The AI-Powered Tools Supercharging Your Imagination | Bilawal Sidhu | TED

69,179 views ใƒป 2023-09-01

TED


์•„๋ž˜ ์˜๋ฌธ์ž๋ง‰์„ ๋”๋ธ”ํด๋ฆญํ•˜์‹œ๋ฉด ์˜์ƒ์ด ์žฌ์ƒ๋ฉ๋‹ˆ๋‹ค.

๋ฒˆ์—ญ: Noah Shin ๊ฒ€ํ† : Hannah Choi
00:04
All right. Good afternoon, yโ€™all.
0
4376
1835
์•ˆ๋…•ํ•˜์„ธ์š”, ์—ฌ๋Ÿฌ๋ถ„.
00:06
Let's talk about blending reality and imagination.
1
6253
2711
ํ˜„์‹ค๊ณผ ์ƒ์ƒ์˜ ์œตํ•ฉ์— ๋Œ€ํ•ด ์ด์•ผ๊ธฐํ•ด ๋ด…์‹œ๋‹ค.
00:10
But first, let's take a step back in time to 2001.
2
10006
4505
์ด์•ผ๊ธฐํ•˜๊ธฐ์— ์•ž์„œ 2001๋…„์œผ๋กœ ๋Œ์•„๊ฐ€ ๋ณด์ฃ .
00:14
As an 11-year-old in India,
3
14553
1459
11์‚ด์ด์—ˆ์„ ๋•Œ ์ €๋Š” ์ธ๋„์—์„œ
00:16
I became obsessed with computer graphics and visual effects.
4
16054
4213
์ปดํ“จํ„ฐ ๊ทธ๋ž˜ํ”ฝ๊ณผ ์‹œ๊ฐ ํšจ๊ณผ์— ์™„์ „ํžˆ ๋น ์ ธ์žˆ์—ˆ์Šต๋‹ˆ๋‹ค.
00:20
Of course, at that age,
5
20308
1293
๋ฌผ๋ก  ๊ทธ ๋‚˜์ด์—๋Š”
00:21
it meant making cheesy videos kind of like this.
6
21601
2544
์ด๋Ÿฐ ์‹์˜ ์‹ธ๊ตฌ๋ ค ๋น„๋””์˜ค๋ฅผ ๋งŒ๋“ค์—ˆ๋‹ค๋Š” ๊ฒƒ์„ ์˜๋ฏธํ–ˆ์ฃ .
00:24
But therein started a foundational theme in my life,
7
24896
3170
๊ทธ๋Ÿฌ๋‚˜ ๊ทธ๋•Œ๊ฐ€ ์ œ ์‚ถ์˜ ํ† ๋Œ€๊ฐ€ ๋˜๋Š” ์ฃผ์ œ์˜ ์‹œ์ž‘์ด์—ˆ์Šต๋‹ˆ๋‹ค.
00:28
the quest to blend reality and imagination.
8
28066
3629
ํ˜„์‹ค๊ณผ ์ƒ์ƒ์˜ ์œตํ•ฉ์„ ํƒ๊ตฌํ•˜๋Š” ๊ฒƒ ๋ง์ด์ฃ .
00:31
And that quest has stayed with me
9
31695
1835
๊ทธ๋ฆฌ๊ณ  ๊ทธ ํƒ๊ตฌ๋Š”
00:33
and permeated across my decade-long career in tech,
10
33530
3045
๊ตฌ๊ธ€๊ณผ ๊ฐ™์€ ๊ธฐ์—…์˜ ์ œํ’ˆ ๊ด€๋ฆฌ์ž๋กœ์„œ ์ผํ•  ๋•Œ๋‚˜
00:36
working as a product manager at companies like Google
11
36575
3253
์œ ํŠœ๋ธŒ๋‚˜ ํ‹ฑํ†ก๊ณผ ๊ฐ™์€ ํ”Œ๋žซํผ์—์„œ ์ฝ˜ํ…์ธ  ํฌ๋ฆฌ์—์ดํ„ฐ๋กœ ํ™œ๋™ํ•  ๋•Œ ๊ฐ™์ด
00:39
and as a content creator on platforms like YouTube and TikTok.
12
39869
4130
์ œ 10๋…„์— ๊ฑธ์นœ ๊ธฐ์ˆ  ๋ถ„์•ผ ๊ฒฝ๋ ฅ๊ณผ ๊ณ„์†ํ•ด์„œ ํ•จ๊ป˜ํ–ˆ์Šต๋‹ˆ๋‹ค.
