Could We Replace Data Centers with … Plant DNA? | Cliff Kapono and Keolu Fox | TED
8,608 views ・ 2024-10-29
请双击下面的英文字幕来播放视频。
翻译人员: Fanglin ren
校对人员: Bruce Wang
00:04
Cliff Kapono: So people
are pretty surprised
0
4417
2128
Cliff Kapono:所以当我告诉人们
00:06
when I tell them I know the world's first
native Hawaiian genome scientist.
1
6586
3837
我认识世界上第一位夏威夷
原住民基因组科学家时,他们会感到非常惊讶。
00:10
Not because of all the texts he sends me
about genome editing and DNA research,
2
10465
6632
不是因为他给我发了所有
关于基因组编辑和DNA研究的短信,
00:17
but mostly because
they're dominated by cat memes.
3
17138
3170
而是主要是因为
它们以猫咪模因为主。
00:20
Keolu Fox: Oh, brother, you know
there's plenty of surfing in there, too.
4
20350
4713
Keolu Fox:哦,兄弟,你知道
里面也有很多冲浪的地方。
00:25
CK: That's true.
Those take up their own folder.
5
25063
2294
CK:确实如此。
它们占用自己的文件夹。
00:29
KF: There's something
you should know about Cliff and I.
6
29025
2670
KF:关于克里夫和我,
你应该知道一些事情
00:32
We've known each other forever,
7
32362
1543
我们已经认识了很长时间,
00:33
and we spend way too much time
staring at our phones,
8
33905
2670
而且我们花了太多时间盯着手机、
00:36
looking at surf videos.
9
36575
1668
看冲浪视频。
00:38
CK: He's one of those screen time people.
10
38243
2461
CK:他是那些对手机很有节制的人之一。
00:40
You know, FaceTime only,
11
40704
1459
你知道,只有FaceTime,
除非他和我们面对面交流,
00:42
won't answer the phone
unless he's giving us face-to-face.
12
42163
3087
否则他不会接电话。
00:45
And if he is texting me,
he's asking me to send him updates
13
45292
2794
如果他在给我发短信,
他就是要我给他发最新消息,
00:48
on what the waves look like
or how the waves were.
14
48128
2377
告诉他海浪是什么样子
或者海浪怎么样了。
00:51
KF: That's true. I love FaceTime,
but that's because Cliff here is unique.
15
51214
5881
KF:确实如此。我喜欢 FaceTime,
但那是因为这里的克里夫是独一无二的。
00:57
As the only professional
surfer to have a PhD
16
57137
2752
作为唯一拥有分析化学博士学位
00:59
in analytical chemistry, no less,
17
59931
2461
的专业冲浪运动员,
01:02
he gets to spend his time
exploring and surfing
18
62434
2794
他可以花时间探索和冲浪地球上
01:05
some of the most incredible
waves on the planet.
19
65270
2252
一些最不可思议的海浪。
01:07
And now that we have the capacity
20
67522
1668
既然我们有能力
01:09
to share that information
from anywhere in the world,
21
69190
2711
从世界任何地方共享这些信息,
01:11
why wouldn't we?
22
71901
1252
为什么不呢?
01:13
CK: Probably because
it's using up all my monthly data,
23
73778
2628
CK:可能是因为它耗尽了
我所有的月度数据,
01:16
sending him images
of what the waves look like
24
76406
2252
给他发送了
01:18
whenever I'm on a surf trip.
25
78658
1752
每当我去冲浪旅行时
海浪是什么样子的照片。
01:20
KF: That brings up a good point.
26
80410
1585
KF:这提出了一个很好的观点。
01:21
We do share and interact
with terabytes of data almost every day,
27
81995
3879
我们几乎每天都会
与数兆字节的数据共享和交互,
01:25
some more than others.
28
85915
1836
有些比其他人还要多。
01:27
CK: Through texts, emails,
and now this thing --
29
87792
2253
CK:通过短信、电子邮件
以及现在的这件事——
01:30
I don't know if you guys heard about AI --
30
90086
2002
我不知道你们有没有听说过人工智能——
01:32
we're entering a new phase
of data production, consumption,
31
92130
3879
我们正在进入数据生产、
01:36
and probably one of the most challenging
issues of our generation -- data storage.
