of the interesting things is Africa seems to be underestimating.

1041
01:13:06.592 --> 01:13:09.215
how much access their population has to electricity.

1042
01:13:09.855 --> 01:13:13.579
But at the same time, Africa has a massive variance.

1043
01:13:13.979 --> 01:13:17.022
This is delineated by continent.

1044
01:13:17.583 --> 01:13:24.770
And you can see that there are some countries where the actual access, according to Black Marble, is a lot higher than what they report.

1045
01:13:25.130 --> 01:13:27.993
But then at the same time, Africa has some that are like negative 10%.

1046
01:13:29.474 --> 01:13:33.778
What they're estimating the population has access to.

1047
01:13:34.098 --> 01:13:37.721
All of the other continents are a little bit sort of constrained.

1048
01:13:39.082 --> 01:13:42.344
Europe, very interestingly, is overestimating.

1049
01:13:42.364 --> 01:13:47.328
So do with that what you will.

1050
01:13:47.388 --> 01:13:50.530
North America is very fairly on the zero.

1051
01:13:50.570 --> 01:13:55.694
We estimated exactly how much... American precision.

1052
01:13:55.714 --> 01:13:56.355
Yeah, yeah.

1053
01:13:57.295 --> 01:14:06.844
It could just be that, like, I don't know this, but, like, what if the estimate is also related to the satellite data?

1054
01:14:07.584 --> 01:14:10.507
That's something that I'm going to ask Miguel, actually.

1055
01:14:10.607 --> 01:14:12.048
That's one of the questions I have.

1056
01:14:12.088 --> 01:14:12.969
That's a part two question.

1057
01:14:12.989 --> 01:14:13.509
Like, yeah.

1058
01:14:13.529 --> 01:14:13.770
Yeah.

1059
01:14:14.270 --> 01:14:17.712
Look, there's a little bit of a, we got to ask.

1060
01:14:17.752 --> 01:14:19.052
Yeah, I got to ask, right?

1061
01:14:19.072 --> 01:14:21.173
Like, it seems like a fair question.

1062
01:14:21.493 --> 01:14:25.755
Anyway, so another map that's useful is the underserved places.

1063
01:14:25.775 --> 01:14:27.076
Like, where are the underserved places?

1064
01:14:27.436 --> 01:14:31.218
This shows a percent of population with access to electricity for each country.

1065
01:14:32.360 --> 01:14:37.305
And you can see that the highest gaps are observed in Africa and parts of Asia and Oceania.

1066
01:14:37.885 --> 01:14:38.206
Yeah.

1067
01:14:38.626 --> 01:14:41.128
So the application is pretty concrete.

1068
01:14:41.148 --> 01:14:47.194
You can have grid planners and public agencies then know where the demand is unmet, for example.

1069
01:14:47.774 --> 01:14:48.495
It's huge, right?

1070
01:14:48.695 --> 01:14:54.220
I mean, my brain is just going into a million different places of applications for this context.

1071
01:14:54.501 --> 01:14:56.422
I mean, you could have a field day.

1072
01:14:56.743 --> 01:14:56.863
Yeah.

1073
01:14:56.883 --> 01:14:56.983
Yeah.

1074
01:14:58.504 --> 01:15:13.255
in terms of there's no shortage of agencies, organizations, at local municipalities, state, even national levels, that can use this in ways that are extremely relevant to a number of their mandates.

1075
01:15:13.755 --> 01:15:15.216
Yeah, yeah, exactly.

1076
01:15:15.236 --> 01:15:15.977
That's quite good.

1077
01:15:16.217 --> 01:15:23.442
Yeah, so our final paper, Paper 5, it's a paper that's published in Nature Proper of this year.

1078
01:15:26.395 --> 01:15:27.016
Here's what they did.

1079
01:15:27.516 --> 01:15:39.669
So using daily observations from 2014 through 2022, nearly a decade, they asked whether the familiar story of an ever brightening earth survives the actual data

1080
01:15:41.722 --> 01:15:43.084
that we've taken, right?

1081
01:15:43.524 --> 01:15:44.826
Light pollution is getting bad.

1082
01:15:45.987 --> 01:15:46.929
That's what everyone says.

1083
01:15:47.289 --> 01:15:48.450
Who says it the most?

1084
01:15:48.631 --> 01:15:49.732
The astronomers?

1085
01:15:49.792 --> 01:15:51.154
Astronomers are trying to say that.

1086
01:15:51.454 --> 01:15:54.357
I mean, they're complaining about the Starlink that's up there.

1087
01:15:55.118 --> 01:15:58.042
But also, they're also saying that light pollution has gotten bad.

1088
01:15:58.102 --> 01:16:00.365
And like all the, you know...

1089
01:16:00.845 --> 01:16:09.271
People who want to live by the stars are saying that light pollution has gotten very bad and I can't see Cassiopeia anymore, right, from New York City.

1090
01:16:09.291 --> 01:16:10.772
It's like maybe don't live in New York City.

1091
01:16:10.792 --> 01:16:11.052
I don't know.

1092
01:16:11.072 --> 01:16:12.533
What does the data say?

1093
01:16:12.553 --> 01:16:12.733
Yeah.

1094
01:16:12.893 --> 01:16:14.054
So let's see what the data says.

1095
01:16:15.235 --> 01:16:15.915
Here's what it says.

1096
01:16:16.636 --> 01:16:17.997
It's not a simple story.

1097
01:16:19.078 --> 01:16:21.880
as nothing really is when it comes to human civilization, right?

1098
01:16:22.260 --> 01:16:31.466
The paper applies continuous change detection to every single pixel that the satellite looks at for the past 10 years, right?

1099
01:16:32.567 --> 01:16:36.489
And it looks at a sequence, not just pairs of annual screenshots.

1100
01:16:37.210 --> 01:16:45.275
And the technical idea is to fit an expected pattern that includes seasonal cycles, so you've got to take care of the seasonal cycles, and...

1101
01:16:45.515 --> 01:16:46.896
You've got to take care of trends.

1102
01:16:46.936 --> 01:16:49.018
You've got to account for the viewing angle, as I was saying.

1103
01:16:49.379 --> 01:16:59.128
You've got to examine observations persistently when they depart from some model that you've created that's based on the moon, the seasonal cycles, and so on and so forth.

1104
01:16:59.568 --> 01:17:03.192
And the headline result is more interesting than a single net percentage.

1105
01:17:03.472 --> 01:17:06.975
So across the study, they show that a brightening contributed

1106
01:17:07.836 --> 01:17:15.483
a radiance gain of approximately 34% of baseline in 2014, but there was a dimming offset of 18%.

1107
01:17:16.284 --> 01:17:19.246
Some areas got brighter, other areas got dimmer.

1108
01:17:19.567 --> 01:17:21.669
It's not that simple, is the idea.

1109
01:17:21.989 --> 01:17:23.470
Everyone's moving away from the coast.

1110
01:17:23.610 --> 01:17:32.779
Yeah, so here, over here, what we're looking at is, the color actually codes the year of the most recently detected abrupt change.

1111
01:17:33.199 --> 01:17:33.499
Okay.

1112
01:17:33.539 --> 01:17:34.240
In luminescence.

1113
01:17:34.300 --> 01:17:35.040
Okay.

1114
01:17:35.180 --> 01:17:39.763
The bluer year means it's closer to 2014 and the redder year means it's closer to 2022.

1115
01:17:39.943 --> 01:17:47.308
And what we're looking at is within a year, was there a significant change in how much activity was happening?

1116
01:17:47.588 --> 01:17:54.893
So from here, we can actually see specific spots on Earth that will show kind of interesting stories.

1117
01:17:55.253 --> 01:18:00.757
So in the next photo, we'll see these are panels that have very specific...

1118
01:18:02.828 --> 01:18:04.789
Spots on the earth highlighted.

1119
01:18:05.189 --> 01:18:09.090
I'm going to go from left to right so you can kind of understand what we're looking at.

1120
01:18:09.531 --> 01:18:11.831
On the left is Guangzhou, China.

1121
01:18:11.931 --> 01:18:15.893
This is the urban area around Hong Kong that's right outside of Hong Kong.

1122
01:18:16.741 --> 01:18:20.403
You can see urbanization happening very, very tightly.

1123
01:18:20.923 --> 01:18:22.604
Let me just go through the colors real quick.

1124
01:18:23.104 --> 01:18:31.509
Blue means that there was a net reduction, and red means there was a net gain in the amount of light that was coming through.

1125
01:18:31.549 --> 01:18:34.591
You can see the Guangzhou, China is insanely... Yep.

1126
01:18:34.631 --> 01:18:34.691
...

1127
01:18:35.511 --> 01:18:35.972
lighting up.

1128
01:18:36.492 --> 01:18:36.752
Okay.

1129
01:18:37.293 --> 01:18:40.315
The next one, armed conflicts in Kiev, Ukraine.

1130
01:18:40.776 --> 01:18:41.636
That's the second row.

1131
01:18:41.917 --> 01:18:43.178
You can see it's mostly blue.

1132
01:18:43.318 --> 01:18:43.618
Yeah.

1133
01:18:44.018 --> 01:18:44.239
Right.

1134
01:18:44.759 --> 01:18:44.919
Yeah.

1135
01:18:44.939 --> 01:18:47.081
Because there's a lot of dimming that's happening.

1136
01:18:47.722 --> 01:18:53.647
The third row shows development of dragon fruit architecture or agriculture in Vietnam.

1137
01:18:53.667 --> 01:18:54.087
Vietnam.

1138
01:18:54.667 --> 01:18:55.227
Interesting.

1139
01:18:55.527 --> 01:18:59.950
So you're seeing a little bit of lighting up and a little bit of dimming based on like where that's happening.

1140
01:19:00.310 --> 01:19:08.374
Environmental policies in Paris, France show a lot of dimming because they've changed the lights that are in the city.

1141
01:19:08.854 --> 01:19:09.114
Okay.

1142
01:19:09.414 --> 01:19:13.276
The next one is gas flare changes in the Middle East region.

1143
01:19:15.458 --> 01:19:20.904
Massive, massive flare-ups and also massive, massive shutdowns of where the gas is happening.

1144
01:19:21.465 --> 01:19:24.408
Next one, you've got power outages in Puerto Rico, as I was showing.

1145
01:19:24.809 --> 01:19:25.770
A lot of that is blue.

1146
01:19:26.130 --> 01:19:29.614
And then some of it is red because it came back up.