00:44
So today, let's deconstruct this quest to blend reality and imagination
13
44833
4796
๊ทธ๋ž˜์„œ ์˜ค๋Š˜์€ ์ด ํ˜„์‹ค๊ณผ ์ƒ์ƒ์˜ ์œตํ•ฉ ํƒ๊ตฌ๋ฅผ ๋ถ„์„ํ•˜๊ณ 
00:49
and explore how itโ€™s getting supercharged -- buzzword alert --
14
49671
3337
์ด๊ฒƒ์ด ์–ด๋–ป๊ฒŒ ๊ธ‰๊ฒฉํ•˜๊ฒŒ ๋ฐœ์ „ํ•˜๊ณ  ์žˆ๋Š”์ง€ ํ•จ๊ป˜ ์•Œ์•„๋ณด๊ฒ ์Šต๋‹ˆ๋‹ค.
00:53
by artificial intelligence.
15
53049
1710
์š”์ฆ˜ ์œ ํ–‰ํ•˜๋Š” ์ธ๊ณต์ง€๋Šฅ์— ์˜ํ•ด์„œ ๋ง์ด์ฃ .
00:55
Let's start with the reality bit.
16
55260
1710
๋จผ์ € ํ˜„์‹ค์— ๋Œ€ํ•ด ์ด์•ผ๊ธฐํ•ด ๋ด…์‹œ๋‹ค.
00:58
You probably heard about photogrammetry.
17
58638
2378
์•„๋งˆ๋„ ์‚ฌ์ง„ ์ธก๋Ÿ‰๋ฒ•์— ๋Œ€ํ•ด ๋“ค์–ด๋ณด์‹  ๋ถ„๋“ค์ด ๊ณ„์‹ค ๊ฒ๋‹ˆ๋‹ค.
01:01
It's the art and science of measuring stuff in the real world
18
61057
3629
์‚ฌ์ง„ ์ธก๋Ÿ‰๋ฒ•์€ ์‚ฌ์ง„๋“ค๊ณผ ์—ฌ๋Ÿฌ ์„ผ์„œ๋“ค์„ ์ด์šฉํ•˜์—ฌ
01:04
using photos and other sensors.
19
64686
2669
ํ˜„์‹ค ์„ธ๊ณ„์˜ ์‚ฌ๋ฌผ์„ ์ธก์ •ํ•˜๋Š” ์˜ˆ์ˆ ์ ์ธ ๊ณผํ•™์ž…๋‹ˆ๋‹ค.
01:07
What required massive data centers and teams of experts in the 2000s
20
67355
5047
2000๋…„๋Œ€์—๋Š” ๊ฑฐ๋Œ€ํ•œ ๋ฐ์ดํ„ฐ ์„ผํ„ฐ์™€ ์ „๋ฌธ๊ฐ€ ํŒ€์ด ํ•„์š”ํ–ˆ๋˜ ๊ฒƒ์ด
01:12
became increasingly democratized by the 2010s.
21
72402
3921
2010๋…„๋Œ€์— ๋“ค์–ด์™€์„œ๋Š” ์ƒ๋‹นํžˆ ๋Œ€์ค‘ํ™”๋˜์—ˆ์ฃ .
01:16
Then, of course, machine learning came along
22
76323
2335
๊ทธ๋ฆฌ๊ณ  ๋‚˜์„  ๊ธฐ๊ณ„ ํ•™์Šต ๊ธฐ์ˆ ์ด ๋“ฑ์žฅํ•ด
01:18
and took things to a whole new level
23
78658
2127
์ด๋ฅผ ์™„์ „ํžˆ ์ƒˆ๋กœ์šด ์ˆ˜์ค€์œผ๋กœ ๋Œ์–ด์˜ฌ๋ ธ์Šต๋‹ˆ๋‹ค.
01:20
with techniques like neural radiance fields, or NeRFs.
24
80827
4254
์‹ ๊ฒฝ๋ง ๋ณต์‚ฌ๋„ ๋ถ„์•ผ(NeRFs)์™€ ๊ฐ™์€ ๊ธฐ์ˆ ์„ ์‚ฌ์šฉํ•ด์„œ ๋ง์ด์ฃ .
01:25
What you're seeing here is an AI model
25
85123
2794
์ง€๊ธˆ ๋ณด๊ณ  ๊ณ„์‹  ์ด๊ฒƒ์€ ์ธ๊ณต์ง€๋Šฅ ๋ชจ๋ธ์ด
01:27
creating a ground-up volumetric 3D representation
26
87959
3462
2D ์ด๋ฏธ์ง€๋งŒ์„ ์‚ฌ์šฉํ•˜์—ฌ ์ฒ˜์Œ๋ถ€ํ„ฐ ๋๊นŒ์ง€
01:31
using 2D images alone.
27
91463
2252
3D ๋ฌ˜์‚ฌ๋ฅผ ์ƒ์„ฑํ•˜๊ณ  ์žˆ๋Š” ์˜์ƒ์ž…๋‹ˆ๋‹ค.
