Miris Streams Photorealistic 3D at Scale

1

00:00:12,929 --> 00:00:13,596

Welcome back to

2

00:00:13,596 --> 00:00:15,473

GTC live Day two

3

00:00:15,473 --> 00:00:17,600

Lisa Martin here with CoreWeave.

4

00:00:17,600 --> 00:00:20,061

You can see some of the energy behind me.

5

00:00:20,061 --> 00:00:22,272

And I gotta tell you, as I've said before

6

00:00:22,272 --> 00:00:23,106

saying it's electric.

7

00:00:23,106 --> 00:00:24,024

I gotta find a new word

8

00:00:24,024 --> 00:00:25,734

because this is next level.

9

00:00:25,734 --> 00:00:27,652

This is the Super Bowl for AI.

10

00:00:27,652 --> 00:00:28,737

People have been describing it

11

00:00:28,737 --> 00:00:31,406

as a carnival, as the heartbeat of AI.

12

00:00:31,406 --> 00:00:32,907

It's all true.

13

00:00:32,907 --> 00:00:34,159

Super excited to be joined

14

00:00:34,159 --> 00:00:37,620

by Miris, Will McDonald the Chief Product Officer.

15

00:00:37,620 --> 00:00:38,455

Well, thank you so much

16

00:00:38,455 --> 00:00:39,330

for spending some time with us.

17

00:00:39,330 --> 00:00:39,748

I caught you

18

00:00:39,748 --> 00:00:41,207

as part of your presentation

19

00:00:41,207 --> 00:00:42,876

and the with yesterday.

20

00:00:42,876 --> 00:00:44,753

Yeah. Thanks for having me.

21

00:00:44,753 --> 00:00:46,212

You are such a cool background,

22

00:00:46,212 --> 00:00:47,881

and I want the audience to know this.

23

00:00:47,881 --> 00:00:52,427

You worked at ILM, Pixar, unity, AWS.

24

00:00:52,552 --> 00:00:53,678

Now you're building

25

00:00:53,678 --> 00:00:55,180

streaming infrastructure

26

00:00:55,180 --> 00:00:57,390

for photorealistic 3D mirrors.

27

00:00:57,390 --> 00:01:00,727

What drew you to to Miris last year?

28

00:01:01,352 --> 00:01:01,978

Yeah. It's funny.

29

00:01:01,978 --> 00:01:03,855

You know, you name all those companies.

30

00:01:03,855 --> 00:01:05,815

I was thinking about that the other day.

31

00:01:05,815 --> 00:01:07,150

Someone might look at my background.

32

00:01:07,150 --> 00:01:07,525

I was like, wow,

33

00:01:07,525 --> 00:01:09,110

this person has shiny objects and

34

00:01:10,904 --> 00:01:12,113

all these different roles and

35

00:01:12,113 --> 00:01:13,740

all these exciting companies,

36

00:01:13,740 --> 00:01:14,991

and I've really had a

37

00:01:14,991 --> 00:01:15,992

really exciting career.

38

00:01:15,992 --> 00:01:17,410

It's been a lot of fun for me.

39

00:01:17,410 --> 00:01:17,619

You know,

40

00:01:17,619 --> 00:01:18,661

I was thinking about

41

00:01:18,661 --> 00:01:21,081

how the common thread connected

42

00:01:21,081 --> 00:01:22,624

amid all those experiences.

43

00:01:22,624 --> 00:01:23,958

And at the heart of it,

44

00:01:23,958 --> 00:01:25,627

I just really like being in a place

45

00:01:25,627 --> 00:01:26,419

where I can help

46

00:01:26,419 --> 00:01:28,505

co-creators and technologists

47

00:01:28,505 --> 00:01:30,632

actually deliver what they wanted to

48

00:01:30,632 --> 00:01:31,883

build, what they want to build.

49

00:01:31,883 --> 00:01:34,094

So that's been kind of the common thread,

50

00:01:34,094 --> 00:01:36,179

starting when I was at Lucasfilm,

51

00:01:36,179 --> 00:01:36,679

all the way

52

00:01:36,679 --> 00:01:38,848

past working today with Miris.

53

00:01:38,848 --> 00:01:40,475

It's interesting the,

54

00:01:40,475 --> 00:01:41,601

you know, earlier in my career,

55

00:01:41,601 --> 00:01:43,520

I was really focused on

56

00:01:43,520 --> 00:01:45,897

how do I help artists really create

57

00:01:45,897 --> 00:01:47,398

amazing things on screen

58

00:01:47,398 --> 00:01:49,234

or amazing experiences.

59

00:01:49,234 --> 00:01:49,859

And one of the things

60

00:01:49,859 --> 00:01:51,027

that led me to Miris,

61

00:01:51,027 --> 00:01:52,862

like fast forwarding to today, is

62

00:01:52,862 --> 00:01:54,614

how do I help people actually deliver

63

00:01:54,614 --> 00:01:55,907

that content at scale?

64

00:01:55,907 --> 00:01:57,951

You built all this amazing work.

65

00:01:57,951 --> 00:01:58,493

How do you get it

66

00:01:58,493 --> 00:02:00,036

to where it needs to be?

67

00:02:00,036 --> 00:02:00,537

So yeah, it's been

68

00:02:00,537 --> 00:02:02,455

to have a nice little full circle

69

00:02:02,455 --> 00:02:03,164

sort of moment.

70

00:02:03,164 --> 00:02:04,415

Like I started

71

00:02:04,415 --> 00:02:05,834

on the creation side

72

00:02:05,834 --> 00:02:06,501

and now I'm over

73

00:02:06,501 --> 00:02:08,336

on the distribution side.

74

00:02:08,336 --> 00:02:11,339

I think it's fun when you think back to

75

00:02:11,422 --> 00:02:12,715

a career that may look zig zaggy.

76

00:02:12,715 --> 00:02:13,424

I have the same thing

77

00:02:13,424 --> 00:02:14,759

I started at NASA back in the day,

78

00:02:14,759 --> 00:02:15,760

and now here I am.

79

00:02:15,760 --> 00:02:17,428

But there are common threads

80

00:02:17,428 --> 00:02:18,680

and I think when you discover those,

81

00:02:18,680 --> 00:02:20,515

I think it just empowers you

82

00:02:20,515 --> 00:02:21,891

to to want to do

83

00:02:21,891 --> 00:02:24,060

an even better job for your customers.