1147
01:19:29.895 --> 01:19:35.681
Urban expansion and decentralization in Charlotte, North Carolina, is the one right after.

1148
01:19:36.522 --> 01:19:38.203
Charlotte's changing a lot, apparently.

1149
01:19:38.243 --> 01:19:41.263
Shout out Tabo, who's in my brother, who's in Charlotte right now.

1150
01:19:41.303 --> 01:19:44.584
Yeah, so that's the second to last column.

1151
01:19:44.644 --> 01:19:49.946
And then finally, the economic collapse of Valencia and Caracas in Venezuela.

1152
01:19:49.966 --> 01:19:50.846
Yeah, wow.

1153
01:19:50.866 --> 01:19:54.087
A lot of dimming happening in Venezuela, as you can see.

1154
01:19:54.507 --> 01:19:58.010
A lot of blue, which means there's a lot of dimming that's happening in Venezuela.

1155
01:19:58.250 --> 01:20:00.012
And this is before 2022.

1156
01:20:00.352 --> 01:20:04.235
So this is before any of the recent shenanigans that have been happening.

1157
01:20:05.496 --> 01:20:07.998
But, you know, it tells a compelling story.

1158
01:20:08.379 --> 01:20:09.500
Each of these locations...

1159
01:20:10.320 --> 01:20:16.023
It's also important to also recognize that, and it kind of reminds me a little bit of our Navier Stokes episode.

1160
01:20:16.043 --> 01:20:20.605
You have the larger system, and then you have the systems, the smaller systems within that larger system.

1161
01:20:20.845 --> 01:20:25.688
And depending on what level of Zoom you look at,

1162
01:20:26.448 --> 01:20:53.584
different stories can be told because when you look at the whole thing it's one story but when you kind of you can tie these real world social political economic cultural events or happenings to again this empirical days this is so good it's so cool this is the the most recent paper that they have so i just want to go on a little bit of a deeper dive here so this is figure two how often does a place change and why

1163
01:20:54.024 --> 01:21:03.089
So first point, if you look at the small pie chart on the lower left, that shows that roughly 80% of places with an abrupt change have it in a single year.

1164
01:21:03.829 --> 01:21:05.550
Things happen fast.

1165
01:21:06.150 --> 01:21:09.132
Only 20% have it over multiple years.

1166
01:21:10.044 --> 01:21:15.088
So things happen fast when it comes to lighting and when it comes to human civilization, which I think is interesting.

1167
01:21:15.168 --> 01:21:18.871
And then they label these like sort of labeled boxes in the world map.

1168
01:21:18.891 --> 01:21:21.273
You've got the gas flares in central United States.

1169
01:21:21.754 --> 01:21:23.915
That's the box that you see in the central United States.

1170
01:21:23.975 --> 01:21:25.357
That's centered around Texas.

1171
01:21:26.597 --> 01:21:30.960
those are the Permian Basin in Texas that has all of the oil.

1172
01:21:31.380 --> 01:21:33.301
You've got the grid failure in Venezuela.

1173
01:21:33.541 --> 01:21:35.402
That's the next one.

1174
01:21:35.582 --> 01:21:39.824
Then you've got conflicts around the Middle East, so a lot of changes happening in the Middle East.

1175
01:21:40.244 --> 01:21:44.867
And then you've got construction and demolition happening in India and in China.

1176
01:21:45.267 --> 01:21:54.392
One of the interesting things that I found about India was a lot of the change was concentrated in the northern part of India, in Uttar Pradesh,

1177
01:21:55.238 --> 01:21:57.796
and Haryana and like that, the Ganges Valley.

1178
01:22:00.355 --> 01:22:05.036
Now I'm talking completely out of my, you know what?

1179
01:22:05.336 --> 01:22:07.937
The time period is 2014 to 2022.

1180
01:22:08.637 --> 01:22:14.978
And I wonder if this has something to do with Modi and the BJP coming to power, right?

1181
01:22:15.418 --> 01:22:18.839
And a lot of their strength is centered in Northern India.

1182
01:22:19.259 --> 01:22:24.780
And they want to like, they diverted a lot of resources into the construction and

1183
01:22:26.916 --> 01:22:32.040
you know, absurgence of those areas in India, because I don't see the similar things happening down South.

1184
01:22:32.060 --> 01:22:32.220
Right.

1185
01:22:32.240 --> 01:22:35.022
And down South is where Hyderabad is Bangalore.

1186
01:22:35.322 --> 01:22:36.183
These are tech cities.

1187
01:22:36.303 --> 01:22:36.623
Right.

1188
01:22:36.883 --> 01:22:37.303
What's going on?

1189
01:22:37.323 --> 01:22:39.444
But I'm not seeing, I'm just saying.

1190
01:22:39.584 --> 01:22:46.345
Now, this is, again, there's, yo, if there's like Indians in the comments who are Modi worshipers, come at me, okay?

1191
01:22:47.005 --> 01:22:48.366
I don't care, all right?

1192
01:22:48.446 --> 01:22:51.486
But I'm just saying that it seems maybe correlated.

1193
01:22:52.087 --> 01:22:55.567
And there's a worthy hypothesis that now this data set allows us to actually ask that question.

1194
01:22:55.607 --> 01:22:55.827
Yeah.

1195
01:22:55.847 --> 01:22:58.688
Because people have been saying that, independent of you bringing it up now.

1196
01:22:58.788 --> 01:23:01.469
Yeah, and if there are alternate explanations, let me know.

1197
01:23:01.549 --> 01:23:03.689
But this seems to me the simplest sort of explanation.

1198
01:23:03.949 --> 01:23:06.670
Like, why is it all centered around the Ganges Valley?

1199
01:23:08.253 --> 01:23:15.226
that's Uttar Pradesh, Bihar, Jharkhand like I know my stuff okay Haryana like these are all BJP strongholds

1200
01:23:16.384 --> 01:23:19.207
Like, why is all of the lighting up happening in BJP Strongholds?

1201
01:23:19.447 --> 01:23:19.827
Sorry.

1202
01:23:19.947 --> 01:23:21.989
I think maybe I spent a little too much time on that.

1203
01:23:22.009 --> 01:23:22.310
No, no, no.

1204
01:23:22.350 --> 01:23:22.710
It's great.

1205
01:23:23.190 --> 01:23:33.941
But I mean, I think this gets back to what I was saying earlier, which is the data gives you a reference point to ask a whole variety of questions.

1206
01:23:35.002 --> 01:23:37.944
And because we can now see this, it's happening.

1207
01:23:38.024 --> 01:23:38.125
Yeah.

1208
01:23:38.145 --> 01:23:38.265
Yeah.

1209
01:23:38.952 --> 01:23:40.413
Like this, it's happening.

1210
01:23:40.453 --> 01:23:41.454
Yeah.

1211
01:23:41.574 --> 01:23:44.416
And the time period, the 2014 to 2022, right?

1212
01:23:44.716 --> 01:23:53.363
This is right around when Modi was elected as prime minister and like his first few decades, I guess, first decade in power, right?

1213
01:23:53.423 --> 01:23:54.264
So I don't know.

1214
01:23:54.444 --> 01:23:55.805
We can start asking these questions.

1215
01:23:55.865 --> 01:23:56.926
Yeah.

1216
01:23:57.326 --> 01:24:00.629
So figure three, why the global average is deceptive.

1217
01:24:00.649 --> 01:24:02.210
You can't just take a global average, right?

1218
01:24:02.230 --> 01:24:04.232
Because the original question was, oh, like the...

1219
01:24:05.021 --> 01:24:06.281
The world is getting brighter.

1220
01:24:06.321 --> 01:24:07.042
Right, right.

1221
01:24:07.302 --> 01:24:11.723
Nah, you know, as we just saw, stuff is getting dimmer, stuff is getting brighter.

1222
01:24:12.083 --> 01:24:14.744
This figure, it's a large map.

1223
01:24:14.824 --> 01:24:18.885
The red is accumulated brightening and the blue is accumulated dimming.

1224
01:24:19.726 --> 01:24:22.687
And the pale and whitish areas are substantial amounts of both.

1225
01:24:22.727 --> 01:24:23.727
What's going on in Europe?

1226
01:24:24.552 --> 01:24:24.832
Right.

1227
01:24:25.432 --> 01:24:25.752
Sorry.

1228
01:24:25.852 --> 01:24:26.232
Dimming.

1229
01:24:26.412 --> 01:24:30.433
I didn't mean to cut you off, but the immediate thing I see, some areas you expect.

1230
01:24:30.593 --> 01:24:30.833
Yeah.

1231
01:24:30.933 --> 01:24:32.514
But the thing I'm looking at is like, what happens?

1232
01:24:32.554 --> 01:24:33.534
A lot of dimming, right?

1233
01:24:34.234 --> 01:24:38.755
That probably has to do with policy changes in the types of lights that they're allowed to make.

1234
01:24:39.035 --> 01:24:39.175
Right.

1235
01:24:39.195 --> 01:24:39.975
That's fair.

1236
01:24:40.555 --> 01:24:43.256
Maybe they're doing more efficient street lights, things like that.

1237
01:24:43.956 --> 01:24:50.617
Maybe the headlights themselves are more efficient and lower power so they don't bleed out into the night sky.

1238
01:24:52.468 --> 01:24:53.850
So I think that's what's happening.

1239
01:24:53.870 --> 01:24:57.655
But if you look, that's a very different situation than Venezuela.

1240
01:24:57.955 --> 01:25:03.422
In Venezuela, there's the same amount of dimming, but that's actually because of the economics and the economic collapse.

1241
01:25:04.123 --> 01:25:08.509
What's interesting is the economics of the region is not necessarily correlated to a brightening or dimming.

1242
01:25:08.669 --> 01:25:09.250
Exactly.

1243
01:25:09.290 --> 01:25:11.392
There can be a number of other reasons.

1244
01:25:11.412 --> 01:25:12.273
It could be policy.

1245
01:25:12.373 --> 01:25:12.554
Right.

1246
01:25:12.834 --> 01:25:12.994
Right.

1247
01:25:13.034 --> 01:25:14.856
There's so many multiple faceted things.

1248
01:25:14.896 --> 01:25:15.197
Yeah.

1249
01:25:15.297 --> 01:25:24.888
And in Europe, actually, we see a 33% reduction in France, 22% reduction in the UK, and 21% in the Netherlands over this study period.

1250
01:25:24.908 --> 01:25:25.188
That's it.