01:33
But unlike older techniques for reality capture,
28
93757
2919
ํ•˜์ง€๋งŒ ์˜ˆ์ „์˜ ๋ฆฌ์–ผ๋ฆฌํ‹ฐ ์บก์ฒ˜ ๊ธฐ์ˆ ๋“ค๊ณผ๋Š” ๋‹ฌ๋ฆฌ,
01:36
NeRFs do a really good job of encapsulating the sheer complexity
29
96718
4880
NeRFs๋Š” ํ˜„์‹ค ์„ธ๊ณ„์˜ ์ˆœ์ˆ˜ํ•œ ๋ณต์žก์„ฑ๊ณผ
01:41
and nuance of reality.
30
101640
1876
๋ฏธ๋ฌ˜ํ•จ์„ ์ž˜ ํฌ์ฐฉํ•ฉ๋‹ˆ๋‹ค.
01:43
The vibe, if you will.
31
103558
1960
์›ํ•˜์‹ ๋‹ค๋ฉด ๋Š๋‚Œ๊ณผ ๋ถ„์œ„๊ธฐ๊นŒ์ง€๋„ ๋ง์ด์ฃ .
01:46
Twelve months later, you can do all of this stuff
32
106311
2544
12๊ฐœ์›” ๋’ค์— ์—ฌ๋Ÿฌ๋ถ„์€ ์ด ๋ชจ๋“  ๊ฒƒ๋“ค์„
01:48
using the iPhone in your pocket,
33
108855
1752
๋ฃจ๋งˆ ๊ฐ™์€ ์•ฑ์„ ์‚ฌ์šฉํ•˜์—ฌ
01:50
using apps like Luma.
34
110607
1877
์ฃผ๋จธ๋‹ˆ ์† ์•„์ดํฐ์œผ๋กœ๋„ ํ•  ์ˆ˜ ์žˆ๊ฒŒ ๋˜์—ˆ์Šต๋‹ˆ๋‹ค.
01:52
It's like 3D screenshots for the real world.
35
112484
2878
๋งˆ์น˜ ํ˜„์‹ค ์„ธ๊ณ„๋ฅผ 3D๋กœ ์Šคํฌ๋ฆฐ์ƒท ํ•˜๋Š” ๊ฒƒ๊ณผ ๊ฐ™์ฃ .
01:55
Capture anything once and reframe it infinitely in postproduction,
36
115362
4504
ํ•œ ๋ฒˆ ์บก์ณํ•˜๋ฉด ํ›„๋ฐ˜ ์ž‘์—…์—์„œ ๋ฌดํ•œ์œผ๋กœ ์žฌ๊ตฌ์„ฑ ํ•  ์ˆ˜ ์žˆ์–ด
01:59
so you can start building that collection of spaces, places and objects
37
119866
4296
์—ฌ๋Ÿฌ๋ถ„์ด ์ง„์‹ฌ์œผ๋กœ ์ข‹์•„ํ•˜์‹œ๋Š”
02:04
that you truly care about
38
124204
1585
๊ณต๊ฐ„, ์žฅ์†Œ ๋ฐ ์‚ฌ๋ฌผ์˜ ์ปฌ๋ ‰์…˜์„ ๋งŒ๋“ค ์ˆ˜ ์žˆ๊ณ 
02:05
and conjure them up in your future creations.
39
125830
2586
๋ฏธ๋ž˜์— ์ฐฝ์ž‘๋ฌผ์„ ๋งŒ๋“ค ๋•Œ ๊ทธ๊ฒƒ๋“ค์„ ์†์‰ฝ๊ฒŒ ๋ถˆ๋Ÿฌ์˜ฌ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
02:09
So that's the reality bit.
40
129751
1752
๋ฐ”๋กœ ์ด๊ฒƒ์ด ํ˜„์‹ค ๋ถ€๋ถ„์ž…๋‹ˆ๋‹ค.
02:11
As NeRFs were popping off last year,
41
131544
2378
์ง€๋‚œํ•ด NeRFs๊ฐ€ ์ฃผ๋ชฉ๋ฐ›์œผ๋ฉฐ
02:13
the AI summer was also in full effect,
42
133963
2878
์ธ๊ณต์ง€๋Šฅ ์—ดํ’์ด ๋ณธ๊ฒฉ์ ์œผ๋กœ ์‹œ์ž‘๋˜์—ˆ์Šต๋‹ˆ๋‹ค.
02:16
with Midjourney, DALL-E 2, Stable Diffusion
43
136883
3462
๋ฏธ๋“œ์ €๋‹ˆ, ๋‹ฌ๋ฆฌ 2, ์Šคํ…Œ์ด๋ธ” ๋””ํ“จ์ „ ๋“ฑ์ด
02:20
all hitting the market around the same time.
44
140387
2669
๊ฑฐ์˜ ๋™์‹œ์— ์‹œ์žฅ์— ๋“ฑ์žฅํ–ˆ์ฃ .
02:23
But what I fell in love with was inpainting.