84

00:02:24,060 --> 00:02:24,602

Well, CoreWeave

85

00:02:24,602 --> 00:02:26,020

has this brand new brand

86

00:02:26,020 --> 00:02:27,981

campaign called Ready for Anything.

87

00:02:27,981 --> 00:02:30,984

I want to you to kind of walk us back

88

00:02:30,984 --> 00:02:33,987

when you were doing R&D for ILM or Pixar,

89

00:02:34,404 --> 00:02:35,780

what did ready for

90

00:02:35,780 --> 00:02:36,865

anything infrastructure...

91

00:02:36,865 --> 00:02:38,158

What did it mean,

92

00:02:38,158 --> 00:02:38,992

what does it look like

93

00:02:38,992 --> 00:02:40,118

behind the scenes back then?

94

00:02:40,118 --> 00:02:40,618

Oh wow.

95

00:02:40,618 --> 00:02:41,744

There's so many stories

96

00:02:41,744 --> 00:02:42,829

I could tell about that.

97

00:02:42,829 --> 00:02:43,121

You know,

98

00:02:43,121 --> 00:02:44,706

the interesting thing about

99

00:02:44,706 --> 00:02:45,707

film production,

100

00:02:45,707 --> 00:02:48,960

be it for VFX or be it for animation,

101

00:02:49,460 --> 00:02:50,545

is ultimately

102

00:02:50,545 --> 00:02:52,630

you're not shipping software.

103

00:02:52,630 --> 00:02:54,465

You're shipping pixels.

104

00:02:54,465 --> 00:02:56,634

All that matters is

105

00:02:56,634 --> 00:02:58,553

how it looks on the screen

106

00:02:58,553 --> 00:02:59,679

or what you're interacting

107

00:02:59,679 --> 00:03:01,139

with in a game.

108

00:03:01,139 --> 00:03:02,432

So oftentimes

109

00:03:02,432 --> 00:03:04,767

production ends up being really messy

110

00:03:04,767 --> 00:03:06,352

from a technology standpoint.

111

00:03:06,352 --> 00:03:08,104

You're building tools that oftentimes

112

00:03:08,104 --> 00:03:09,564

you're having to throw out quickly

113

00:03:09,564 --> 00:03:11,608

because it's just used to solve

114

00:03:11,608 --> 00:03:12,442

a particular problem,

115

00:03:12,442 --> 00:03:14,360

one problem for one shot, right?

116

00:03:14,360 --> 00:03:16,529

And then after that, off it goes.

117

00:03:16,529 --> 00:03:16,696

You know

118

00:03:16,696 --> 00:03:17,697

we're always trying to do our best

119

00:03:17,697 --> 00:03:18,406

to build things

120

00:03:18,406 --> 00:03:20,575

that are more generalized, that work

121

00:03:20,575 --> 00:03:22,535

for lots of different spaces.

122

00:03:22,535 --> 00:03:24,120

But in production, where in the end

123

00:03:24,120 --> 00:03:25,622

you just need to get it on the screen

124

00:03:25,622 --> 00:03:26,122

and looks

125

00:03:26,122 --> 00:03:27,207

really, really good,

126

00:03:27,207 --> 00:03:28,291

you're kind of doing whatever

127

00:03:28,291 --> 00:03:29,876

it takes to do that.

128

00:03:29,876 --> 00:03:32,128

So that was always a lot of fun,

129

00:03:32,128 --> 00:03:33,129

being able to do that.

130

00:03:33,129 --> 00:03:36,132

Like working with creatives

131

00:03:36,174 --> 00:03:37,342

who are really just,

132

00:03:37,342 --> 00:03:38,843

you know, at the height of their powers,

133

00:03:38,843 --> 00:03:41,304

building incredible content.

134

00:03:41,304 --> 00:03:42,513

I've always just loved

135

00:03:42,513 --> 00:03:44,724

helping that sort of person

136

00:03:44,724 --> 00:03:45,642

realize what they want to realize

137

00:03:45,642 --> 00:03:46,434

in terms of content.

138

00:03:46,434 --> 00:03:48,019

I imagine that they're very

139

00:03:48,019 --> 00:03:49,187

inspirational to you

140

00:03:49,187 --> 00:03:51,022

because they have these grand visions

141

00:03:51,022 --> 00:03:53,316

that they want to see brought to life.

142

00:03:53,316 --> 00:03:54,859

And you got to figure out how to do that.

143

00:03:54,859 --> 00:03:55,652

But I imagine

144

00:03:55,652 --> 00:03:57,320

the momentum that they kind of

145

00:03:57,320 --> 00:03:59,155

carry forward is inspiring.

146

00:03:59,155 --> 00:04:00,448

It is. Absolutely.

147

00:04:00,448 --> 00:04:02,283

And so, like your earlier question,

148

00:04:02,283 --> 00:04:04,285

kind of what brought me to Miris,

149

00:04:04,285 --> 00:04:05,036

like a lot of the things

150

00:04:05,036 --> 00:04:06,746

I was noticing and a lot of folks in

151

00:04:06,746 --> 00:04:08,623

Miris have noticed the same inspired

152

00:04:08,623 --> 00:04:11,334

a lot of us from VFX and animation

153

00:04:11,334 --> 00:04:13,670

and from the cloud world and so on.

154

00:04:13,670 --> 00:04:15,088

Like, one of the things that we noticed

155

00:04:15,088 --> 00:04:17,090

is that people can build

156

00:04:17,090 --> 00:04:19,801

and visualize incredible work, right?

157

00:04:19,801 --> 00:04:21,386

But then when it comes time

158

00:04:21,386 --> 00:04:22,971

to actually getting it to where it needs

159

00:04:22,971 --> 00:04:25,181

to be really, really quickly

160

00:04:25,181 --> 00:04:26,432

and at scale,

161

00:04:26,432 --> 00:04:27,100

they're kind of stuck

162

00:04:27,100 --> 00:04:29,394

between these two options, either

163

00:04:29,394 --> 00:04:31,354

decimating it down into like a web

164

00:04:31,354 --> 00:04:33,314

format that can render in the browser,

165

00:04:33,314 --> 00:04:34,315

or they're doing things

166

00:04:34,315 --> 00:04:35,400

like pixel streaming

167

00:04:35,400 --> 00:04:37,610

where they really just can't scale up

168

00:04:37,610 --> 00:04:39,612

to like a consumer grade internet.