1251
01:25:25.248 --> 01:25:28.452
And that connects to this broader pattern of lighting technology and policy.

1252
01:25:28.512 --> 01:25:35.481
Venezuela, as I said, strong dimming, but that's probably because of economic and infrastructure decline.

1253
01:25:35.882 --> 01:25:42.130
The few things that I want to mention here, the United States, there's a difference between the Western brightening that's happening and...

1254
01:25:42.590 --> 01:25:44.051
And the eastern parts that are dimming.

1255
01:25:44.151 --> 01:25:47.032
In the west, you've got the Permian Basin in Texas.

1256
01:25:47.392 --> 01:25:50.194
You've got Bakken in North Dakota that has the oil fields.

1257
01:25:51.094 --> 01:25:55.076
Extraction and flaring can create these alternating increases and decreases.

1258
01:25:55.436 --> 01:25:58.778
And those are some of the stuff that you're seeing in the North Dakota and the Texas areas.

1259
01:26:00.455 --> 01:26:08.060
Again, in India and China, you zoom in onto the eastern parts of China and then India, again, big northern part.

1260
01:26:08.761 --> 01:26:16.645
I'm just saying the authors describe strong development-related brightening with different timing in northern and south-southern India.

1261
01:26:17.006 --> 01:26:25.011
So the authors of this paper are noticing what I was noticing, which is that northern and southern India have a difference that is pretty stark in the data.

1262
01:26:26.274 --> 01:26:34.845
I think that's also for folks who, you know, may be into larger geopolitical conversations, right?

1263
01:26:34.885 --> 01:26:38.030
There's this, there's been this, um,

1264
01:26:39.383 --> 01:27:04.077
conversation around, you know, re on shoring in the United States as an example, because we've offshored so much of our manufacturing capacity to a variety of areas in regions where we're seeing the brightness increasing significantly, the rise of China, a lot of these other, again, conversations and narratives, but the data is very like, it's, it's so interesting.

1265
01:27:04.597 --> 01:27:05.298
The West is,

1266
01:27:06.346 --> 01:27:09.227
you know, is in a very different place.

1267
01:27:09.427 --> 01:27:09.667
Yeah.

1268
01:27:10.788 --> 01:27:26.554
As it relates to the data set, you can make, you can extrapolate a variety of things from that versus sort of the rise of areas that people talk about all time bricks, you know, and this, and it becoming now a block that is challenging on the world stage.

1269
01:27:27.115 --> 01:27:29.476
And it is interesting for me as someone who's sort of like a,

1270
01:27:30.957 --> 01:27:39.143
a political junkie to see empirical data that kind of has interesting implications as it relates to how people think about that conversation.

1271
01:27:39.544 --> 01:27:39.784
Yeah.

1272
01:27:40.104 --> 01:27:46.209
And you can see the stark difference between like sort of the West and the BRICS nations and the, and the developing nation.

1273
01:27:46.549 --> 01:27:50.853
If you look at the aggregate top plot and the, the aggregate population,

1274
01:27:51.163 --> 01:27:52.184
plot on the right-hand side.

1275
01:27:52.224 --> 01:27:55.446
So those are showing by... Is this...

1276
01:27:56.066 --> 01:27:56.907
This is longitude?

1277
01:27:57.347 --> 01:27:58.988
Oh, we got a comment on this last time.

1278
01:27:59.028 --> 01:28:01.390
Yeah, no, this is longitude, this is latitude, right?

1279
01:28:01.690 --> 01:28:05.393
And so by longitude, you're seeing the Western Hemisphere...

1280
01:28:06.165 --> 01:28:13.713
Like America and South America don't have that much change compared to the massive changes happening because of the Middle East.

1281
01:28:13.993 --> 01:28:14.293
100%.

1282
01:28:14.433 --> 01:28:15.755
And it's happening because of China.

1283
01:28:16.035 --> 01:28:16.655
Yep.

1284
01:28:16.936 --> 01:28:17.857
On the longitude scale.

1285
01:28:18.137 --> 01:28:22.501
And on the latitude scale, well, China, India, and the Middle East are all on the same sort of longitude.

1286
01:28:22.561 --> 01:28:27.126
So you're getting massive longitude like in the tropics in the northern hemisphere and nothing elsewhere.

1287
01:28:27.426 --> 01:28:37.740
And then the global South in this context is effectively in both, both, in both views is, and again, this is the conversation where the global South is trying to become a part of the conversation.

1288
01:28:38.300 --> 01:28:42.766
Again, there is, I think the interesting thing you note here is the Middle East is,

1289
01:28:43.407 --> 01:29:04.759
on both axes is just massive is massive but that's that's a that's a factor of the fact that they have oil so like the refineries and the and the oil fields are extremely bright yeah at night you know and as well as a whole variety of other things they're not necessarily uh net positive exactly so it's it's it's complex yeah

1290
01:29:05.419 --> 01:29:07.741
But it's incredible that we have the data, right?

1291
01:29:07.801 --> 01:29:08.621
Oh, man.

1292
01:29:08.701 --> 01:29:13.385
Because we are no longer bound by what we gather from satellites.

1293
01:29:13.805 --> 01:29:21.470
We have now a very sophisticated algorithm that can take care of all of the riffraff and tell us what the human activity actually is.

1294
01:29:21.491 --> 01:29:26.434
So we can start answering these questions and know that it is not because of artifacts.

1295
01:29:26.814 --> 01:29:27.995
It's not because of the moon.

1296
01:29:28.355 --> 01:29:29.456
It's not because of clouds.

1297
01:29:29.496 --> 01:29:31.778
It's not because of snow or the aurora borealis.

1298
01:29:31.818 --> 01:29:34.240
No, this is human activity that we're seeing.

1299
01:29:35.380 --> 01:29:58.019
from from space yeah it's that's quite it's quite nice cool yeah it's quite and the last thing i want to do is just return to the first image that we have of puerto rico after maria it's visually arresting to see the darkened island it tells us immediately that something terrible has happened and it tells us the potential of this kind of technology if there are future

1300
01:30:00.192 --> 01:30:26.132
disasters like this this is something that we have an immediate data access to right 100 so um it's it's actually incredible the stuff that we can do for the betterment of humanity not just to study how humans are living on this planet yeah but you know as this algorithm gets more and more sophisticated as it starts incorporating more and more data i mean right now it's just incorporating satellite data and some population level data imagine

1301
01:30:27.631 --> 01:30:30.073
it can start incorporating localized data.

1302
01:30:30.533 --> 01:30:32.754
That was immediately what I was thinking about.

1303
01:30:33.095 --> 01:30:49.566
The more sophisticated this algorithm gets, the more granular we're going to have an understanding of human disasters, human influence on the environment, human influence on our natural environment like forests.

1304
01:30:50.504 --> 01:30:54.025
forest deforestation, like all sorts of stuff that I really care about.

1305
01:30:54.445 --> 01:30:57.787
Actually, because you talked about the trees earlier.

1306
01:30:58.047 --> 01:31:01.248
Deforestation seems like an obvious one that is going to be very easy to see.

1307
01:31:01.288 --> 01:31:01.428
Yeah.

1308
01:31:01.708 --> 01:31:15.633
And the other one that I didn't really talk about because it's not highlighted in one of the papers, but they were talking about how offshore fishing, illegal offshore fishing, is something that they've like literally caught.

1309
01:31:16.343 --> 01:31:26.346
Where there are countries that will, you know, kind of try and bend the rules and go into places where they shouldn't and try to fish and kill species that are protected.

1310
01:31:27.527 --> 01:31:31.408
We've all agreed to protect on the open ocean.

1311
01:31:32.409 --> 01:31:36.350
And this kind of data has access to...

1312
01:31:38.030 --> 01:31:42.893
you know, finding out whether, whether that's actually happening and if we can actually do something about it.

1313
01:31:42.973 --> 01:31:44.714
What's, and I'll make this the last note here.

1314
01:31:44.734 --> 01:31:49.858
What's interesting about that is, you know, folks have seen the imagery because a lot of it's a lot of times it's night fishing.

1315
01:31:50.138 --> 01:31:52.159
And so they need these, yeah, they need lighting.

1316
01:31:52.219 --> 01:31:52.820
They need lighting.

1317
01:31:52.840 --> 01:31:53.320
Yeah.

1318
01:31:53.640 --> 01:31:54.621
And it's over the ocean.

1319
01:31:54.641 --> 01:31:55.301
Deadliest catch.

1320
01:31:55.462 --> 01:31:56.362
Right, right, right, right, right.

1321
01:31:56.582 --> 01:31:57.983
And it's over the ocean versus land.

1322
01:31:58.023 --> 01:31:58.243
Yeah.

1323
01:31:58.364 --> 01:31:58.704
And so,

1324
01:31:59.402 --> 01:32:00.363
That's like obvious.

1325
01:32:00.423 --> 01:32:01.484
Right.

1326
01:32:02.504 --> 01:32:04.126
Where did this come from?

1327
01:32:04.166 --> 01:32:05.867
There's no land there.

1328
01:32:05.887 --> 01:32:08.049
And they usually travel in packs.

1329
01:32:08.269 --> 01:32:09.169
So, you know.

1330
01:32:09.690 --> 01:32:15.534
It would be the fundamentals of how they do it make it actually very visible in this data set.

1331
01:32:15.594 --> 01:32:21.298
Congratulations to the NASA Black Marble team on the Golden Goose.

1332
01:32:22.880 --> 01:32:28.424
Again, we will be having a conversation with Miguel about some of these questions.

1333
01:32:28.924 --> 01:32:31.305
Very interested to dig into it, but I think we did great coverage.

1334
01:32:31.645 --> 01:32:37.686
We're going to move into, we somehow always find a way to be making two to three hour episodes.

1335
01:32:38.046 --> 01:32:38.806
Despite our...

1336
01:32:38.826 --> 01:32:40.967
This was supposed to be a show of one, sorry.

1337
01:32:41.027 --> 01:32:43.367
Despite our endeavor to try to make it tight.

1338
01:32:43.387 --> 01:32:45.487
But there's just so much interesting things to talk about.

1339
01:32:45.868 --> 01:32:50.989
We will be back with our last Golden Goose Award recipient.

1340
01:32:51.969 --> 01:32:55.570
And talking about something that's a little bit different than our first two stories.

1341
01:32:55.970 --> 01:32:57.290
Which is Disordered Systems.