45
143098
3253
ํ•˜์ง€๋งŒ ์ œ๊ฐ€ ๋งค๋ฃŒ๋œ ๊ฒƒ์€ ์ธํŽ˜์ธํŒ…์ด์—ˆ์Šต๋‹ˆ๋‹ค.
02:26
This technique allows you to take existing imagery
46
146351
3086
์ด ๊ธฐ์ˆ ์„ ์ด์šฉํ•˜๋ฉด ๊ธฐ์กด ์ด๋ฏธ์ง€์—
02:29
and augment it with whatever you like,
47
149437
2419
์›ํ•˜๋Š” ๊ฒƒ์„ ๋ง๋ถ™์ผ ์ˆ˜ ์žˆ๊ณ 
02:31
and the results are photorealistically fantastic.
48
151856
3337
๊ทธ ๊ฒฐ๊ณผ๋ฌผ์€ ์‚ฌ์‹ค์ ์ด๊ณ ๋„ ํ™˜์ƒ์ ์ž…๋‹ˆ๋‹ค.
02:35
It blew my mind
49
155193
1293
์ด๊ฑด ์ €์—๊ฒŒ ์ •๋ง ๋†€๋ผ์šด ์ผ์ด์—ˆ์–ด์š”.
02:36
because stuff that would have taken me like three hours in classical workflows
50
156486
4213
๊ธฐ์กด ๋ฐฉ์‹์˜ ์ž‘์—…์—์„  3์‹œ๊ฐ„์ด ๊ฑธ๋ฆด ์ผ์„
02:40
I could pull off in just three minutes.
51
160699
2877
3๋ถ„๋งŒ์— ํ•  ์ˆ˜ ์žˆ๊ฒŒ ๋˜์—ˆ๊ธฐ ๋•Œ๋ฌธ์ด์ฃ .
02:44
But I wanted more.
52
164285
1252
ํ•˜์ง€๋งŒ ์ €๋Š” ๋” ๋งŽ์€ ๊ฒƒ์„ ์›ํ–ˆ๊ธฐ์—
02:45
Enter ControlNet,
53
165578
1252
์ปจํŠธ๋กค๋„ท์„ ์‚ฌ์šฉํ•˜๊ธฐ ์‹œ์ž‘ํ–ˆ์Šต๋‹ˆ๋‹ค.
02:46
a game-changing technique by Stanford researchers
54
166871
2962
์ปจํŠธ๋กค๋„ท์€ ์Šคํƒ ํผ๋“œ ์—ฐ๊ตฌ์›๋“ค์ด ๊ฐœ๋ฐœํ•œ ํ˜์‹ ์ ์ธ ๊ธฐ์ˆ ๋กœ
02:49
that allows you to use various input conditions
55
169874
3337
๋‹ค์–‘ํ•œ ์ž…๋ ฅ ์กฐ๊ฑด์„ ์‚ฌ์šฉํ•˜์—ฌ
02:53
to guide and control the AI image generation process.
56
173253
4004
์ธ๊ณต์ง€๋Šฅ ์ด๋ฏธ์ง€ ์ƒ์„ฑ ๊ณผ์ •์„ ์œ ๋„ํ•˜๊ณ  ์ œ์–ดํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
02:57
So in my case, I could take the depth information
57
177298
3212
์ œ ๊ฒฝ์šฐ์—๋Š” 3D ์Šค์บ”์œผ๋กœ๋ถ€ํ„ฐ
03:00
and the texture detail from my 3D scans
58
180552
2961
๊นŠ์ด ์žˆ๋Š” ์ •๋ณด์™€ ๊ฐ์ด‰ ๋””ํ…Œ์ผ์„ ๊ฐ€์ ธ์™€์„œ
03:03
and use it to literally reskin reality.
59
183555
3170
ํ˜„์‹ค์„ ์‹ค์ œ๋กœ ์žฌ๊ตฌ์„ฑํ•  ์ˆ˜ ์žˆ์—ˆ์Šต๋‹ˆ๋‹ค.
03:06
Now, this isn't just cool video.
60
186766
1543
์ž, ์ง€๊ธˆ ๋ณด์‹œ๋Š” ๊ฒƒ์€ ๋‹จ์ˆœํžˆ ๋ฉ‹์ง„ ๋น„๋””์˜ค๊ฐ€ ์•„๋‹ˆ์—์š”.
03:08
Thereโ€™s a lot of useful use cases, too.
61
188309
2294
์‹ค์šฉ์ ์ธ ์ด์šฉ ์‚ฌ๋ก€๋“ค๋„ ๋งŽ์Šต๋‹ˆ๋‹ค.