169

00:04:39,612 --> 00:04:40,905

And so we thought, hey,

170

00:04:40,905 --> 00:04:41,531

how can we help

171

00:04:41,531 --> 00:04:42,949

creators and developers

172

00:04:42,949 --> 00:04:44,242

actually be able to reach

173

00:04:44,242 --> 00:04:45,493

that billion user scale,

174

00:04:45,493 --> 00:04:47,203

be able to deliver really,

175

00:04:47,203 --> 00:04:49,247

really nice looking content

176

00:04:49,247 --> 00:04:50,748

without compromise?

177

00:04:50,748 --> 00:04:52,208

So yeah, again, like one of the reasons

178

00:04:52,208 --> 00:04:53,543

I'm really excited about being in

179

00:04:53,543 --> 00:04:54,794

Miris is like,

180

00:04:54,794 --> 00:04:56,129

this allows us to help

181

00:04:56,129 --> 00:04:58,589

the right set of creators and developers

182

00:04:58,589 --> 00:04:59,132

realize that.

183

00:04:59,132 --> 00:05:00,717

Without compromise,

184

00:05:00,717 --> 00:05:01,759

that's kind a kind of a key thing

185

00:05:01,759 --> 00:05:02,510

that you mentioned.

186

00:05:02,510 --> 00:05:05,555

When we think about VFX, 3D pipelines,

187

00:05:06,139 --> 00:05:06,806

they've always been

188

00:05:06,806 --> 00:05:08,308

really compute intensive,

189

00:05:08,308 --> 00:05:10,435

but now enter AI.

190

00:05:10,435 --> 00:05:13,438

How is it changing the game in real time?

191

00:05:13,479 --> 00:05:14,063

Yeah, I mean,

192

00:05:14,063 --> 00:05:15,189

I came from a world,

193

00:05:15,189 --> 00:05:16,816

particularly with vFX,

194

00:05:16,816 --> 00:05:18,359

it's very CPU bound.

195

00:05:18,359 --> 00:05:20,320

A lot of the render engines

196

00:05:20,320 --> 00:05:22,071

hadn’t started to leverage GPUs.

197

00:05:22,071 --> 00:05:24,157

And then going back to kind the idea

198

00:05:24,157 --> 00:05:25,992

that it's never really been

199

00:05:25,992 --> 00:05:27,201

so much about the how,

200

00:05:27,201 --> 00:05:27,660

per se,

201

00:05:27,660 --> 00:05:29,037

it's all been about

202

00:05:29,037 --> 00:05:31,497

how do we get the best pixels on screen.

203

00:05:31,497 --> 00:05:32,332

So when we start thinking

204

00:05:32,332 --> 00:05:33,499

about factoring in things

205

00:05:33,499 --> 00:05:34,792

like generative AI,

206

00:05:34,792 --> 00:05:37,795

and AI just more broadly, it's all about

207

00:05:38,254 --> 00:05:39,088

are they produce

208

00:05:39,088 --> 00:05:40,214

or are they providing

209

00:05:40,214 --> 00:05:42,175

tooling in a way of being able

210

00:05:42,175 --> 00:05:43,760

to get to that final image?

211

00:05:43,760 --> 00:05:45,887

That's just like a really exciting time,

212

00:05:45,887 --> 00:05:46,304

I think,

213

00:05:46,304 --> 00:05:47,889

in the industry at large,

214

00:05:47,889 --> 00:05:50,516

not only for VFX and animation,

215

00:05:50,516 --> 00:05:51,976

but also more broadly into things

216

00:05:51,976 --> 00:05:53,770

like physical AI,

217

00:05:53,770 --> 00:05:55,772

digital twins, retail, right?

218

00:05:55,772 --> 00:05:56,314

Like they're all

219

00:05:57,482 --> 00:05:58,649

similar sort of problems,

220

00:05:58,649 --> 00:05:59,400

which is like, you know,

221

00:05:59,400 --> 00:06:02,362

if I'm a, retailer or a brand,

222

00:06:02,487 --> 00:06:05,448

Sameh Louis Vuitton or I feed Spain,

223

00:06:05,448 --> 00:06:06,574

I want to be a little bit

224

00:06:06,574 --> 00:06:09,577

show the, or the products that I have

225

00:06:09,744 --> 00:06:11,120

and the best way possible.

226

00:06:11,120 --> 00:06:13,206

I want things like the fabrics

227

00:06:13,206 --> 00:06:14,665

to really come through.

228

00:06:14,665 --> 00:06:15,833

I might not be doing that.

229

00:06:15,833 --> 00:06:17,085

And so a lot of these companies

230

00:06:17,085 --> 00:06:20,088

are leveraging traditional 3D approaches,

231

00:06:20,088 --> 00:06:20,421

like what

232

00:06:20,421 --> 00:06:23,424

we've used in VFX animation for ages

233

00:06:23,466 --> 00:06:24,801

alongside generative

234

00:06:24,801 --> 00:06:26,094

AI approaches,

235

00:06:26,094 --> 00:06:27,178

and that convergence

236

00:06:27,178 --> 00:06:29,097

of technology together for music.

237

00:06:29,097 --> 00:06:30,973

Really exciting results.

238

00:06:30,973 --> 00:06:31,391

I love that

239

00:06:31,391 --> 00:06:32,975

you mentioned some brands

240

00:06:32,975 --> 00:06:36,062

and I'm fans, but this is more than VFX.

241

00:06:36,479 --> 00:06:37,772

This is and

242

00:06:37,772 --> 00:06:39,690

this is really transcendental

243

00:06:39,690 --> 00:06:41,067

across industries.

244

00:06:41,067 --> 00:06:42,777

Yes, we were with the Aston

245

00:06:42,777 --> 00:06:44,612

Martin Formula one folks this morning.

246

00:06:44,612 --> 00:06:46,030

Call this a sponsor of the team.

247

00:06:46,030 --> 00:06:47,240

And they got one fan.

248

00:06:47,240 --> 00:06:49,575

And one of the things that struck me

249

00:06:49,575 --> 00:06:50,952

is that

250

00:06:50,952 --> 00:06:53,871

what happens in the AI race at F1 races

251

00:06:53,871 --> 00:06:54,956

and with the car

252

00:06:54,956 --> 00:06:58,376

is so influential in other industries.

253

00:06:58,709 --> 00:07:00,628

Talk more about the parallels there.

254

00:07:00,628 --> 00:07:03,923

You mentioned your experiences with VFX,

255

00:07:03,923 --> 00:07:04,799

3D pipelines,

256

00:07:04,799 --> 00:07:07,176

but you also mentioned retail commerce.