1342
01:32:58.030 --> 01:33:12.382
Our final story is Everyday Phenomena and Finding Beauty in Nature's Disordered Systems, recognizing Dr. Sidney Nagel, the Steinfreuler Distinguished Service Professor of Physics at the University of Chicago.

1343
01:33:13.123 --> 01:33:18.145
Nagel has spent decades investigating things that we rarely stop to question every day.

1344
01:33:18.485 --> 01:33:21.166
Why does spilled coffee leave a ring?

1345
01:33:21.526 --> 01:33:26.968
Why can sand pour like a liquid and then sit in a pile and hold its shape?

1346
01:33:27.328 --> 01:33:30.150
A question I've never asked myself.

1347
01:33:30.570 --> 01:33:38.373
With the support of the National Science Foundation and the Department of Energy, he and his collaborators have turned everyday puzzles into science.

1348
01:33:39.253 --> 01:33:43.936
These experiments that reveal deeper physical principles.

1349
01:33:43.996 --> 01:33:49.059
And this is something I've certainly never thought about, but is a very interesting question.

1350
01:33:49.619 --> 01:33:54.582
What we're looking at here is a dried coffee drop about two centimeters across.

1351
01:33:55.382 --> 01:33:56.843
from the team's original research.

1352
01:33:57.923 --> 01:34:03.205
As we can notice, there are these darker regions around the edge.

1353
01:34:03.765 --> 01:34:08.827
Those particles all started out spread through the liquid in some contiguous fashion.

1354
01:34:08.867 --> 01:34:14.008
So why did they ultimately end up concentrating and creating that outline?

1355
01:34:14.568 --> 01:34:15.969
Nagel and his colleagues...

1356
01:34:17.886 --> 01:34:25.869
showed how evaporation can flow inside the drying drop, which I've never thought about, but now makes total sense.

1357
01:34:25.909 --> 01:34:29.810
Carrying particles through towards the edge and concentrating them there.

1358
01:34:30.590 --> 01:34:35.772
Understanding that process matters for things that actually matter to us every day, like inkjet printing.

1359
01:34:36.452 --> 01:34:44.875
If you want to drop a liquid ink on a paper, you need to know how it's going to ultimately land and spread so you get the actual letters that you want.

1360
01:34:46.096 --> 01:34:47.117
when the ink dries.

1361
01:34:47.517 --> 01:34:52.803
And across his work with droplets and grains, there's a bigger question.

1362
01:34:52.863 --> 01:35:03.155
How do we find predictable behavior in materials whose parts are not neatly arranged, which is a very, very interesting question.

1363
01:35:03.588 --> 01:35:06.989
Yeah, and physics has a reputation for not caring about that.

1364
01:35:07.770 --> 01:35:08.050
Okay?

1365
01:35:09.170 --> 01:35:16.293
Physicists have a reputation for not caring about neatly, or for not caring about not neatly arranged stuff.

1366
01:35:16.713 --> 01:35:16.893
Right?

1367
01:35:16.913 --> 01:35:20.255
We care about stuff that's near zero Kelvin, like quantum properties.

1368
01:35:20.275 --> 01:35:22.376
We've done the quantum computing stuff.

1369
01:35:22.556 --> 01:35:25.177
Like, nice systems that obey planets.

1370
01:35:26.300 --> 01:35:32.827
Nice laws, you know, the cow is a sphere is a favorite sort of comic that we talk about.

1371
01:35:33.228 --> 01:35:36.311
The formation of the universe, everything is nice and isotropic.

1372
01:35:36.691 --> 01:35:41.316
Collisions of black holes, again, black holes have no hairs and it's just like two little bodies.

1373
01:35:41.376 --> 01:35:46.782
Okay, fine, the equations of GR are a little chaotic, but it's not really...

1374
01:35:47.618 --> 01:35:49.299
Disordered in a sense.

1375
01:35:49.419 --> 01:35:54.781
And Sidney Nagel was curious about much more modest things in his life.

1376
01:35:55.161 --> 01:36:03.045
The texture of our world, the physics of a raindrop as it stretches and falls from an overhang, the spill of a coffee cup.

1377
01:36:03.790 --> 01:36:10.815
And in one of his interviews, he says that no one thought you'd find basic truths by looking at these sorts of things.

1378
01:36:10.915 --> 01:36:17.679
But the fact that we don't know how to think about answering these questions means that there's some fundamental physics that we don't know.

1379
01:36:18.420 --> 01:36:22.362
And those are the types of questions that he's trying to answer.

1380
01:36:22.863 --> 01:36:23.783
What's the mechanism?

1381
01:36:24.384 --> 01:36:30.708
What carries coffee particles to the edge to give us that nice black marking?

1382
01:36:30.728 --> 01:36:31.008
Yeah.

1383
01:36:31.220 --> 01:36:39.683
And the answers require us to look at flows, forces, geometries, and time scales that maybe our everyday intuition barely ever registers.

1384
01:36:40.423 --> 01:36:46.785
Now, Nagel is a physicist at the University of Chicago and one of the pioneers of a field called soft matter physics.

1385
01:36:46.885 --> 01:36:52.887
In 2023, he received the American Physical Society's Medal for Exceptional Advancement in Research.

1386
01:36:53.848 --> 01:36:58.429
And the work that we're going to follow spans about three decades of work, many collaborators.

1387
01:37:01.362 --> 01:37:05.885
in order from the falling drop in 1994 to the stuff that he's been doing in 2025.

1388
01:37:05.946 --> 01:37:08.948
We're going to start with the falling drop.

1389
01:37:09.168 --> 01:37:13.872
The first paper we will cover actually got him on the cover of science.

1390
01:37:13.972 --> 01:37:16.213
Still a great white buffalo for me.

1391
01:37:16.694 --> 01:37:17.374
Not just a cover.

1392
01:37:17.434 --> 01:37:19.136
I mean, I just want to get into science.

1393
01:37:20.457 --> 01:37:26.201
Many of you may have noticed in between us, someone who made it to the cover of nature,

1394
01:37:27.601 --> 01:37:29.902
happens to be behind us in studio.

1395
01:37:30.402 --> 01:37:32.703
But that's not what we're focused on this episode.

1396
01:37:32.723 --> 01:37:33.384
But it is there.

1397
01:37:33.424 --> 01:37:34.104
But it is there.

1398
01:37:34.244 --> 01:37:40.486
So he's been on the cover of Science, and this is what it's been about.

1399
01:37:40.946 --> 01:37:50.630
It's about how drops disattach or de-attach from a faucet.

1400
01:37:51.502 --> 01:38:00.395
You know, like imagine you've got your faucet and you turn it up to just a tiny bit so that like individual drops are coming out of the faucet.

1401
01:38:00.475 --> 01:38:01.156
You get a little drip.

1402
01:38:01.216 --> 01:38:02.138
You get a little drip, drip.

1403
01:38:03.039 --> 01:38:04.642
What is the physics behind that?

1404
01:38:05.122 --> 01:38:05.823
In the 1994 science paper...

1405
01:38:08.133 --> 01:38:11.515
It's called A Cascade of Structure in Drop Falling from a Faucet.

1406
01:38:11.615 --> 01:38:14.517
A great, great title, by the way.

1407
01:38:14.537 --> 01:38:15.718
We talk about titles a lot.

1408
01:38:15.738 --> 01:38:16.479
This is a good one.

1409
01:38:16.499 --> 01:38:17.019
On the show.

1410
01:38:17.099 --> 01:38:17.379
Yeah.

1411
01:38:17.940 --> 01:38:25.565
So what he ended up doing with his students there is photographing drops made from mixtures of water and glycerol.

1412
01:38:26.185 --> 01:38:26.886
Why glycerol?

1413
01:38:26.946 --> 01:38:32.009
Because glycerol has a really high viscosity compared to water.

1414
01:38:32.950 --> 01:38:37.473
The question they're asking is, does viscosity affect the dynamics of how this drop

1415
01:38:38.916 --> 01:38:42.518
detaches from the parent drop that's on the faucet.

1416
01:38:42.618 --> 01:38:44.078
How drippy is my drip?

1417
01:38:44.178 --> 01:38:53.563
Yeah, and this is a classic golden goose type of word because at first anyone would be like, really?

1418
01:38:54.263 --> 01:38:54.923
Why do we care?

1419
01:38:54.963 --> 01:39:00.066
Well, there's a lot of really interesting physics that's happening there that I didn't quite appreciate, I have to say.

1420
01:39:00.106 --> 01:39:02.887
So let's take a look at this particular image, okay?

1421
01:39:03.127 --> 01:39:05.008
This is figure one from their paper.

1422
01:39:06.509 --> 01:39:07.049
In A...

1423
01:39:07.445 --> 01:39:08.265
That's pure water.

1424
01:39:08.886 --> 01:39:09.146
Okay.

1425
01:39:09.606 --> 01:39:12.687
And you can see that the drop is kind of a deattaching.

1426
01:39:12.747 --> 01:39:16.489
And these are all photos that are taken with a high speed camera.

1427
01:39:16.529 --> 01:39:20.531
There's a strobe light behind them that's illuminating the stuff.

1428
01:39:20.551 --> 01:39:23.972
And then the camera is kind of taking photo photographs that way.

1429
01:39:26.453 --> 01:39:29.694
In a, you're seeing a water droplet disattached.

1430
01:39:29.714 --> 01:39:30.595
This is pure water.

1431
01:39:30.915 --> 01:39:35.557
And you can see that the sort of trail, the part of the part

1432
01:39:35.998 --> 01:39:42.124
that's hanging between the faucet and the droplet that's... Before it... Before it pinches off.

1433
01:39:42.144 --> 01:39:43.104
Before it pinches off.

1434
01:39:43.124 --> 01:39:45.126
Yeah, before it deattaches is very short.

1435
01:39:45.647 --> 01:39:48.029
In E, that's pure glycerol.

1436
01:39:48.990 --> 01:39:50.871
Look at how large the...

1437
01:39:54.530 --> 01:40:14.364
trail is yeah the little spittle yeah a little which and again would you think who who thinks yeah and and this shows clearly that the more viscous the liquid the neck right we're gonna call it the neck the neck is gonna be very long and thin and near the the lower end the smaller the neck appeared

1438
01:40:15.333 --> 01:40:16.154
This is the key thing.

1439
01:40:16.615 --> 01:40:24.008
He showed that as the drop approaches the spot of separation, the neck doesn't simply disappear.

1440
01:40:24.609 --> 01:40:27.855
There's a lot of very cool dynamics that are happening right at the instant

1441
01:40:28.364 --> 01:40:30.626
where the detachment is happening.