03:10
For example, in this case I'm taking a 3D scan of my parents' drawing room,
62
190603
5506
์˜ˆ๋ฅผ ๋“ค๋ฉด ์—ฌ๊ธฐ์„œ ์ €๋Š” ๋ถ€๋ชจ๋‹˜์˜ ๋ฏธ์ˆ ์‹ค์„ 3D ์Šค์บ”ํ•˜์—ฌ
03:16
as my mother likes to call it,
63
196109
1585
์ €ํฌ ์–ด๋จธ๋‹ˆ๊ฐ€ ๋ฏธ์ˆ ์‹ค์ด๋ผ๊ณ  ๋ถ€๋ฅด์‹œ๋”๋ผ๊ณ ์š”.
03:17
and reskinning it to different styles of Indian decor
64
197694
4004
๋‹ค์–‘ํ•œ ์ธ๋„ ์žฅ์‹ ์Šคํƒ€์ผ๋กœ ์žฌ๊ตฌ์„ฑํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.
03:21
and doing so while respecting the spatial context and the layout
65
201698
4212
๋‚ด๋ถ€ ๊ณต๊ฐ„์˜ ๊ณต๊ฐ„์  ๋งฅ๋ฝ๊ณผ ๋ฐฐ์น˜๋ฅผ ์‚ด๋ฆฌ๋ฉฐ
03:25
of the interior space.
66
205952
2085
์žฌ๊ตฌ์„ฑํ•˜๊ณ  ์žˆ๋Š” ๊ฒƒ์ด์ฃ .
03:28
If you squint, I'm sure you can see
67
208079
2252
๋ˆˆ์„ ์‚ด์ง ๊ฐ๊ณ  ์ž์„ธํžˆ ์ƒ๊ฐํ•ด ๋ณด์‹œ๋ฉด
03:30
how this is going to transform architecture
68
210373
2336
์ด๋Ÿฌํ•œ ๊ธฐ์ˆ ์ด ์–ด๋–ป๊ฒŒ ๊ฑด์ถ•๊ณผ ์ธํ…Œ๋ฆฌ์–ด ๋””์ž์ธ์„ ๋ณ€ํ™”์‹œํ‚ฌ์ง€
03:32
and interior design forever.
69
212751
2210
์•„์‹ค ์ˆ˜ ์žˆ์„ ๊ฒ๋‹ˆ๋‹ค.
03:35
You could take that 2016 scan of a Buddha statue
70
215503
3212
2016๋…„์— ์ฐ์€ ๋ถ€์ฒ˜์ƒ์˜ ์Šค์บ”์„
03:38
and reskin it to be gloriously golden
71
218757
2377
์ฐฌ๋ž€ํ•œ ํ™ฉ๊ธˆ๋น›์œผ๋กœ ๋ณ€์‹ ์‹œํ‚ฌ ์ˆ˜ ์žˆ๊ณ 
03:41
while pulling off these impossible camera moves
72
221176
2544
๋˜ํ•œ ๋ถˆ๊ฐ€๋Šฅํ•œ ์นด๋ฉ”๋ผ ์›€์ง์ž„์„ ์„ ๋ณด์ด๋ฉฐ
03:43
you just couldn't do any other way.
73
223762
2043
๋‹ค๋ฅธ ๊ธฐ์ˆ ๋กœ๋Š” ํ•  ์ˆ˜ ์—†๋Š” ์ผ์„ ํ•ด๋‚ผ ์ˆ˜ ์žˆ์„ ๊ฒƒ์ž…๋‹ˆ๋‹ค.
03:46
Or you could take that vacation footage from your trip to Tokyo
74
226765
3837
๋˜๋Š” ์—ฌ๋Ÿฌ๋ถ„์€ ๋„์ฟ„ ์—ฌํ–‰ ์ค‘์— ์ดฌ์˜ํ•œ ํœด๊ฐ€ ์˜์ƒ์„ ๊ฐ€์ ธ์™€
03:50
and bring these cherry blossoms to life in a whole new way.
75
230602
3587
์ด ๋ฒš๊ฝƒ์„ ์™„์ „ํžˆ ์ƒˆ๋กœ์šด ๋ฐฉ์‹์œผ๋กœ ์‚ด๋ ค๋‚ผ ์ˆ˜ ์žˆ์„ ๊ฒƒ์ž…๋‹ˆ๋‹ค.
03:54
And let me tell you, cherry blossoms look really good during the day,
76
234189
3295
๊ทธ๋ฆฌ๊ณ  ์ด ๋ง์”€์„ ๋“œ๋ฆด๊ฒŒ์š”. ๋ฒš๊ฝƒ์€ ๋‚ฎ์—๋„ ์ •๋ง ์•„๋ฆ„๋‹ต๊ฒŒ ๋ณด์ด์ง€๋งŒ
03:57
but they look even better at night.
77
237484
2752
๋ฐค์—๋Š” ๋”์šฑ๋” ์•„๋ฆ„๋‹ค์›Œ์ง‘๋‹ˆ๋‹ค.
04:00
Oh, my God. They sure are glowing.
78
240236
2503
์˜ค, ์ •๋ง ๋น›๋‚˜๊ณ  ์žˆ๋„ค์š”.