32
96051
4546
消费的新阶段 ,也可能是我们这一代
最具挑战性的问题之一 ——数据存储。
01:40
KF: But we think we have a solution.
33
100639
1793
KF:但是我们认为我们有解决方案。
01:42
And the answer is an Indigenous one.
34
102474
2752
是个固有的答案。
01:45
We believe we can package text, images
and even surf videos
35
105268
4171
我们相信我们可以将文本、
图像甚至冲浪视频
01:49
in the genomes of living organisms
36
109439
2211
打包到活生物体的基因组中,
01:51
and access them at a later time.
37
111650
2335
以后再进行访问。
01:53
CK: Because, as Keolu likes to remind me,
what's a genome if not a giant hard drive?
38
113985
4338
CK:因为,正如 Keolu喜欢提醒我的那样,
如果不是巨型硬盘,基因组又会是什么?
01:58
Our difference is we want to put
this information into local plants,
39
118323
3962
我们的不同之处在于,我们希望将
这些信息输入当地工厂,
02:02
sequestering atmospheric CO2,
40
122285
2169
封存大气中的二氧化碳,
02:04
and also reducing the energy required
in a rapidly changing climate.
41
124454
4588
同时减少快速变化的气候所需的能量。
02:09
KF: You see, the climate crisis
has a new problem.
42
129084
4045
KF:你看,气候危机
有了一个新问题。
02:13
Data centers are rapidly contributing
to this climate crisis.
43
133171
3879
数据中心正在迅速助长
这场气候危机。
02:17
And data forecasters estimate
44
137092
2586
数据预测者估计,
02:19
that by next year,
there will be over 50 billion
45
139719
2503
到明年,将有超过500亿台
02:22
IoT devices connected to one another,
46
142263
2503
物联网设备相互连接,
02:24
generating over 80 billion
terabytes of information.
47
144808
4546
产生超过800亿太字节的信息。
02:29
And to put things into context,
48
149396
1543
更具体地说,
02:30
that new iPhone in your pocket,
49
150939
1793
你口袋里的那部新苹果手机
02:32
that can hold a total
of around one terabyte of information.
50
152732
4088
总共可以容纳大约一太字节的信息。
02:36
CK: But we're not just storing information
on our phones anymore.
51
156820
3461
CK:但是我们不再只是在手机
上存储信息了。
02:40
We're uploading them
onto that cloud, wherever it is.
52
160281
4380
我们正在将它们上传
到云端,无论它在哪里。
02:44
And heat is a byproduct of cloud storage.
53
164661
2669
而热量是云存储的副产品。
02:47
It's estimated between three to seven
kilowatt-hours of energy are needed
54
167372
4254
据估计,每传输和存储一 千兆字节的信息
需要
02:51
for every gigabyte of information
transferred and stored.
55
171668
4254
三到七千瓦时的能量。
02:55
To put that into context,
watching about two minutes of 4K video
56
175964
5756
综上所述,
观看大约两分钟
03:01
is the same amount of energy
57
181761
1377
的 4K视频相当于
03:03
that's required to power
your phone for one year.
58
183179
3629
为手机供电一年所需的能量。
03:06
And if we are continuing
to back up this information,
59
186850
4504
而且,如果我们
继续像鼓励我们
03:11
as we are encouraged to do,
60
191354
2086
那样备份这些信息,
03:13
the heat expenditure
could increase exponentially,
61
193440
3753
那么未来热量消耗
03:17
moving into the future.
62
197193
1669
可能会呈指数级增长。
03:18
KF: Now data centers combat
these thermodynamic challenges
63
198862
3587
KF:现在,数据中心
依靠冷却机制(比如空调)
03:22
by relying on cooling mechanisms,
kind of like air conditioners,
64
202449
3670
来应对这些热力学挑战,
03:26
but they take up around 40 percent
of their energy demands.
65
206119
3920
但它们占据了大约40%的能源需求。
03:30
And because of this,
66
210081
1710
正因为如此,
03:31
their energy footprint is greater
than the airline industry's as a whole.
67
211833
6089
它们的能源足迹
比整个航空业的能源足迹还要大。
03:37
CK: Imagine how much energy was required
by the entire US alone.
68
217964
4004
CK:想象一下,
光是整个美国就需要多少能源。
03:42
KF: Funny you should ask, Clifford.