1442
01:40:30.906 --> 01:40:32.507
And it has everything to do with viscosity.

1443
01:40:32.947 --> 01:40:41.233
The reason why we want to start introducing glycerol is because it increases the viscosity, but it doesn't increase the surface tension in the same way.

1444
01:40:41.533 --> 01:40:45.656
So now we're sort of disentangling viscosity versus surface tension.

1445
01:40:45.716 --> 01:40:50.260
And you would think that surface tension is everything when it comes to making a drop what it is.

1446
01:40:50.280 --> 01:40:56.584
I mean, that's sort of what we learn about in high school and undergrad physics is like surface tension is what keeps a drop spherical.

1447
01:40:59.168 --> 01:41:04.419
water wants to minimize the amount of surface area and the best way to do that is turn into a sphere.

1448
01:41:04.902 --> 01:41:12.846
Is this when you've seen probably the memes of people playing the game where they try to fill the cup of water to the top and then whoever makes the first drip go over the edge?

1449
01:41:12.866 --> 01:41:13.067
Yes.

1450
01:41:13.127 --> 01:41:18.650
And then there's a little bit and it's like everyone's always so shocked at how much water is still in the cup before it drips over the edge.

1451
01:41:18.710 --> 01:41:24.293
That's a classic experiment where you get like a penny and you try to jam as many drops as possible.

1452
01:41:24.453 --> 01:41:24.613
Yeah.

1453
01:41:24.633 --> 01:41:24.773
Right?

1454
01:41:25.093 --> 01:41:32.818
And the classic explanation, and it's true, is that it's because of the surface tension of the water that you don't want – the water doesn't want to break –

1455
01:41:33.598 --> 01:41:36.359
the drop.

1456
01:41:37.599 --> 01:41:40.940
This experiment is showing that surface tension is not everything.

1457
01:41:40.980 --> 01:41:42.460
Viscosity has a lot to do with it.

1458
01:41:43.100 --> 01:41:47.582
We've talked a lot about viscosity in the previous episode, so I think you'll get a kick out of this.

1459
01:41:48.722 --> 01:41:57.324
Here we're seeing a mixture of glycerol and water, and what they're doing is zooming in on that moment of separation.

1460
01:41:58.010 --> 01:42:06.141
It's kind of like in the last episode when we talked about how we can rescale time as a countdown clock to some critical phenomenon.

1461
01:42:06.502 --> 01:42:14.312
In this case, there's a critical phenomenon that's happening, which is the moment that the water droplet separates from its parent water droplet.

1462
01:42:15.333 --> 01:42:19.895
As we get closer and closer to that moment, what happens to the neck?

1463
01:42:20.516 --> 01:42:20.756
Okay?

1464
01:42:21.056 --> 01:42:23.797
On the left-hand side, we're seeing a sort of zoomed-out picture.

1465
01:42:25.338 --> 01:42:31.481
And the other two, B and C, we're seeing timestamps getting closer and closer to that moment of separation.

1466
01:42:31.821 --> 01:42:33.122
And here's what I want you to notice.

1467
01:42:33.822 --> 01:42:35.743
In the middle, you've got a single neck.

1468
01:42:37.544 --> 01:42:41.766
As we get closer, the single neck forms an even smaller neck.

1469
01:42:42.607 --> 01:42:42.907
You see?

1470
01:42:42.927 --> 01:42:43.447
Mm-hmm.

1471
01:42:44.765 --> 01:42:51.370
it's kind of like a fractally self-similar type of behavior that's happening.

1472
01:42:52.151 --> 01:42:57.295
In A, there's a one long neck, but the one long neck creates a smaller neck.

1473
01:42:57.695 --> 01:43:00.718
And then that smaller neck is going to create an even smaller neck.

1474
01:43:01.603 --> 01:43:01.823
Right?

1475
01:43:02.224 --> 01:43:04.686
There's this critical phenomenon that's happening.

1476
01:43:04.906 --> 01:43:06.548
It's like a fractal-like cascade.

1477
01:43:07.489 --> 01:43:08.350
I'm going to be annoying.

1478
01:43:08.390 --> 01:43:10.973
As tau approaches zero, we're going to infinity.

1479
01:43:11.113 --> 01:43:11.593
Yeah, yeah.

1480
01:43:11.633 --> 01:43:17.439
The number, yeah, you could think about the number of cascades is going to go to infinity in some sense.

1481
01:43:17.459 --> 01:43:17.760
Right, right.

1482
01:43:17.780 --> 01:43:22.405
Now, obviously, this is, it's not going to happen until infinity because just like the Navier-Stokes equations.

1483
01:43:22.425 --> 01:43:22.605
Yes.

1484
01:43:22.905 --> 01:43:28.728
At some point, you're going to run into the molecular barrier, and you're going to be like, you can't get smaller than the molecules and things like that.

1485
01:43:28.788 --> 01:43:37.173
But it's still very cool that, you know, for the regimes that we care about, there is this self-similar type of response that's happening.

1486
01:43:37.553 --> 01:43:39.454
And it's so funny, we just did the Navier-Stokes episode.

1487
01:43:39.474 --> 01:43:40.435
And the scale is shrinking.

1488
01:43:40.715 --> 01:43:45.578
And so in that paper, one, he showed the experimental evidence of this type of stuff happening.

1489
01:43:45.878 --> 01:43:48.440
And then he solved numerically the Navier-Stokes equations.

1490
01:43:48.820 --> 01:43:52.602
And he showed that this does happen, but only if there is noise in the system.

1491
01:43:52.982 --> 01:43:54.202
This is what's kind of interesting.

1492
01:43:54.883 --> 01:44:08.527
If there's no noise, then the neck becomes this self-similar sort of power law thing that doesn't have these discrete types of, you know, tuck, tuck, and then it becomes smaller and smaller and smaller.

1493
01:44:08.567 --> 01:44:10.127
It's not this rhythmic construction.

1494
01:44:10.287 --> 01:44:17.030
But if you introduce noise into the system, then on the right-hand side, you see the self-similar thing where you've got a big thing and then you zoom in.

1495
01:44:17.330 --> 01:44:18.250
That's the second panel.

1496
01:44:18.470 --> 01:44:26.373
And then the second panel has an inset that zooms in over there and you see an even more self-similar sort of tiny neck happening.

1497
01:44:26.513 --> 01:44:31.075
And the details change from run to run, but you can repeatedly create these necks.

1498
01:44:31.255 --> 01:44:32.356
And that is robust.

1499
01:44:32.436 --> 01:44:33.496
The details can change.

1500
01:44:33.856 --> 01:44:35.357
But the fact that these necks exist...

1501
01:44:36.258 --> 01:44:38.519
And they're self-similar, smaller and smaller and smaller.

1502
01:44:38.819 --> 01:44:40.059
That is a robust phenomenon.

1503
01:44:40.719 --> 01:44:41.880
We love to see it.

1504
01:44:42.480 --> 01:44:50.743
Again, it's these signals in observation of something interesting.

1505
01:44:51.203 --> 01:45:03.647
And then a lot of times I feel like when we talk about it, it's like who zoomed in more to just double click to make sure that just because we see it like this on the outside, it actually looks like that on the inside.

1506
01:45:03.667 --> 01:45:03.767
Yeah.

1507
01:45:03.787 --> 01:45:03.947
Yeah.

1508
01:45:04.507 --> 01:45:09.912
And I mean, from a physicist perspective, this is already very cool.

1509
01:45:10.092 --> 01:45:10.352
Yeah.

1510
01:45:10.612 --> 01:45:10.872
Okay.

1511
01:45:10.973 --> 01:45:17.338
Like this is a phenomenon that like you're seeing like a fractally nature happening in like the disentangling of droplets.

1512
01:45:17.418 --> 01:45:18.619
That's kind of cool.

1513
01:45:18.739 --> 01:45:21.201
From practical standpoints, it's not just like inkjet printers.

1514
01:45:21.221 --> 01:45:26.625
You can have sprays, microfluidic devices, which are used a lot of times in biomedical applications.

1515
01:45:27.046 --> 01:45:32.290
When you want to like separate out cells, you want to separate out DNA, small little things.

1516
01:45:33.416 --> 01:45:37.399
micro thingies for a lack of a word.

1517
01:45:37.419 --> 01:45:38.700
We got a thingy in this episode.

1518
01:45:38.741 --> 01:45:42.063
You did four hours last time with not a single thingy and people are mad.

1519
01:45:42.083 --> 01:45:43.504
Yeah, people are mad.

1520
01:45:43.524 --> 01:45:47.067
Well, you know, there's tremendous applications here.

1521
01:45:47.347 --> 01:45:53.432
So it is quite interesting that like such rich phenomenon can come from something that never really thought about.

1522
01:45:53.452 --> 01:45:56.615
1997, this is the next paper that we're going to talk about.

1523
01:45:56.635 --> 01:45:58.116
This is the coffee ring effect.

1524
01:45:58.676 --> 01:46:00.398
There's a whole Wikipedia article about it.

1525
01:46:00.658 --> 01:46:00.798
Mm-hmm.

1526
01:46:01.543 --> 01:46:10.928
Um, the idea is that when you have coffee stains and they dry out, the edges of the coffee stains are always much darker than the interior.

1527
01:46:11.208 --> 01:46:14.170
I've actually always wondered this, but not actually thought about it.

1528
01:46:14.190 --> 01:46:14.410
Yeah.

1529
01:46:14.470 --> 01:46:15.271
I've like noticed it.

1530
01:46:15.371 --> 01:46:15.591
Yeah.

1531
01:46:15.631 --> 01:46:15.831
Yeah.

1532
01:46:15.851 --> 01:46:15.951
Yeah.

1533
01:46:16.191 --> 01:46:20.954
And then, and then now because of the golden goose, I'm like diving deep and I'm like, oh, this is quite interesting.

1534
01:46:21.614 --> 01:46:24.356
Um, so they came out with a paper in nature.

1535
01:46:24.956 --> 01:46:31.418
that had another amazing title, capillary flow as the cause of ring stains from dried liquid drops.

1536
01:46:32.759 --> 01:46:33.679
Here is the idea.

1537
01:46:34.280 --> 01:46:43.703
So let's imagine a liquid drop with a bunch of coffee stains, like coffee, liquid drop with coffee.

1538
01:46:45.524 --> 01:46:47.785
What is going to happen as it evaporates?