04:02
It's almost like this dreamlike quality
79
242739
1918
์ธ๊ณต์ง€๋Šฅ์„ ์ด์šฉํ•ด
04:04
where you can use AI to accentuate the best aspects of the real world.
80
244657
4672
ํ˜„์‹ค ์„ธ๊ณ„์˜ ๊ฐ€์žฅ ์ข‹์€ ์ธก๋ฉด์„ ๊ฐ•์กฐํ•  ์ˆ˜ ์žˆ๋Š” ์ •๋ง ๊ฟˆ๊ฐ™์€ ์ผ์ž…๋‹ˆ๋‹ค.
04:10
Natural landscapes look just as beautiful.
81
250205
2377
์ž์—ฐ ํ’๊ฒฝ ๋˜ํ•œ ๋งค์šฐ ์•„๋ฆ„๋‹ค์›Œ์š”.
04:12
Like this waterfall that could be on another planet.
82
252624
3211
๋‹ค๋ฅธ ํ–‰์„ฑ์— ์กด์žฌํ•  ๊ฒƒ ๊ฐ™์€ ์ด ํญํฌ๋ฅผ ๋ณด์„ธ์š”.
04:15
But of course, you could go over the hills and far away
83
255877
2628
๋ฌผ๋ก  ์—ฌ๋Ÿฌ๋ถ„์€ ์–ธ์ œ๋“ ์ง€ ๋‹ค๋ฅธ ์ฐจ์›์—์„œ ์ด ์–ธ๋•์„ ๋„˜์–ด
04:18
to the French Alps from another dimension.
84
258546
2002
ํ”„๋ž‘์Šค ์•Œํ”„์Šค๋กœ ์—ฌํ–‰ํ•˜์‹ค ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
04:21
But it's not just static scenes.
85
261883
1960
ํ•˜์ง€๋งŒ ์ •์ ์ธ ์žฅ๋ฉด์— ํ•œ์ •๋œ ๊ฒƒ์€ ์•„๋‹™๋‹ˆ๋‹ค.
04:23
You can do this stuff with video, too.
86
263885
2878
๋น„๋””์˜ค์—๋„ ์ ์šฉ์‹œํ‚ฌ ์ˆ˜ ์žˆ์–ด์š”.
04:26
I can't wait till this technology is running at 30 frames per second
87
266805
3920
์ €๋Š” ์ด ๊ธฐ์ˆ ์ด ์ดˆ๋‹น 30ํ”„๋ ˆ์ž„์œผ๋กœ ๋™์ž‘ํ•˜๋Š” ๋‚ ์ด ๋นจ๋ฆฌ ์™”์œผ๋ฉด ์ข‹๊ฒ ์Šต๋‹ˆ๋‹ค.
04:30
because it's going to transform augmented reality and 3D rendering.
88
270767
3712
์ด ๊ธฐ์ˆ ์€ ์ฆ๊ฐ• ํ˜„์‹ค๊ณผ 3D ๋ Œ๋”๋ง์„ ๋ณ€ํ™”์‹œํ‚ฌ ๊ฒƒ์ด๊ธฐ ๋•Œ๋ฌธ์ด์ฃ .
04:34
I mean, how soon until we're channel-surfing realities
89
274479
3587
๋‹ค์‹œ ๋งํ•ด, ํ˜„์‹ค์„ ์‹ค์„ธ๊ณ„ ์œ„์— ๊ฒน์ณ๋†“์€ ์ฑ„๋„์˜ ์„œํ•‘์„
04:38
layered on top of the real world?
90
278066
2461
์šฐ๋ฆฌ๋Š” ์–ธ์ œ์ฏค ํ•  ์ˆ˜ ์žˆ๊ฒŒ ๋ ๊นŒ์š”?
04:41
Of course, just like reality capture got democratized,
91
281236
2627
๋ฌผ๋ก  ๋ฆฌ์–ผ๋ฆฌํ‹ฐ ์บก์ณ๊ฐ€ ๋Œ€์ค‘ํ™”๋œ ๊ฒƒ์ฒ˜๋Ÿผ
04:43
all these tools from last year are getting even easier.
92
283863
3462
์ง€๋‚œํ•ด์— ์‚ฌ์šฉ๋œ ์ด ๋ชจ๋“  ๋„๊ตฌ๋“ค์ด ๋”์šฑ ๋‹ค๋ฃจ๊ธฐ ์‰ฌ์›Œ์ง€๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.