69
222010
1710
KF:有趣的是,你应该问一下,克利福德。
03:45
The data industry requires
approximately 200 terawatt-hours annually
70
225388
5589
数据行业每年需要大约
200 太瓦时,再加上手机、
03:51
and with the combination of smart devices,
71
231019
3003
手表和个人计算机
03:54
like phones, watches
and personal computers,
72
234022
3587
等智能设备,
03:57
it accounts for two percent
of global CO2 emissions.
73
237609
3587
它占全球二氧化碳排放量的百分之二。
04:01
But here's the kicker.
74
241196
1251
但关键在此,
04:03
With the introduction of GPUs
75
243531
2253
随着 GPU
04:07
and large language models,
76
247160
3253
和大型语言模型的引入
04:10
and deep learning,
77
250413
1960
以及深度学习,
04:12
heat emissions from data centers
78
252415
2503
数据中心的热排放
04:14
are poised to surpass
the fossil fuel industry
79
254959
2753
有望超过化石燃料行业,
04:17
as the number one contributor
to the climate crisis in our lifetimes.
80
257754
4588
成为我们一生中造成气候危机的头号因素。
04:22
But there's a better way. DNA.
81
262383
3212
但是有更好的方法。DNA。
04:25
After all, life has been storing
information in the organisms' genomes
82
265637
5589
毕竟,生命在生物体的基因组中存储
04:31
for billions of years,
83
271267
1794
信息已有数十亿年了,存储文本、图像
04:33
and storing text, images
and videos is nothing new.
84
273102
3838
和视频并不是什么新鲜事。
04:39
CK: Working off the research
by Daniel Gibson,
85
279776
3003
CK:根据丹尼尔·吉布森的研究,
由塞思·希普曼和乔·戴维斯
04:42
a monumental study in 2017,
86
282779
2586
领导的2017年的一项
04:45
led by Seth Shipman and Joe Davis,
87
285365
1835
具有里程碑意义的研究向全世界表明,
04:47
showed the world that we can put a video
into the genome of a bacteria.
88
287200
4588
我们可以将视频放入细菌的基因组中。
04:51
They showed us that every image pixel
can be represented
89
291788
3128
他们向我们展示了每个图像像素都可以
04:54
by a unique RGB color code.
90
294958
3211
由唯一的RGB颜色代码表示。
04:58
KF: And that color code can be ciphered
into the genetic code,
91
298211
3253
KF:而且这种颜色代码可以加密
05:01
and those sequences can be inserted
92
301506
2252
到遗传密码中,
05:03
into the noncoding portion
of an organism's genome,
93
303800
3003
使用你们中许多人熟悉的基因组编辑技术,
可以将这些序列
05:06
using genome editing techniques
that many of you are familiar with.
94
306845
3962
插入生物体基因组的非编码部分。
05:10
CK: And all we need to do
to access this information at a later time
95
310849
3420
CK:要在以后访问
这些信息,我们所需要做的
05:14
is to sequence the genes
that hold that information
96
314310
3462
就是对保存
这些信息的基因进行测序,
05:17
and cipher it back into the pixeled image.
97
317772
4129
然后将其重新加密回像素化图像。
05:21
And we see the potential of this research,
98
321901
2086
我们看到了这项研究的潜力,
05:23
and we want to expand beyond just
single-cellular organisms like bacteria,
99
323987
4796
我们希望将范围扩展到细菌等
单细胞生物之外,
05:28
and move into multicellular organisms,
100
328783
2461
转向多细胞生物,像甘蔗等
05:31
multicellular organisms like sugarcane.
101
331244
3670
多细胞生物。
05:34
KF: What once was a symbol
of colonialism across our islands,
102
334956
3211
KF:曾经是我们岛屿上殖民主义的象征,
05:38
can now be viewed as a symbol
for a more sustainable future.
103
338209
3504
现在可以看作是更可持续未来的象征。
05:41
CK: If you're not catching
the colonial reference there,
104
341754
3295
CK:如果你没有看到那里的殖民地文字,
05:45
let us remind you of some of the history.