1539
01:46:49.485 --> 01:46:50.926
That's the question we're trying to answer.

1540
01:46:51.266 --> 01:46:51.506
Okay.

1541
01:46:51.966 --> 01:46:54.287
So if we look at a coffee stain...

1542
01:46:54.973 --> 01:46:58.615
On the upper left, that's a liquid drop, let's say.

1543
01:47:00.517 --> 01:47:11.984
On the bottom left, what they've done is introduce microspheres into the coffee, and they image it to track how the microspheres are moving around in the coffee.

1544
01:47:12.004 --> 01:47:17.007
And what they notice is all of the microspheres are moving out to the edge.

1545
01:47:17.468 --> 01:47:22.471
There is movement that is happening towards the edge of this coffee stain.

1546
01:47:22.551 --> 01:47:23.652
Now, why is that?

1547
01:47:24.546 --> 01:47:25.467
Here is the idea.

1548
01:47:26.748 --> 01:47:36.496
Picture the drop as a low dome, like, you know, the experiment that we were talking about with the penny, and then you drop, like, water on it.

1549
01:47:36.877 --> 01:47:39.479
Similarly, coffee just on a piece of paper or something like that.

1550
01:47:39.559 --> 01:47:43.862
The boundary is going to be established based on how much coffee you spilled.

1551
01:47:44.823 --> 01:47:50.148
But as it evaporates, the coffee stain doesn't shrink.

1552
01:47:50.168 --> 01:47:50.328
Right.

1553
01:47:51.580 --> 01:47:53.701
the area stays exactly the same.

1554
01:47:54.221 --> 01:47:55.401
So what's going to happen now?

1555
01:47:56.462 --> 01:48:06.365
If the area stays exactly the same, the liquid is leaving, and by conservation of mass, you need to replace the liquid at the edges.

1556
01:48:08.185 --> 01:48:08.385
Right?

1557
01:48:08.746 --> 01:48:13.287
And so if we go back to photo number 58, I just want to show you the model that they have.

1558
01:48:14.888 --> 01:48:15.908
Yep, and there.

1559
01:48:16.682 --> 01:48:17.963
The liquid is leaving, right?

1560
01:48:18.023 --> 01:48:20.004
It's evaporating on the right-hand side.

1561
01:48:20.024 --> 01:48:23.085
They've got this figure that shows the liquid is evaporating.

1562
01:48:23.125 --> 01:48:31.209
But because the liquid is evaporating, you need a flow from the interior to the exterior that is going to replace that liquid.

1563
01:48:31.769 --> 01:48:35.331
Because the surface area of the drop has to remain the same.

1564
01:48:36.996 --> 01:48:38.037
Oh, man.

1565
01:48:38.077 --> 01:48:38.277
Right?

1566
01:48:38.637 --> 01:48:42.679
And because of that, you see these microspheres move towards the edge.

1567
01:48:42.759 --> 01:48:49.683
Well, that also means all of the coffee grounds or whatever the stuff that makes the coffee is moving towards the edge.

1568
01:48:49.963 --> 01:48:56.166
And by the time that everything evaporates, all of the stuff has concentrated to the edge.

1569
01:48:56.186 --> 01:48:56.727
The edge.

1570
01:48:57.827 --> 01:48:58.047
Right.

1571
01:48:58.468 --> 01:48:59.849
It makes total sense.

1572
01:49:00.209 --> 01:49:02.191
And it's like obvious in retrospect.

1573
01:49:02.211 --> 01:49:02.471
Yeah.

1574
01:49:02.491 --> 01:49:02.671
Yeah.

1575
01:49:02.691 --> 01:49:02.951
Yeah.

1576
01:49:02.991 --> 01:49:05.994
But it's such a, it's such a clean argument.

1577
01:49:06.014 --> 01:49:06.894
A hundred percent.

1578
01:49:07.135 --> 01:49:07.415
Yeah.

1579
01:49:07.675 --> 01:49:09.056
It's a hundred percent.

1580
01:49:09.116 --> 01:49:13.059
This is something that I could imagine in an advanced undergrad class of fluid mechanics.

1581
01:49:13.139 --> 01:49:14.360
Like this is something you solve.

1582
01:49:14.420 --> 01:49:14.660
Yeah.

1583
01:49:14.680 --> 01:49:17.403
Like this is a homework problem that would be very nice.

1584
01:49:18.163 --> 01:49:19.004
That's so good.

1585
01:49:19.024 --> 01:49:22.967
You know, the contact line is pinned basically.

1586
01:49:22.987 --> 01:49:24.288
And then the evaporation is,

1587
01:49:24.587 --> 01:49:30.132
removes the solvent, so mass conservation creates this outward flow.

1588
01:49:31.033 --> 01:49:32.053
This is the first treatment.

1589
01:49:32.073 --> 01:49:44.104
Their later treatment, which is this paper over here, this was in the Physical Review E in 2000, three years later, they had a much longer treatment where they developed the flow and deposit theory in much greater detail.

1590
01:49:44.604 --> 01:49:51.710
And what you can do is you can change the pinning, you can change particle interactions, you can change the internal flow, and then you can change the pattern accordingly.

1591
01:49:53.353 --> 01:49:56.456
So they developed it a lot more further in that paper.

1592
01:49:57.136 --> 01:50:05.384
And I think that quantification is very, very important because if you're printing with some kind of functional ink, right, you might want a uniform deposit.

1593
01:50:06.164 --> 01:50:10.508
If you're trying to print with an ink printer, you don't want only the edges to show up.

1594
01:50:10.548 --> 01:50:13.451
So the next question obviously becomes, how do we prevent this?

1595
01:50:14.308 --> 01:50:14.908
from happening.

1596
01:50:15.549 --> 01:50:25.197
And the way we prevent this from happening is we change the characteristics of the coloring, the coffee grounds or whatever, that is inside the liquid.

1597
01:50:27.058 --> 01:50:31.622
What we end up doing is, in order to suppress this outward flow,

1598
01:50:32.402 --> 01:50:47.073
You can engineer the shape of the particles inside that will create traffic jams inside the flow that will make the particles, your ink or whatever thing that is coloring it, stay put and not actually move.

1599
01:50:47.093 --> 01:50:49.015
So maybe the water will move, right?

1600
01:50:49.075 --> 01:50:49.675
The other fluid.

1601
01:50:50.016 --> 01:50:53.338
But the stuff that is creating the coloring doesn't move because it's jammed up.

1602
01:50:53.979 --> 01:50:59.563
You're sort of engineering the outcome by creating differences in the substance such that...

1603
01:51:00.379 --> 01:51:03.482
the only outcome that arises is the desired outcome.

1604
01:51:04.823 --> 01:51:05.923
This is interesting.

1605
01:51:06.204 --> 01:51:09.206
Yeah, no, this is good.

1606
01:51:10.327 --> 01:51:17.052
So in 2023, sorry, 2003, they wanted to start understanding jamming because they've sort of introduced this idea of jamming and fluids.

1607
01:51:17.412 --> 01:51:18.734
So what does jamming actually do?

1608
01:51:18.774 --> 01:51:20.995
So imagine this is what jamming actually means.

1609
01:51:21.035 --> 01:51:26.199
Like, have you ever held those hourglasses where sometimes the sand gets stuck?

1610
01:51:26.420 --> 01:51:26.560
Mm-hmm.

1611
01:51:27.180 --> 01:51:27.980
Now, it's sand.

1612
01:51:28.020 --> 01:51:37.905
It should flow like a fluid, but sometimes the sand gets stuck, and you've got to jiggle it a little bit to have the sand fall from the top of the hourglass to the bottom of the hourglass.

1613
01:51:38.265 --> 01:51:39.966
The question he's asking is why, right?

1614
01:51:40.106 --> 01:51:45.028
If the spheres occupy only a small fraction of the volume, then why is it jamming?

1615
01:51:45.048 --> 01:51:46.329
Well, that's exactly the case.

1616
01:51:46.969 --> 01:51:53.872
If the spheres are occupying only a small fraction of the volume, then they can be rearranged like it's nobody's business.

1617
01:51:53.932 --> 01:51:54.332
Okay, yeah.

1618
01:51:56.890 --> 01:52:05.836
of how much sphere there is versus how much empty space there is, if that is higher than some certain amount, there's going to be jamming.

1619
01:52:06.277 --> 01:52:11.901
And he quantified exactly what the statistics of that jamming is.

1620
01:52:13.708 --> 01:52:18.652
The idea being we have in an hourglass, there's a narrower space where the jamming arises.

1621
01:52:18.892 --> 01:52:29.981
And so the actual dynamics of the amount of space those smaller particles are in, there's some XY, smaller space, higher jamming potential.

1622
01:52:30.761 --> 01:52:38.487
And we now have effectively the formula to understand at the rate at which jamming increases.

1623
01:52:38.547 --> 01:52:41.770
Yeah, and that's why if you have hourglasses that have a fluid in it,

1624
01:52:42.488 --> 01:52:47.230
you almost never have jamming because the fluid creates a lower packing ratio one.

1625
01:52:47.750 --> 01:52:52.572
And the fluid has tiny bits of turbulence that'll like move around the jamming, right?

1626
01:52:52.632 --> 01:52:57.674
This is the kind of physics that he was focused on in this next paper that we're talking about here.

1627
01:52:58.074 --> 01:53:09.899
It's in the physical review E in 2003, jamming at zero temperature and zero applied stress, the epitome of disorder, zero temperature, meaning no jiggling and zero applied stress, meaning you're not like pushing it in any way.

1628
01:53:10.554 --> 01:53:12.035
How does the jamming actually happen?

1629
01:53:12.776 --> 01:53:15.598
And the control knob happens to be the packing fraction.

1630
01:53:16.098 --> 01:53:17.920
Now, what are the practical uses?

1631
01:53:18.500 --> 01:53:21.342
This is one of the craziest things that I've seen, actually.

1632
01:53:21.762 --> 01:53:23.124
And this is actually quite old.

1633
01:53:24.805 --> 01:53:34.232
This next paper that I'm about to show you, 2010 PNAS out of Cornell University, they used this idea to create a robot that can pick anything up.

1634
01:53:35.572 --> 01:53:35.952
Okay.

1635
01:53:36.193 --> 01:53:36.453
Okay.

1636
01:53:36.733 --> 01:53:41.277
So here we're looking at a video of a robot that's built from a human hand.

1637
01:53:41.717 --> 01:53:43.819
It can't pick up those random items, right?