257

00:07:07,176 --> 00:07:09,971

Yeah, physical. Are the next frontier.

258

00:07:09,971 --> 00:07:12,181

Walk me through how Miris is helping

259

00:07:12,181 --> 00:07:14,183

those types of organizations

260

00:07:14,183 --> 00:07:16,144

truly accelerate in this age of AI,

261

00:07:16,144 --> 00:07:17,895

and to be able to not compromise,

262

00:07:17,895 --> 00:07:18,604

as you said.

263

00:07:18,604 --> 00:07:19,313

That's right. Yeah.

264

00:07:19,313 --> 00:07:20,106

I mean, I've always been

265

00:07:20,106 --> 00:07:21,315

a really big believer,

266

00:07:21,315 --> 00:07:22,316

no matter

267

00:07:22,316 --> 00:07:23,901

where I've been throughout my career,

268

00:07:23,901 --> 00:07:25,027

and Miris in particular,

269

00:07:25,027 --> 00:07:25,820

in that

270

00:07:25,820 --> 00:07:26,988

the best way to drive

271

00:07:26,988 --> 00:07:29,157

innovation is alongside a customer.

272

00:07:29,157 --> 00:07:31,075

So like when you see innovation

273

00:07:31,075 --> 00:07:32,660

for like an F1,

274

00:07:32,660 --> 00:07:33,453

they're trying to solve

275

00:07:33,453 --> 00:07:35,746

very particular business problems

276

00:07:35,746 --> 00:07:37,081

and they're innovating their way

277

00:07:37,081 --> 00:07:37,498

through that,

278

00:07:37,498 --> 00:07:38,958

which is always just so exciting

279

00:07:38,958 --> 00:07:40,001

to see that sort of thing take shape.

280

00:07:40,001 --> 00:07:40,918

Yeah, yeah,

281

00:07:40,918 --> 00:07:42,545

certainly when I was at industrial,

282

00:07:42,545 --> 00:07:44,005

like magic places

283

00:07:44,005 --> 00:07:45,506

like Electronic Arts

284

00:07:45,506 --> 00:07:46,090

again, like we're

285

00:07:46,090 --> 00:07:47,967

just trying to get like the end product,

286

00:07:47,967 --> 00:07:49,135

the end result.

287

00:07:49,135 --> 00:07:50,595

And there's like a lot of innovation

288

00:07:50,595 --> 00:07:53,139

that takes place as part of that.

289

00:07:53,139 --> 00:07:54,056

I found that,

290

00:07:54,056 --> 00:07:57,018

you know, in unity or at Amazon,

291

00:07:57,018 --> 00:07:57,810

the best way

292

00:07:57,810 --> 00:07:59,061

we were able to really drive,

293

00:07:59,061 --> 00:07:59,896

like a lot of big

294

00:07:59,896 --> 00:08:01,397

leaps in our technology,

295

00:08:01,397 --> 00:08:02,315

is really around

296

00:08:02,315 --> 00:08:03,566

finding those customers

297

00:08:03,566 --> 00:08:04,859

that are trying to do

298

00:08:04,859 --> 00:08:05,818

things that have never been done

299

00:08:05,818 --> 00:08:06,569

before or had,

300

00:08:06,569 --> 00:08:09,197

like a really gnarly scale problem

301

00:08:09,197 --> 00:08:12,200

or a speed problem problem, you name it.

302

00:08:12,617 --> 00:08:13,659

Like jumping in there

303

00:08:13,659 --> 00:08:14,744

headfirst into the customer

304

00:08:14,744 --> 00:08:16,412

to be able to solve that together.

305

00:08:16,412 --> 00:08:17,205

That's where you see

306

00:08:17,205 --> 00:08:18,164

the really interesting

307

00:08:18,164 --> 00:08:19,749

technology innovation.

308

00:08:19,749 --> 00:08:21,626

Like doing it in common Ivory tower.

309

00:08:21,626 --> 00:08:23,586

People can do that and you can see

310

00:08:23,586 --> 00:08:25,129

interesting things from that.

311

00:08:25,129 --> 00:08:26,547

But the rubber really hits the road

312

00:08:26,547 --> 00:08:27,590

when you've got a partner

313

00:08:27,590 --> 00:08:30,051

or a customer can be with together.

314

00:08:30,051 --> 00:08:31,511

Yeah. That shared vision.

315

00:08:31,511 --> 00:08:33,596

Absolutely. Well it's motivating.

316

00:08:33,596 --> 00:08:34,430

It's inspiring.

317

00:08:34,430 --> 00:08:36,682

It's it's probably helping you push past

318

00:08:36,682 --> 00:08:37,934

what we thought was possible

319

00:08:37,934 --> 00:08:38,726

because there's there's

320

00:08:38,726 --> 00:08:40,436

a strong demand for it

321

00:08:40,436 --> 00:08:41,938

in these other industries.

322

00:08:41,938 --> 00:08:42,772

What is driving

323

00:08:42,772 --> 00:08:46,108

the demand for spatial content in retail

324

00:08:46,108 --> 00:08:49,278

and in, physical AI in commerce?

325

00:08:49,278 --> 00:08:50,863

Yeah, there's a common thread.

326

00:08:50,863 --> 00:08:52,573

I have a fun little story about this.

327

00:08:52,573 --> 00:08:54,408

So it actually is tied to

328

00:08:54,408 --> 00:08:55,952

why I came to Microsoft.

329

00:08:55,952 --> 00:08:57,995

So not a year ago.

330

00:08:57,995 --> 00:09:00,665

I have a ten year old and a five year old

331

00:09:00,665 --> 00:09:01,290

at best. Boy.

332

00:09:02,333 --> 00:09:03,543

I walk into my ten year old

333

00:09:03,543 --> 00:09:04,335

into his room

334

00:09:04,335 --> 00:09:05,795

and he's at his computer

335

00:09:05,795 --> 00:09:07,797

and he's making 3D content, right.

336

00:09:07,797 --> 00:09:10,383

So I went to school to make 3D content,

337

00:09:10,383 --> 00:09:12,051

you know, four year college.

338

00:09:12,051 --> 00:09:13,344

Like I'm learning all these tools,

339

00:09:13,344 --> 00:09:14,679

like my, blender.

340

00:09:14,679 --> 00:09:16,556

Yeah, everything about materials

341

00:09:16,556 --> 00:09:16,973

and lighting.