04:47
So instead of me spending hours
93
287325
2336
๊ทธ๋ž˜์„œ ๋ช‡ ์‹œ๊ฐ„ ๋™์•ˆ
04:49
weaving together a bunch of different tools,
94
289702
2169
์ˆ˜๋งŽ์€ ๋„๊ตฌ๋“ค์„ ํ•จ๊ป˜ ์‚ฌ์šฉํ•ด์•ผ ํ•˜๋Š” ๋Œ€์‹ 
04:51
tools like Runway and Kaiber let you do exactly the same stuff
95
291913
4171
๋Ÿฐ์›จ์ด์™€ ์นด์ด๋ฒ„ ๊ฐ™์€ ๋„๊ตฌ๋“ค์€ ๋ช‡ ๋ฒˆ์˜ ํด๋ฆญ๋งŒ์œผ๋กœ
04:56
with just a couple clicks.
96
296125
1585
๋˜‘๊ฐ™์€ ์ž‘์—…์„ ํ•  ์ˆ˜ ์žˆ๊ฒŒ ํ•ด์ค๋‹ˆ๋‹ค.
04:57
Want to go from day to night? No problemo.
97
297752
2920
๋‚ฎ์—์„œ ๋ฐค์œผ๋กœ ๋ฐ”๊พธ๊ณ  ์‹ถ์œผ์‹ ๊ฐ€์š”? ๋ฌธ์ œ ์—†์Šต๋‹ˆ๋‹ค.
05:00
Want to get that retro 90s aesthetic from "Full House"?
98
300713
3295
โ€˜ํ’€ ํ•˜์šฐ์Šคโ€™์˜ ๋ ˆํŠธ๋กœ 90๋…„๋Œ€ ๊ฐ์„ฑ์„ ์›ํ•˜์‹œ๋‚˜์š”?
05:04
You can do that too.
99
304050
1460
๊ทธ๊ฒƒ๋„ ๊ฐ€๋Šฅํ•ฉ๋‹ˆ๋‹ค.
05:06
But it goes beyond reality capture.
100
306427
2628
ํ•˜์ง€๋งŒ ์ด๋Š” ๋ฆฌ์–ผ๋ฆฌํ‹ฐ ์บก์ณ๋ฅผ ๋„˜์–ด์„œ๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.
05:09
Companies like Wonder Dynamics
101
309097
1877
์›๋” ๋‹ค์ด๋‚ด๋ฏน์Šค์™€ ๊ฐ™์€ ๊ธฐ์—…๋“ค์€
05:11
are turning video into this immaculate form of performance capture
102
311015
4463
์˜์ƒ์„ ์™„๋ฒฝํ•œ ํ˜•ํƒœ์˜ ํผํฌ๋จผ์Šค ์บก์ณ๋กœ ๋ฐ”๊ฟ”์„œ
05:15
so you can embody fantastical creatures using the phone in your pocket.
103
315478
4963
์ฃผ๋จธ๋‹ˆ์˜ ํœด๋Œ€์ „ํ™”๋กœ ํ™˜์ƒ์ ์ธ ์ƒ๋ฌผ์„ ๊ตฌํ˜„ํ•  ์ˆ˜ ์žˆ๋„๋ก ํ•ด์ค๋‹ˆ๋‹ค.
05:20
This is stuff that James Cameron only dreamt about in the 2000s.
104
320441
3629
์ด๊ฒƒ์€ ์ œ์ž„์Šค ์นด๋ฉ”๋ก ์ด 2000๋…„๋Œ€์— ๊ฟˆ๊พธ๊ธฐ๋งŒ ํ–ˆ๋˜ ๊ฒƒ๋“ค์ด์—์š”.
05:24
And now you could do it with your iPhone? Thatโ€™s absolutely wild to me.
105
324070
3337
์ด์   ์•„์ดํฐ๋งŒ์œผ๋กœ ํ•  ์ˆ˜ ์žˆ๋‹ค๋‹ˆ ์ •๋ง ๋ฏฟ๊ธฐ ์–ด๋ ต๋„ค์š”.
05:28
So when I look back at the past two decades
106
328241
3629
์ง€๋‚œ 20๋…„์„ ๋˜๋Œ์•„๋ณด๋ฉด์„œ
05:31
and this ill-tailored tapestry of tools that I've had to learn,
107
331911
4129
์ œ๊ฐ€ ๋ฐฐ์›Œ์•ผ ํ–ˆ๋˜ ์ด๋Ÿฌํ•œ ๋‚œ์žกํ•˜๊ฒŒ ์–ฝํžŒ ๋„๊ตฌ๋“ค์„ ์ƒ๊ฐํ•ด ๋ณด๋ฉด
05:36
I feel a sense of optimism for what lies ahead
108
336082
2836
๋‹ค์Œ ์„ธ๋Œ€์˜ ํฌ๋ฆฌ์—์ดํ„ฐ๋“ค์—๊ฒŒ ์–ด๋–ค ๊ฐ€๋Šฅ์„ฑ์ด ๊ธฐ๋‹ค๋ฆฌ๊ณ  ์žˆ์„์ง€
05:38
for the next generation of creators.
109
338960
1752
๊ธฐ๋Œ€๊ฐ€ ๋ฉ๋‹ˆ๋‹ค.