1638
01:53:43.859 --> 01:53:54.028
But instead, they've got a sphere with a bunch of coffee grounds that they can control whether it jams or not.

1639
01:53:55.742 --> 01:54:09.432
The key idea here is that because you've got a sphere with all of these little tiny grains, and you can tune the transition between whether it's jamming or not, using some kind of external drive,

1640
01:54:10.697 --> 01:54:12.558
You can now fit that sphere.

1641
01:54:12.799 --> 01:54:16.762
You can fit it to any arbitrary shape, and you can pick up anything you want.

1642
01:54:17.302 --> 01:54:19.244
Here it's picking up an LED light, putting it inside.

1643
01:54:19.264 --> 01:54:20.104
This is brilliant.

1644
01:54:20.144 --> 01:54:21.825
Yeah, here it's picking up a spring.

1645
01:54:22.666 --> 01:54:24.027
This is notoriously hard to do.

1646
01:54:24.167 --> 01:54:31.073
And we used to think that the solution to creating random hands is like emulating the human hands.

1647
01:54:31.133 --> 01:54:34.035
But, mate, we've had millions of years of evolution.

1648
01:54:34.055 --> 01:54:34.155
Yeah.

1649
01:54:34.715 --> 01:54:34.916
Right?

1650
01:54:34.996 --> 01:54:36.457
It's incredibly difficult to do this.

1651
01:54:36.537 --> 01:54:39.661
I mean, now, with the advent of AI and modern robotics, sure.

1652
01:54:39.941 --> 01:54:41.102
But this is 2010.

1653
01:54:41.142 --> 01:54:42.343
This is 16 years ago.

1654
01:54:43.044 --> 01:54:43.765
They had figured out

1655
01:54:44.483 --> 01:54:47.566
It's also clever how they got there because it's not intuitive to me that.

1656
01:54:48.226 --> 01:54:48.446
Yeah.

1657
01:54:48.486 --> 01:54:50.608
A ball, a ball of coffee grounds.

1658
01:54:50.648 --> 01:54:51.429
Cause that's what this is.

1659
01:54:51.529 --> 01:54:51.689
Right.

1660
01:54:51.729 --> 01:54:53.050
It's a ball of coffee grounds.

1661
01:54:53.410 --> 01:54:53.570
Right.

1662
01:54:53.731 --> 01:54:54.751
It's like a hacky sack.

1663
01:54:55.432 --> 01:55:00.976
It's, it's a surface that basically can arbitrarily based on whatever the object is that it comes in contact with.

1664
01:55:01.116 --> 01:55:01.337
Yeah.

1665
01:55:01.357 --> 01:55:03.799
Create the exact construction in order to be able to lift it.

1666
01:55:03.899 --> 01:55:04.099
Yeah.

1667
01:55:04.219 --> 01:55:04.399
Yeah.

1668
01:55:04.459 --> 01:55:06.241
And there's, there's other examples in here.

1669
01:55:06.641 --> 01:55:07.562
Those are the objects.

1670
01:55:07.622 --> 01:55:09.323
This is a standard hand back in the day.

1671
01:55:09.343 --> 01:55:09.563
Yeah.

1672
01:55:09.703 --> 01:55:10.604
It couldn't lift those up.

1673
01:55:10.624 --> 01:55:10.784
Right.

1674
01:55:10.804 --> 01:55:13.446
But this, a single thing is just going to go and lift it up.

1675
01:55:13.687 --> 01:55:14.107
Brilliant.

1676
01:55:14.307 --> 01:55:41.463
isn't that cool that is very and this is directly from that jamming transition yeah that's very that we had understood using using this math that is very cool because again it's it wasn't intuitive to me that that's where you go with it but obviously it makes again yeah in retrospect it's like oh yeah yeah exactly so the next paper that we're going to talk about is 2005 this is in physical review letters um and how it's about how

1677
01:55:42.300 --> 01:55:45.496
the air decides whether a drop splashes or not.

1678
01:55:46.864 --> 01:55:49.125
This was a little bit non-intuitive to me, actually.

1679
01:55:49.165 --> 01:55:49.606
Can I just say...

1680
01:55:49.846 --> 01:55:50.766
He's on a heater.

1681
01:55:50.906 --> 01:55:52.848
He's on a heater.

1682
01:55:52.948 --> 01:55:53.348
Yeah.

1683
01:55:54.529 --> 01:55:56.030
Just in terms of, like, the questions.

1684
01:55:56.090 --> 01:55:56.510
Yeah.

1685
01:55:56.650 --> 01:55:58.271
And they're, like, fundamental questions.

1686
01:55:58.291 --> 01:56:01.933
The drip, the stain, and now air.

1687
01:56:02.053 --> 01:56:02.333
Yeah.

1688
01:56:02.373 --> 01:56:03.194
Like, these are...

1689
01:56:03.554 --> 01:56:05.755
These are, like, fundamental questions that people ask.

1690
01:56:06.376 --> 01:56:07.937
Like, what causes a splash?

1691
01:56:08.117 --> 01:56:09.238
And then you just move on.

1692
01:56:09.258 --> 01:56:10.398
Right.

1693
01:56:10.438 --> 01:56:11.699
This guy's like, no, no, no, really.

1694
01:56:11.739 --> 01:56:16.442
Like, I'm gonna... And so, here's an experiment that they do where they have...

1695
01:56:17.208 --> 01:56:23.909
Drops of water splash onto, let's say, a Petri dish or a table or something, but in different air pressures.

1696
01:56:24.210 --> 01:56:25.850
So the ambient air pressure is different.

1697
01:56:26.170 --> 01:56:31.391
And what they notice is that at atmosphere, that's 100 kilopascals at the top, you get a lot of splashing.

1698
01:56:31.691 --> 01:56:34.592
Those are all the drips that you're seeing, like, in the third column.

1699
01:56:35.272 --> 01:56:39.413
But as you lower the pressure, no more splashing.

1700
01:56:40.527 --> 01:56:52.990
This is very non-intuitive to me because I would have thought from our Navier-Stokes episode and from what I know from fluid mechanics that, you know, the Navier-Stokes equations tell me everything about what the fluid is going to do, right?

1701
01:56:53.950 --> 01:56:55.070
I know the viscosity.

1702
01:56:55.130 --> 01:56:56.150
I know the pressure around.

1703
01:56:56.471 --> 01:56:58.851
But the interface pressure, right?

1704
01:56:59.911 --> 01:57:07.153
The difference in pressure between the ambient air outside and that interface is actually what's causing this.

1705
01:57:07.867 --> 01:57:08.087
Right?

1706
01:57:08.987 --> 01:57:11.328
This is a reworking of some fundamental physics.

1707
01:57:12.168 --> 01:57:15.009
You know exactly what my brain is going to when I see this.

1708
01:57:15.690 --> 01:57:18.650
This is how the UAPs do their transmedium travel.

1709
01:57:18.871 --> 01:57:20.871
Yeah.

1710
01:57:20.931 --> 01:57:22.292
There it is.

1711
01:57:22.332 --> 01:57:25.873
Just like I didn't have a thingy, you didn't have a UAP in a while.

1712
01:57:25.893 --> 01:57:27.213
I haven't had a UAP in a while.

1713
01:57:27.233 --> 01:57:28.334
I had to squeeze it in there.

1714
01:57:28.414 --> 01:57:28.854
There it is.

1715
01:57:29.234 --> 01:57:32.495
Yeah, but I think it's pretty cool because like...

1716
01:57:33.523 --> 01:57:34.364
I wouldn't have thought of that.

1717
01:57:34.444 --> 01:57:37.068
I wouldn't have thought that pressure had something to do with splashing.

1718
01:57:37.228 --> 01:57:37.748
Yeah.

1719
01:57:37.969 --> 01:57:39.230
But it turns out it does.

1720
01:57:39.350 --> 01:57:41.093
And the authors themselves, they point to...

1721
01:57:42.248 --> 01:57:51.392
applications of this, they show that like, you know, fuel sprays, spray drying, inkjet printing, washing and coating, their settings where splash behavior matters.

1722
01:57:52.453 --> 01:58:04.618
And to say that, you know, putting all printing lines under a vacuum, maybe that's not the idea, but the experiment gives an engineer a variable that they previously did not account for, right?

1723
01:58:04.758 --> 01:58:08.380
No one was thinking about ambient pressure when it comes to splashing.

1724
01:58:08.700 --> 01:58:14.244
And again, the paper is not just, oh, there's a threshold pressure, right?

1725
01:58:14.284 --> 01:58:19.508
He goes into the physics and he talks about why exactly the pressure matters.

1726
01:58:19.809 --> 01:58:20.969
And there's a scaling argument.

1727
01:58:20.989 --> 01:58:24.272
There's a threshold of pressure for different viscosities.

1728
01:58:24.598 --> 01:58:29.480
Like for a certain viscosity, a certain pressure is the threshold pressure where it goes splash, no splash.

1729
01:58:29.821 --> 01:58:33.302
I reduce the viscosity, the pressure threshold changes, right?

1730
01:58:33.362 --> 01:58:35.643
And there's a systematic relationship there.

1731
01:58:36.984 --> 01:58:53.092
The cookbook is being written piece by piece, and engineers have hopefully picked up on a lot of – because, again, I can immediately, as has already happened and will continue to happen, see the practical applications to these fundamental understandings.

1732
01:58:53.272 --> 01:58:53.873
Yeah, yeah.

1733
01:58:54.133 --> 01:58:57.857
And that brings us to our final paper, 2025.

1734
01:58:58.577 --> 01:59:05.725
This is a training and retraining liquid crystal elastomer metamaterials for pluripotent functionality.

1735
01:59:06.365 --> 01:59:10.850
That word pluripotent, I've only seen in the context of pluripotent

1736
01:59:11.247 --> 01:59:12.128
stem cells.

1737
01:59:12.208 --> 01:59:12.468
Same.

1738
01:59:12.508 --> 01:59:12.708
Right?

1739
01:59:12.788 --> 01:59:16.570
And so now I'm seeing it in metamaterials, and I was obviously peaked.

1740
01:59:17.010 --> 01:59:19.272
I was like, what exactly are we talking about?

1741
01:59:19.312 --> 01:59:24.875
So this is a paper about disordered lattices cut from liquid crystal elastomers.

1742
01:59:24.995 --> 01:59:27.556
And elastomer is like a rubber-like polymer.

1743
01:59:27.716 --> 01:59:32.619
Okay, so it's like it has some kind of deformability in some sense.