342

00:09:16,973 --> 00:09:18,015

It's quite technical, right?

343

00:09:18,015 --> 00:09:19,267

It always has that.

344

00:09:19,267 --> 00:09:19,642

And here

345

00:09:19,642 --> 00:09:20,518

I catch my ten year

346

00:09:20,518 --> 00:09:21,894

old using one of these generative

347

00:09:21,894 --> 00:09:23,396

AI tools to make, like,

348

00:09:23,396 --> 00:09:26,107

very convincing 3D content, like, wow,

349

00:09:26,107 --> 00:09:27,441

this is amazing, right?

350

00:09:27,441 --> 00:09:30,278

And it dawned on me in that moment that

351

00:09:30,278 --> 00:09:31,445

what we're about to see

352

00:09:31,445 --> 00:09:32,738

and what's already here, it's

353

00:09:32,738 --> 00:09:33,197

just this

354

00:09:33,197 --> 00:09:34,282

absolutely tidal

355

00:09:34,282 --> 00:09:35,783

wave of 3D content,

356

00:09:35,783 --> 00:09:38,160

because this barrier of entry

357

00:09:38,160 --> 00:09:41,080

to creating the content is lowering.

358

00:09:41,080 --> 00:09:41,581

Yes.

359

00:09:41,581 --> 00:09:43,416

It kind of reminds me of

360

00:09:43,416 --> 00:09:43,833

you know,

361

00:09:43,833 --> 00:09:45,543

the democratization of video

362

00:09:45,543 --> 00:09:47,920

that happened at the early March 2000.

363

00:09:47,920 --> 00:09:48,421

That's right.

364

00:09:48,421 --> 00:09:48,629

Yeah.

365

00:09:48,629 --> 00:09:50,172

All things like video

366

00:09:50,172 --> 00:09:53,175

cameras and, video editing software

367

00:09:53,426 --> 00:09:54,760

like The Price has Come down

368

00:09:54,760 --> 00:09:56,053

become accessible, right?

369

00:09:56,053 --> 00:09:56,971

Since, like me,

370

00:09:56,971 --> 00:09:59,181

like younger, like picking up camcorders

371

00:09:59,181 --> 00:10:00,891

and making their movies and editing them.

372

00:10:00,891 --> 00:10:01,100

Yeah.

373

00:10:01,100 --> 00:10:02,310

And they worked all of that, right?

374

00:10:02,310 --> 00:10:03,853

Those things like YouTube

375

00:10:03,853 --> 00:10:05,187

and Twitch text.

376

00:10:05,187 --> 00:10:06,355

Yeah. Netflix.

377

00:10:06,355 --> 00:10:07,189

On and on it goes.

378

00:10:07,189 --> 00:10:08,608

Because the amount of content

379

00:10:08,608 --> 00:10:10,901

being created just, oh, rocketed.

380

00:10:10,901 --> 00:10:11,861

And then all of a sudden people

381

00:10:11,861 --> 00:10:13,613

wanted more channels of distribution.

382

00:10:13,613 --> 00:10:16,115

Discord. Substack. Yep yep yep.

383

00:10:16,115 --> 00:10:18,409

So, you know, fast forward to Boris.

384

00:10:18,409 --> 00:10:18,868

Like, now

385

00:10:18,868 --> 00:10:20,161

that I'm seeing this tidal wave

386

00:10:20,161 --> 00:10:22,163

on the content creation side

387

00:10:22,163 --> 00:10:22,872

really increased

388

00:10:22,872 --> 00:10:23,748

because of generative

389

00:10:23,748 --> 00:10:25,916

AI and everything going on there.

390

00:10:25,916 --> 00:10:27,668

There's just this amazing opportunity

391

00:10:27,668 --> 00:10:29,128

to kind of catch me

392

00:10:29,128 --> 00:10:30,379

downstream of this tidal

393

00:10:30,379 --> 00:10:31,756

wave and catch that.

394

00:10:31,756 --> 00:10:32,423

And so I really,

395

00:10:32,423 --> 00:10:33,215

really tried to solve

396

00:10:33,215 --> 00:10:34,842

that problem of like, hey,

397

00:10:34,842 --> 00:10:36,177

you got all this content

398

00:10:36,177 --> 00:10:37,219

being created by,

399

00:10:37,219 --> 00:10:39,180

you know, people that are always been

400

00:10:39,180 --> 00:10:40,765

incredibly talented around

401

00:10:40,765 --> 00:10:42,141

or even the new creators

402

00:10:42,141 --> 00:10:42,683

that they don't have

403

00:10:42,683 --> 00:10:43,434

any experience with it.

404

00:10:43,434 --> 00:10:44,352

I mean, the

405

00:10:44,352 --> 00:10:45,895

the content creator gig

406

00:10:45,895 --> 00:10:47,730

economy is massively growing.

407

00:10:47,730 --> 00:10:48,939

It's just like

408

00:10:48,939 --> 00:10:50,858

that barrier of entry is lowering where,

409

00:10:50,858 --> 00:10:51,984

you know, I wouldn't be surprised.

410

00:10:51,984 --> 00:10:54,528

You know, my mom increased my 3D content.

411

00:10:54,528 --> 00:10:57,365

Shout out to my mom like she

412

00:10:57,365 --> 00:10:58,282

because my mom is too.

413

00:10:58,282 --> 00:10:59,992

Oh she is. She calls it shut ABC.

414

00:10:59,992 --> 00:11:01,410

But I know what she's talking about.

415

00:11:01,410 --> 00:11:02,912

Oh, that was amazing.

416

00:11:02,912 --> 00:11:04,288

Yeah, I love that.

417

00:11:04,288 --> 00:11:05,706

Tools like that,

418

00:11:05,706 --> 00:11:06,707

like that barrier of entry

419

00:11:06,707 --> 00:11:08,292

and using them is so low.

420

00:11:08,292 --> 00:11:10,795

So when I see that happening for 3D,

421

00:11:10,795 --> 00:11:12,922

it's just a really exciting time to say,

422

00:11:12,922 --> 00:11:13,964

how do we help people

423

00:11:13,964 --> 00:11:16,175

create that at scale? Right.

424

00:11:16,175 --> 00:11:16,884

And that's always been

425

00:11:16,884 --> 00:11:18,344

a bit of the problem

426

00:11:18,344 --> 00:11:20,179

when you're either stuck

427

00:11:20,179 --> 00:11:21,806

decimating your content to be able

428

00:11:21,806 --> 00:11:23,099

to distribute that at scale

429

00:11:23,099 --> 00:11:24,266

and then wait for the downside

430

00:11:24,266 --> 00:11:26,102

and then you compromise, right?