105
345091
2127
那就让我们提醒你一些历史吧。
05:47
In 1893, a group of American
sugarcane plantation owners
106
347260
3837
1893年,一群美国甘蔗种植园主
05:51
illegally overthrew the Hawaiian Kingdom,
107
351139
3170
非法推翻了夏威夷王国,部分原因是
05:54
in part to avoid tax import
on sugar production,
108
354350
3587
为了逃避糖生产的进口税,
在我们的岛屿上
05:57
leaving behind 131 years
of illegal occupation
109
357937
3837
留下了131年的非法占领
06:01
and systemic oppression
across our islands.
110
361774
2711
和系统性压迫。
06:05
KF: And for many of us Hawaiians,
111
365111
1668
KF:对于我们许多夏威夷人来说,
06:06
sugar has left a bitter taste
in our mouths.
112
366779
3212
糖,在口中留下了苦味。
06:09
CK: But we want to reshape this narrative
and take back agency
113
369991
3545
CK:但,我们想重塑这种说法,
06:13
by putting data into the very thing
that nearly wiped out our people.
114
373536
4088
通过将数据投入到差点消灭我们员工
的事物中来收回代理权。
06:17
Additionally, we see sugar as something
115
377665
2878
此外,
06:20
that is beyond just a way
to decolonize the data industry,
116
380585
3962
我们认为糖不仅是实现
数据行业非殖民化的一种方式,
06:24
but also a way
to show something practical.
117
384589
3337
还是一种展示实用价值的方式。
06:27
Its genome is enormous,
and it has multiple genes within it.
118
387967
4213
它的基因组非常庞大,
里面有多个基因。
06:33
KF: Which theoretically
is great for data storage,
119
393473
3378
KF:从理论上讲,
这对于数据存储、
06:36
duplication and integrity.
120
396893
2044
复制和完整性非常有用。
06:38
Imagine storing over 200,000 terabytes
of information in a single gram of DNA.
121
398978
5339
想象一下,在一克DNA中存储
超过20万太字节的信息。
06:45
CK: And you would think
with this sort of technology,
122
405443
2544
CK:你可能会认为有了这种技术,
06:47
we'd be out there planting
acres and acres of these sugar drives,
123
407987
4046
我们就会在那里种植数英亩的糖厂,
06:52
like those greedy fuckers of the past.
124
412033
3337
就像过去那些贪婪的混蛋一样。
06:55
(Laughter)
125
415370
1918
(笑声)
06:57
But we think we have something different.
We want to do it differently.
126
417288
3379
但是我们认为我们有一些不同的东西。
我们想以不同的方式做这件事。
07:02
KF: We believe that sugar drives
are just a small part of the solution.
127
422251
4088
KF:我们认为糖驱动器
只是解决方案的一小部分。
07:06
We’re also exploring other Indigenous
varieties of breadfruit, coconut
128
426381
3878
我们还在探索其他本土
品种的面包果、椰子
07:10
and other microbiota.
129
430301
1543
和其他微生物群。
07:11
We have an ambitious dream
of revitalizing Indigenous land systems
130
431886
6840
我们有一个雄心勃勃的梦想,
07:18
that double as biological data centers.
131
438768
4213
那就是振兴兼作生物数据中心的
土著土地系统。
07:24
CK: And we hope everyone remembers
132
444065
2294
CK:我们希望每个人都记住,
07:26
that indigeneity is just not
about being native peoples.
133
446359
3378
本土化与成为原住民无关。
07:29
It's about understanding the unique roles
134
449737
2044
这是关于了解多个物种
07:31
that multiple species have
within our ecosystem.
135
451781
3462
在我们生态系统中的独特作用。
07:35
KF: All of us come from unique places
136
455243
2252
KF:在座的所有人都来自独特的地方,
07:37
with a tremendous amount
of native species.
137
457495
3420
有大量的本地物种。
07:40
Collectively, all of those species
help those systems to function better.
138
460957
3962
总的来说,所有这些物种都
有助于这些系统更好地运作。
07:44
CK: All of us Indigenous beings
have to reckon with the fact
139
464961
2878
CK:我们所有的原住民都
必须考虑这样一个事实,
07:47
that we all must return to the earth,
140
467880
2169
那就是我们都必须回到地球,
07:50
and that MEGA Land is a place
where we hope our data can go.