05:41
The 11-year-olds of today
110
341504
2002
์˜ค๋Š˜๋‚ ์˜ 11์‚ด ์•„์ด๋“ค์€
05:43
don't have to worry about all of that crap.
111
343548
2294
๊ทธ ๋ชจ๋“  ๊ฒƒ๋“ค์— ๋Œ€ํ•ด ๊ฑฑ์ •ํ•  ํ•„์š”๊ฐ€ ์—†์–ด์š”.
05:45
All they need to do is have a creative vision
112
345884
2961
์–ด๋ฆฐ์ด๋“ค์€ ๊ทธ์ € ์ฐฝ์˜์ ์ธ ๋น„์ „์„ ๊ฐ€์ง€๊ณ 
05:48
and a knack for working in concert with these AI models,
113
348887
3336
์ด๋Ÿฌํ•œ ์ธ๊ณต์ง€๋Šฅ ๋ชจ๋ธ๋“ค๊ณผ ํ˜‘์—…ํ•˜๋Š” ์š”๋ น์„ ๊ฐ–๊ธฐ๋งŒ ํ•˜๋ฉด ๋ฉ๋‹ˆ๋‹ค.
05:52
these AI models that are truly a distillation
114
352265
2836
์ด๋Ÿฌํ•œ ์ธ๊ณต์ง€๋Šฅ ๋ชจ๋ธ๋“ค์€ ์ •๋ง๋กœ ์ธ๊ฐ„์˜ ์ง€์‹๊ณผ ์ฐฝ์˜์„ฑ์„
05:55
of human knowledge and creativity.
115
355101
2586
์ •์ œํ•œ ๊ฒƒ์ด๋ผ๊ณ  ๋ณผ ์ˆ˜ ์žˆ์–ด์š”.
05:57
And that's a future I'm excited about,
116
357687
2002
๊ทธ๊ฒƒ์ด ๋ฐ”๋กœ ์ œ๊ฐ€ ๊ธฐ๋Œ€ํ•˜๋Š” ๋ฏธ๋ž˜์ž…๋‹ˆ๋‹ค.
05:59
a future where you can blend reality and imagination
117
359689
3212
์—ฌ๋Ÿฌ๋ถ„์˜ ๋ฏฟ์Œ์งํ•œ ์ธ๊ณต์ง€๋Šฅ ์กฐ์ˆ˜์™€ ํ•จ๊ป˜
06:02
with your trusty AI copilot.
118
362901
1918
ํ˜„์‹ค๊ณผ ์ƒ์ƒ์„ ์กฐํ™”๋กญ๊ฒŒ ์–ด์šฐ๋Ÿฌ์ง€๊ฒŒ ํ•  ์ˆ˜ ์žˆ๋Š” ๋ฏธ๋ž˜ ๋ง์ด์ฃ .
06:05
Thank you very much.
119
365361
1293
๊ฐ์‚ฌํ•ฉ๋‹ˆ๋‹ค.
06:06
(Applause)
120
366654
2086
(๋ฐ•์ˆ˜)
์ด ์›น์‚ฌ์ดํŠธ ์ •๋ณด

์ด ์‚ฌ์ดํŠธ๋Š” ์˜์–ด ํ•™์Šต์— ์œ ์šฉํ•œ YouTube ๋™์˜์ƒ์„ ์†Œ๊ฐœํ•ฉ๋‹ˆ๋‹ค. ์ „ ์„ธ๊ณ„ ์ตœ๊ณ ์˜ ์„ ์ƒ๋‹˜๋“ค์ด ๊ฐ€๋ฅด์น˜๋Š” ์˜์–ด ์ˆ˜์—…์„ ๋ณด๊ฒŒ ๋  ๊ฒƒ์ž…๋‹ˆ๋‹ค. ๊ฐ ๋™์˜์ƒ ํŽ˜์ด์ง€์— ํ‘œ์‹œ๋˜๋Š” ์˜์–ด ์ž๋ง‰์„ ๋”๋ธ” ํด๋ฆญํ•˜๋ฉด ๊ทธ๊ณณ์—์„œ ๋™์˜์ƒ์ด ์žฌ์ƒ๋ฉ๋‹ˆ๋‹ค. ๋น„๋””์˜ค ์žฌ์ƒ์— ๋งž์ถฐ ์ž๋ง‰์ด ์Šคํฌ๋กค๋ฉ๋‹ˆ๋‹ค. ์˜๊ฒฌ์ด๋‚˜ ์š”์ฒญ์ด ์žˆ๋Š” ๊ฒฝ์šฐ ์ด ๋ฌธ์˜ ์–‘์‹์„ ์‚ฌ์šฉํ•˜์—ฌ ๋ฌธ์˜ํ•˜์‹ญ์‹œ์˜ค.

https://forms.gle/WvT1wiN1qDtmnspy7