431

00:11:26,102 --> 00:11:27,645

Yeah. Yeah. Quality.

432

00:11:27,645 --> 00:11:28,270

Yeah, yeah.

433

00:11:28,270 --> 00:11:29,480

You're compromising quality

434

00:11:29,480 --> 00:11:31,941

or you're compromising scale.

435

00:11:31,941 --> 00:11:32,358

Right.

436

00:11:32,358 --> 00:11:34,318

And so again, we're kind of in that

437

00:11:34,318 --> 00:11:35,778

like big fat middle

438

00:11:35,778 --> 00:11:36,487

where

439

00:11:36,487 --> 00:11:38,072

we're able to provide like that

440

00:11:38,072 --> 00:11:39,949

really high fidelity content,

441

00:11:39,949 --> 00:11:40,783

but then be able

442

00:11:40,783 --> 00:11:42,368

to uphold things like scale

443

00:11:42,368 --> 00:11:44,120

and then have like 2D economics

444

00:11:44,120 --> 00:11:46,706

and so on. Yeah, it's exciting to talk.

445

00:11:46,706 --> 00:11:47,832

I mentioned that I caught

446

00:11:47,832 --> 00:11:48,833

some of your presentation

447

00:11:48,833 --> 00:11:49,750

downstairs in the car.

448

00:11:49,750 --> 00:11:50,793

We did talk to me

449

00:11:50,793 --> 00:11:52,002

about the CoreWeave partnership.

450

00:11:52,002 --> 00:11:52,586

What are they

451

00:11:52,586 --> 00:11:54,296

how are they enabling you

452

00:11:54,296 --> 00:11:56,215

to enable the creators

453

00:11:56,215 --> 00:11:58,718

to just do what's in their minds?

454

00:11:58,718 --> 00:11:59,176

Yeah.

455

00:11:59,176 --> 00:11:59,844

You know,

456

00:11:59,844 --> 00:12:00,594

the things I mentioned

457

00:12:00,594 --> 00:12:02,179

in my presentation,

458

00:12:02,179 --> 00:12:03,681

there's no for 20 years anymore.

459

00:12:03,681 --> 00:12:05,933

We keep you awake.

460

00:12:05,933 --> 00:12:06,892

The key unlock.

461

00:12:06,892 --> 00:12:08,060

And the reason why we partner

462

00:12:08,060 --> 00:12:09,478

so closely with CoreWeave,

463

00:12:09,478 --> 00:12:10,730

yes it's the GPUs,

464

00:12:10,730 --> 00:12:13,357

but it's actually the platform

465

00:12:13,357 --> 00:12:14,567

as a whole, right?

466

00:12:14,567 --> 00:12:16,736

Like you can go and rent cloud

467

00:12:16,736 --> 00:12:19,155

GPUs from anywhere if you want to.

468

00:12:19,155 --> 00:12:20,322

But like the magic of flex,

469

00:12:20,322 --> 00:12:21,866

where we bring in while we're partners

470

00:12:21,866 --> 00:12:23,242

so closely with them

471

00:12:23,242 --> 00:12:24,535

is things like, hey,

472

00:12:24,535 --> 00:12:26,662

they design their storage solution

473

00:12:26,662 --> 00:12:28,330

such that it has very tightly

474

00:12:28,330 --> 00:12:29,498

to GP workloads.

475

00:12:29,498 --> 00:12:31,375

We're getting that nice performance,

476

00:12:31,375 --> 00:12:32,668

things like how we spin up

477

00:12:32,668 --> 00:12:33,419

our Kubernetes

478

00:12:33,419 --> 00:12:35,087

cluster to be able to

479

00:12:35,087 --> 00:12:36,464

handle the scale that we want,

480

00:12:36,464 --> 00:12:37,715

because remember,

481

00:12:37,715 --> 00:12:38,215

Miris

482

00:12:38,215 --> 00:12:40,176

like we're looking at assets

483

00:12:40,176 --> 00:12:41,552

coming through a pipeline

484

00:12:41,552 --> 00:12:42,928

where we can be dealing with

485

00:12:42,928 --> 00:12:44,513

thousands, tens of thousands,

486

00:12:44,513 --> 00:12:46,182

millions of assets to process

487

00:12:46,182 --> 00:12:48,350

essentially in a single day.

488

00:12:48,350 --> 00:12:49,477

Like we need a platform

489

00:12:49,477 --> 00:12:52,062

that's not only to scale up cheap views,

490

00:12:52,062 --> 00:12:54,523

but like 0.3 storage

491

00:12:54,523 --> 00:12:57,318

to scale the infrastructure as a whole.

492

00:12:57,318 --> 00:12:57,818

So for me,

493

00:12:57,818 --> 00:12:59,361

it's been really exciting to work with,

494

00:12:59,361 --> 00:13:01,739

mainly because they're an AI platform,

495

00:13:01,739 --> 00:13:03,574

they're optimizing around workloads

496

00:13:03,574 --> 00:13:05,451

that we're optimizing around.

497

00:13:05,451 --> 00:13:06,619

And so like that partnership

498

00:13:06,619 --> 00:13:07,703

together has worked

499

00:13:07,703 --> 00:13:09,413

really, really well for us.

500

00:13:09,413 --> 00:13:10,372

And what's the

501

00:13:10,372 --> 00:13:11,540

what are the problems

502

00:13:11,540 --> 00:13:13,167

that that architecture

503

00:13:13,167 --> 00:13:14,585

is helping you to eliminate

504

00:13:14,585 --> 00:13:15,836

for your customers?

505

00:13:15,836 --> 00:13:17,213

It's scale mainly.

506

00:13:17,213 --> 00:13:17,338

Right.

507

00:13:17,338 --> 00:13:19,298

Like scale is like over the years

508

00:13:19,298 --> 00:13:20,466

it's huge driver of that.

509

00:13:20,466 --> 00:13:22,343

So you can think of Miris

510

00:13:22,343 --> 00:13:24,303

and you think about something like

511

00:13:24,303 --> 00:13:25,888

pixel streaming historically.

512

00:13:25,888 --> 00:13:27,640

What happens is here

513

00:13:27,640 --> 00:13:29,266

you have users on the other side and say,

514

00:13:29,266 --> 00:13:30,351

hey, I want to view this

515

00:13:30,351 --> 00:13:32,269

3D asset on the back end.