1744
01:59:32.759 --> 01:59:39.263
The liquid crystal component is because you've got all of these little crystal subunits

1745
01:59:39.883 --> 01:59:42.426
that can be kind of liquid and that they can move around.

1746
01:59:42.746 --> 01:59:46.610
And that's how we get LCDs, like LCD displays, liquid crystal displays.

1747
01:59:46.630 --> 01:59:55.098
That's because you have these pneumatic liquid crystals that can change orientation based on some type of electric field that will let through light and block light.

1748
01:59:55.138 --> 01:59:56.560
That's how we get the calculator.

1749
01:59:56.960 --> 01:59:57.801
LCD displays.

1750
01:59:58.061 --> 02:00:02.606
So it's a crystal-like nature that is not rigid but can sort of move around.

1751
02:00:02.646 --> 02:00:04.968
That's why it's liquid in some sense.

1752
02:00:05.368 --> 02:00:08.251
So the network is a sheet of these connected struts and nodes.

1753
02:00:08.691 --> 02:00:11.834
And the researchers make a flexible irregular lattice.

1754
02:00:11.854 --> 02:00:12.455
Yeah.

1755
02:00:12.695 --> 02:00:17.481
And they hold it compressed so it adapts to this compressed environment.

1756
02:00:17.722 --> 02:00:20.606
And then they test whether it bends differently afterwards.

1757
02:00:21.106 --> 02:00:22.568
So there's a memory here.

1758
02:00:22.708 --> 02:00:25.112
And that's where the pluripotent part comes in.

1759
02:00:25.252 --> 02:00:25.552
Okay.

1760
02:00:25.792 --> 02:00:26.033
Okay?

1761
02:00:26.273 --> 02:00:27.455
Because you can change...

1762
02:00:28.295 --> 02:00:34.799
how the material is going to behave based on the stresses and the environment that you put it through.

1763
02:00:34.839 --> 02:00:42.863
Very similar to how in pluripotent stem cells, you can change what the stem cells become based on the types of transcription factors

1764
02:00:43.280 --> 02:00:46.103
and the environment and the genetics that you expose these stem cells.

1765
02:00:46.163 --> 02:00:57.713
The same stem cell could go on and become a bone marrow or a skin cell or a blood cell, so on and so forth, based on the types of genetic factors that you put it.

1766
02:00:57.853 --> 02:01:03.518
We're giving it multifunctionality based on environmental factors, which obviously has...

1767
02:01:05.173 --> 02:01:06.013
It makes it a lot.

1768
02:01:06.333 --> 02:01:06.953
You could do more.

1769
02:01:07.294 --> 02:01:07.754
Exactly.

1770
02:01:07.774 --> 02:01:08.014
Yeah.

1771
02:01:08.274 --> 02:01:09.594
So let's look at figure number one.

1772
02:01:09.694 --> 02:01:09.954
Okay.

1773
02:01:09.994 --> 02:01:12.055
This is what the lattice is made out of.

1774
02:01:12.095 --> 02:01:18.956
The blue elongated shapes there, they're regularly aligned liquid crystal units within some kind of stretchy polymer.

1775
02:01:19.516 --> 02:01:19.756
Okay.

1776
02:01:20.097 --> 02:01:25.578
Polydomain means different little regions that have different orientations to begin with.

1777
02:01:25.618 --> 02:01:29.179
That's, again, akin to like the liquid crystal displays that we were talking about.

1778
02:01:29.619 --> 02:01:30.659
And this is again in PNAS.

1779
02:01:35.380 --> 02:02:01.550
elastomers so they can like actually interact with one another but not be so rigid that they can't move around okay and on the right hand side we're actually seeing how this elastomer interacts with this environment and remembers the environment that it was in before training on the top row you you push down and the sides move outward right and this is the usual behavior of a material under compression

1780
02:02:02.818 --> 02:02:07.121
On the bottom row, this is after training, you push down and the lattice actually moves inward.

1781
02:02:07.701 --> 02:02:10.363
The lattice contracts in both directions.

1782
02:02:10.403 --> 02:02:13.204
This is called an auxetic response.

1783
02:02:13.385 --> 02:02:18.708
So it's actually a little bit different than what I just stated, which is like it retains its original construction or connections.

1784
02:02:19.568 --> 02:02:28.576
under different compressed or pressure environments versus becoming victim to the external force that's being applied to it.

1785
02:02:28.596 --> 02:02:30.598
That then changes its construction.

1786
02:02:30.618 --> 02:02:31.178
Yeah, yeah, yeah.

1787
02:02:31.298 --> 02:02:38.164
It has a memory and it can like, now that you've compressed it already, it can now do the thing that you want it to do, right?

1788
02:02:38.224 --> 02:02:39.485
It has like a memory.

1789
02:02:39.965 --> 02:02:46.551
And for an engineering audience, like the implication is very appealing because these are materials whose response can be customized

1790
02:02:47.489 --> 02:02:48.350
after fabrication.

1791
02:02:48.370 --> 02:02:50.432
So you just get it from the factory.

1792
02:02:50.892 --> 02:02:55.156
You can enact stresses to this material and you can customize it.

1793
02:02:55.176 --> 02:03:03.304
And perhaps you can adapt it to structures or soft robotic components like the robotic component that we just saw.

1794
02:03:04.145 --> 02:03:06.006
And it demonstrates a lab route

1795
02:03:06.835 --> 02:03:09.599
to retrainable mechanical function, right?

1796
02:03:09.979 --> 02:03:13.464
Because you can take the same polymer now and reuse it.

1797
02:03:13.764 --> 02:03:18.490
You can heat it up, and it'll come back to its original shape, and then you can do something else with it.

1798
02:03:18.991 --> 02:03:23.016
You can put different stresses to it that'll make its response different.

1799
02:03:24.108 --> 02:03:24.928
In some sense.

1800
02:03:24.988 --> 02:03:25.809
Right?

1801
02:03:26.609 --> 02:03:38.533
I'm actually, the through line from the fundamental question we started with on the story and then the applications, that delta for me is the largest.

1802
02:03:39.735 --> 02:03:47.979
Like, I never would have thought that you could get to, because these are all very practical, like, units of something.

1803
02:03:48.079 --> 02:03:48.339
Yeah.

1804
02:03:48.919 --> 02:03:59.784
It might not be, unlike the other stories where the end product is so discreet or well-defined, this is like a platform for any number of ideas.

1805
02:03:59.804 --> 02:04:00.525
Soft matter.

1806
02:04:00.585 --> 02:04:01.785
Soft matter physics, right?

1807
02:04:01.805 --> 02:04:01.985
Yeah.

1808
02:04:02.065 --> 02:04:05.587
And at the time that he started, soft matter physics was definitely looked down upon.

1809
02:04:06.327 --> 02:04:09.769
Now it's very much a field in its own right.

1810
02:04:10.429 --> 02:04:13.511
And he was a big pioneer in that field.

1811
02:04:13.571 --> 02:04:15.752
He helped establish the legitimacy of it.

1812
02:04:17.102 --> 02:04:27.709
You can tell just from that last one, right, with the retrainable lattice, like a disordered network now becomes a way to specify a mechanical response rather than an obstacle to precision.

1813
02:04:27.729 --> 02:04:28.230
Yeah.

1814
02:04:28.270 --> 02:04:28.470
Right?

1815
02:04:28.530 --> 02:04:31.112
Like the deformable network is the point.

1816
02:04:31.352 --> 02:04:31.832
Exactly.

1817
02:04:32.432 --> 02:04:32.913
Exactly.

1818
02:04:33.113 --> 02:04:33.333
Right?

1819
02:04:33.513 --> 02:04:34.234
Exactly.

1820
02:04:34.874 --> 02:04:38.016
And so these practical outcomes, they differ in their maturity.

1821
02:04:38.096 --> 02:04:40.298
Obviously, the coffee ring mechanism...

1822
02:04:41.105 --> 02:04:49.372
it started out as just a super curiosity driven, but now any aerosol paint, any, any liquid deposition, it's going to matter.

1823
02:04:49.932 --> 02:04:50.152
Right.

1824
02:04:50.352 --> 02:04:52.894
Um, it's, it's pretty incredible.

1825
02:04:52.995 --> 02:04:59.540
So, so this guy, I'm very much looking forward to, um, these are going to my interview with them.

1826
02:04:59.640 --> 02:05:09.828
These, again, we have prerecorded these under embargo, um, this episode under embargo, as we prepare to make our way to Washington, DC, uh,

1827
02:05:10.228 --> 02:05:30.337
big thanks to triple as for um allowing us to have some time with all three of the either individuals or teams that came forward with these stories i just my brain is i know you're gonna have a field day yeah i'm so excited i know you're gonna have a field day um this was part one

1828
02:05:31.120 --> 02:05:35.482
of our coverage, our special coverage of the 2026 Golden Goose Awards.

1829
02:05:36.023 --> 02:05:47.329
Hopefully for some of you who may not have heard about the Golden Goose Awards before, you have an appreciation for what they're trying to elevate in terms of some of our best and brightest.

1830
02:05:47.949 --> 02:05:52.452
Federally funded, basic research, stuff that is good.

1831
02:05:53.862 --> 02:05:56.925
has impacted and will continue to impact so many of our lives.

1832
02:05:57.005 --> 02:06:15.978
And it's an incredible stewardship that AAAS has taken over the responsibility of making sure that those who have committed decades, this is decades of work here, and are creating things that really, again, are going to have and will continue to create great impacts in our lives today.

1833
02:06:16.118 --> 02:06:23.444
We are going to report to you live from Washington, D.C. in our follow-up part two for this episode.

1834
02:06:23.504 --> 02:06:32.571
But don't forget, we are also a week away from one of the greatest other events of the year.

1835
02:06:33.151 --> 02:06:39.256
Another opportunity to highlight many of the best and brightest among us, the Nobel Prizes.

1836
02:06:39.296 --> 02:06:44.560
And so this is going to be an action-packed October as we have...

1837
02:06:45.100 --> 02:06:48.723
back-to-back-to-back-to-back banger episodes.

1838
02:06:49.244 --> 02:06:52.427
If you've not already smashed that subscribe button, please do.

1839
02:06:52.447 --> 02:07:03.076
My name is Lester Nare, joined as always by my co-host and our resident PhD, continuing to blow the minds of myself and the rest of the listeners in FFP Nation.

1840
02:07:03.537 --> 02:07:06.179
We will see you all next week.