141
470091
3128
而那个巨型大地是我们希望
我们的数据能够存放的地方。
07:54
KF: Resting on the banks
of our town's largest living estuary,
142
474303
3671
KF:这片占地两英亩的土地
07:58
this two-acre parcel of land
143
478016
2002
坐落在我们镇最大的活河口岸边,
08:00
is a beacon of hope
for Indigenous people around the world.
144
480059
3462
是世界各地土著人民的希望灯塔。
08:03
CK: What was destined
to be another hotel, condominium or marina
145
483521
3128
CK:原本注定是另一家酒店、
公寓或码头的地方,
08:06
is now becoming a thriving
Indigenous land use center.
146
486649
4797
现在正在变成一个蓬勃发展的
土著土地利用中心。
08:11
KF: Here, there exists
a tremendous amount of information
147
491446
3003
KF:这里有大量的信息
08:14
that can be digitized and stored locally,
148
494449
3128
可以数字化并存储在当地,
08:17
just like our community's
oral histories, hula and chants.
149
497577
4212
就像我们社区的
口述历史、呼啦舞和圣歌一样。
08:21
CK: Biodiversity metrics can be taken
alongside environmental observations
150
501831
4630
CK:依赖该区域的人们可以将生物多样性指标
与环境观测
08:26
by the people who depend on this area
151
506502
2253
一起采用,
08:28
and put into organisms that potentially
hold the key to environmental resilience.
152
508796
5839
并投入到可能
成为环境适应能力关键的生物体中。
08:34
KF: Data no longer need
be taken away from these places
153
514677
3587
KF:数据不再需要从这些地方带走
08:38
and stored in hot, dark,
noisy rooms miles away.
154
518306
3462
并存储在数英里之外炎热、黑暗、
嘈杂的房间里。
08:41
CK: It can remain connected to the land,
155
521809
2378
CK:它可以保持与土地的联系,
08:44
like the stories of old
and the Indigenous peoples of today.
156
524228
3420
就像古老的故事
和当今土著人民的故事一样。
08:49
KF: We're talking
about doing something new
157
529233
3254
KF:我们正在谈论做一些我们
08:52
that we know to be very, very old.
158
532487
2627
知道已经非常非常古老的新事物。
08:55
CK: And this might not solve
the global data storage crisis,
159
535948
5798
CK:这可能无法解决
全球数据存储危机,
09:01
but as we move into a future
where there's digital species
160
541746
5005
但是随着我们进入一个有数字物种
09:06
and AI influencers,
161
546793
3295
和人工智能创作者的未来,
09:10
we hope that some of the solutions
that we so desperately seek
162
550129
3128
我们希望我们迫切寻求的一些解决方案
09:13
aren't only found in a deep algorithm
in some supercomputer,
163
553299
4171
不仅可以在某些超级计算机的深度算法中找到,
09:17
but they might be in a community
that is standing in front of you today.
164
557512
3962
而且可能存在于当今站在你面前的社区中。
09:22
(Applause)
165
562850
2378
(掌声)
09:30
Oh, you guys are getting crazy. OK.
166
570149
1961
哦,你们疯了。好吧。
09:32
(Applause)
167
572110
1876
(掌声)
09:37
But probably the biggest hope,
for me at least,
168
577698
3003
但是,至少
对我来说,整个实验最大的希望可能
09:40
of this whole experiment --
169
580701
1877
是——
09:42
(Laughter)
170
582578
1252
(笑声)
09:46
Is to one day be able
to respond to Keolu's annoying texts.
171
586457
5714
有一天能够回复Keolu的烦人短信。
09:53
KF: How were the waves?
172
593214
1752
KF:海浪怎么样?
09:55
CK: Take a hike at the MEGA Land,
173
595007
2419
CK:在 MEGA Land 徒步旅行,
对一些 DNA 进行测试,
09:57
sequence some DNA and see it for yourself.
174
597468
2920
然后亲眼看看。
10:00
Both: Thank you.
175
600429
1293
两个:谢谢。
10:01
(Applause)
176
601764
2211
(掌声)
10:04
(Laughter)
177
604016
2461
(笑声)
New videos
Original video on YouTube.com
关于本网站
这个网站将向你介绍对学习英语有用的YouTube视频。你将看到来自世界各地的一流教师教授的英语课程。双击每个视频页面上显示的英文字幕,即可从那里播放视频。字幕会随着视频的播放而同步滚动。如果你有任何意见或要求,请使用此联系表与我们联系。