516

00:13:32,269 --> 00:13:33,395

A service might say, I'm

517

00:13:33,395 --> 00:13:34,522

going to spin up

518

00:13:34,522 --> 00:13:36,398

all this cloud infrastructure

519

00:13:36,398 --> 00:13:37,817

and then deliver that to you.

520

00:13:37,817 --> 00:13:39,360

And like, hey, subscriptions, right?

521

00:13:39,360 --> 00:13:40,528

Like a buffer.

522

00:13:40,528 --> 00:13:41,070

The problem is

523

00:13:41,070 --> 00:13:43,405

it's like you're having to spin up a GQ

524

00:13:43,405 --> 00:13:44,990

for a user, right? Okay.

525

00:13:44,990 --> 00:13:46,617

So what matters is what we do with for we

526

00:13:46,617 --> 00:13:48,077

because we flip that equation,

527

00:13:48,077 --> 00:13:49,286

we say, hey,

528

00:13:49,286 --> 00:13:50,746

we're gonna take your asset

529

00:13:50,746 --> 00:13:51,580

and we're going to spin up

530

00:13:51,580 --> 00:13:54,291

a GPU upstream, upstream.

531

00:13:54,291 --> 00:13:56,961

And so we're running all these GPUs

532

00:13:56,961 --> 00:13:57,253

and you know,

533

00:13:57,253 --> 00:13:58,462

we're using the latest

534

00:13:58,462 --> 00:14:00,381

hardware from Nvidia

535

00:14:00,381 --> 00:14:02,299

for we to be able to do that.

536

00:14:02,299 --> 00:14:03,467

But we do that once it's

537

00:14:03,467 --> 00:14:05,427

actively for that.

538

00:14:05,427 --> 00:14:06,095

And then when it comes

539

00:14:06,095 --> 00:14:07,346

time to actually deliver that

540

00:14:07,346 --> 00:14:09,181

to customers, for clients,

541

00:14:09,181 --> 00:14:10,933

it's like GPUs are effectively

542

00:14:10,933 --> 00:14:12,643

out of the equation at that point. Okay.

543

00:14:12,643 --> 00:14:13,978

So like that, that ends up

544

00:14:13,978 --> 00:14:14,562

allowing us

545

00:14:14,562 --> 00:14:16,981

to deliver at billion user scale.

546

00:14:16,981 --> 00:14:17,773

And it's not because

547

00:14:17,773 --> 00:14:19,191

we're not using GPUs.

548

00:14:19,191 --> 00:14:20,192

It's mainly that,

549

00:14:20,192 --> 00:14:20,609

again,

550

00:14:20,609 --> 00:14:21,819

with the support in the sense

551

00:14:21,819 --> 00:14:23,821

that we're using it upstream.

552

00:14:23,821 --> 00:14:25,030

And that's the big

553

00:14:25,030 --> 00:14:25,573

I imagine

554

00:14:25,573 --> 00:14:27,116

that's a pretty big differentiator

555

00:14:27,116 --> 00:14:27,867

for mirrors

556

00:14:27,867 --> 00:14:29,785

that flipping the script

557

00:14:29,785 --> 00:14:31,078

and going upstream.

558

00:14:31,078 --> 00:14:32,746

Yes. Yeah, it's a big deal.

559

00:14:32,746 --> 00:14:35,332

And like the other part of this is like

560

00:14:35,332 --> 00:14:37,459

the reason why it's an AI and ML sort of

561

00:14:37,459 --> 00:14:38,669

workload is

562

00:14:38,669 --> 00:14:40,254

what comes out the other

563

00:14:40,254 --> 00:14:41,881

side of Miris is not meshes.

564

00:14:41,881 --> 00:14:42,756

It's not the things

565

00:14:42,756 --> 00:14:43,799

that we've been using over

566

00:14:43,799 --> 00:14:45,718

the last many decades. Right.

567

00:14:45,718 --> 00:14:47,469

What comes out the other side of Miris

568

00:14:47,469 --> 00:14:49,555

is actually volumetric content. Okay.

569

00:14:49,555 --> 00:14:51,557

We're using things like radiance fields

570

00:14:51,557 --> 00:14:52,099

and so on.

571

00:14:52,099 --> 00:14:53,893

To be able to effectively construct

572

00:14:53,893 --> 00:14:55,853

a representation of your asset

573

00:14:55,853 --> 00:14:57,897

now allows us a lot of flexibility

574

00:14:57,897 --> 00:14:58,606

in the ability

575

00:14:58,606 --> 00:15:00,649

to deliver the content at scale

576

00:15:00,649 --> 00:15:02,693

and in a very adaptive way.

577

00:15:02,693 --> 00:15:04,194

And that's actually paid off

578

00:15:04,194 --> 00:15:05,029

huge for us,

579

00:15:05,029 --> 00:15:05,946

because it allows us to do

580

00:15:05,946 --> 00:15:07,072

a lot of creative things

581

00:15:07,072 --> 00:15:09,366

in terms of optimization,

582

00:15:09,366 --> 00:15:10,284

if you think about things

583

00:15:10,284 --> 00:15:12,995

like display, AI or robotics. Yeah.

584

00:15:12,995 --> 00:15:14,288

One of the things we hear often

585

00:15:14,288 --> 00:15:16,206

from those customers are, hey,

586

00:15:16,206 --> 00:15:17,666

I want to run all these simulations,

587

00:15:17,666 --> 00:15:18,584

but I kind of moved

588

00:15:18,584 --> 00:15:20,044

all this 3D

589

00:15:20,044 --> 00:15:22,379

gigabytes, terabytes worth of data.

590

00:15:22,379 --> 00:15:24,340

I only have so much Vram on these

591

00:15:24,340 --> 00:15:25,341

GPUs now.

592

00:15:25,341 --> 00:15:25,966

I can only run

593

00:15:25,966 --> 00:15:27,593

maybe one simulation at a time.

594

00:15:27,593 --> 00:15:30,262

Oh, because Miris is able to stream

595

00:15:30,262 --> 00:15:31,430

now all of a sudden

596

00:15:31,430 --> 00:15:34,058

we can very creatively kind of impact

597

00:15:34,058 --> 00:15:35,184

the Ram of

598

00:15:35,184 --> 00:15:36,143

what's not being used,

599

00:15:36,143 --> 00:15:37,603

what the robot isn't seeing in

600

00:15:37,603 --> 00:15:38,812

any given time.

601

00:15:38,812 --> 00:15:39,438

Now all of a sudden,

602

00:15:39,438 --> 00:15:40,898

these companies can run

603

00:15:40,898 --> 00:15:43,651

2030 simulations on the same.

604

00:15:43,651 --> 00:15:45,444

So for them, it's a huge cost savings

605

00:15:45,444 --> 00:15:45,986

that allows them

606

00:15:45,986 --> 00:15:47,196

to actually scale up even more.

607

00:15:47,196 --> 00:15:49,406

That's the scale story. That's right.

608

00:15:49,406 --> 00:15:50,783

And so what are you guys demoing

609

00:15:50,783 --> 00:15:52,076

in the CoreWeave booth downstairs?

610

00:15:52,076 --> 00:15:53,243

Yeah, we've got a couple interesting

611

00:15:53,243 --> 00:15:53,953

things we're showing.

612

00:15:53,953 --> 00:15:54,536

So namely

613

00:15:54,536 --> 00:15:54,954

we're showing

614

00:15:54,954 --> 00:15:56,997

a digital twin presentation.

615

00:15:56,997 --> 00:15:58,457

So really high fidelity

616

00:15:58,457 --> 00:16:00,292

digital content that's being streamed.

617

00:16:00,292 --> 00:16:01,919

So it has the same visual fidelity

618

00:16:01,919 --> 00:16:02,586

that you might expect

619

00:16:02,586 --> 00:16:04,296

from like a pixel stream.

620

00:16:04,296 --> 00:16:05,339

But as we've discussed, it's

621

00:16:05,339 --> 00:16:06,590

not a pixel stream.

622

00:16:06,590 --> 00:16:07,675

It's being rasterized

623

00:16:07,675 --> 00:16:09,843

right there on the computer.

624

00:16:09,843 --> 00:16:10,803

But we're even a stream

625

00:16:10,803 --> 00:16:12,346

that in a very adaptive way.

626

00:16:12,346 --> 00:16:13,347

So, you know,

627

00:16:13,347 --> 00:16:15,265

say we have bad Wi-Fi in here.

628

00:16:15,265 --> 00:16:15,891

It's fine.

629

00:16:15,891 --> 00:16:16,141

Right?

630

00:16:16,141 --> 00:16:16,976

Like little adapt

631

00:16:16,976 --> 00:16:18,185

based on those conditions.

632

00:16:18,185 --> 00:16:18,602

Okay.

633

00:16:18,602 --> 00:16:18,978

Let's say

634

00:16:18,978 --> 00:16:21,271

you want to look at it up real close

635

00:16:21,271 --> 00:16:22,982

or stream that content.

636

00:16:22,982 --> 00:16:25,234

So we're showing a really high detail

637

00:16:25,234 --> 00:16:27,778

turbine, right?

638

00:16:27,778 --> 00:16:28,529

Oh yes yes yes

639

00:16:28,529 --> 00:16:29,530

I saw that in your presentation.

640

00:16:29,530 --> 00:16:30,698

Yeah. That was awesome. Right?

641

00:16:30,698 --> 00:16:32,783

I haven't even better one in the demo.

642

00:16:32,783 --> 00:16:34,493

So you have to see it. Okay.

643

00:16:34,493 --> 00:16:34,743

You know,

644

00:16:34,743 --> 00:16:36,078

you can zoom in and it, like,

645

00:16:36,078 --> 00:16:37,746

automatically resolves, right?

646

00:16:37,746 --> 00:16:38,956

It's really, really cool.

647

00:16:38,956 --> 00:16:40,207

You showed I forget

648

00:16:40,207 --> 00:16:41,041

what the who the comparison

649

00:16:41,041 --> 00:16:42,251

was on the left side.

650

00:16:42,251 --> 00:16:44,837

The speed difference was dramatic.

651

00:16:44,837 --> 00:16:45,087

Yeah.

652

00:16:45,087 --> 00:16:48,799

With what general Acme versus Miris.

653

00:16:48,799 --> 00:16:50,592

It was like oh exactly.

654

00:16:50,592 --> 00:16:50,843

Wow.

655

00:16:50,843 --> 00:16:52,761

Factor in like that's so important

656

00:16:52,761 --> 00:16:55,889

because when we talk to like, retailers,

657

00:16:55,889 --> 00:16:56,974

like people building

658

00:16:56,974 --> 00:16:57,558

like product

659

00:16:57,558 --> 00:17:00,561

integrators, or advertisers, right.

660

00:17:00,894 --> 00:17:01,937

Their big thing is like, hey,

661

00:17:01,937 --> 00:17:03,439

I want to get my customers

662

00:17:03,439 --> 00:17:05,065

into this experience.

663

00:17:05,065 --> 00:17:06,608

But in the current like web

664

00:17:06,608 --> 00:17:08,736

download model, they're waiting

665

00:17:08,736 --> 00:17:10,779

three seconds, five seconds, 10s,

666

00:17:10,779 --> 00:17:12,239

maybe even up to 30s.

667

00:17:12,239 --> 00:17:13,699

And what they want is, is like

668

00:17:13,699 --> 00:17:14,366

we see people

669

00:17:14,366 --> 00:17:16,201

just churning out of those experiences

670

00:17:16,201 --> 00:17:17,619

because they're so used to the experience

671

00:17:17,619 --> 00:17:19,580

being instant or non-voting.

672

00:17:19,580 --> 00:17:20,664

Yeah. Yes.

673

00:17:20,664 --> 00:17:22,791

And they churn. Yeah. Yes.

674

00:17:22,791 --> 00:17:24,084

The fact that we can show some interest.

675

00:17:24,084 --> 00:17:25,377

Oh yes, it

676

00:17:25,377 --> 00:17:26,628

is a game changer for them

677

00:17:26,628 --> 00:17:27,379

because all of a sudden

678

00:17:27,379 --> 00:17:28,672

you've got that customer

679

00:17:28,672 --> 00:17:29,548

in the experience.

680

00:17:29,548 --> 00:17:32,134

They can start using it right away.

681

00:17:32,134 --> 00:17:33,594

And like you can kind of extrapolate

682

00:17:33,594 --> 00:17:35,471

that out from like retail customers

683

00:17:35,471 --> 00:17:36,680

to physical eye

684

00:17:36,680 --> 00:17:39,016

digital twin media entertainment.

685

00:17:39,016 --> 00:17:39,433

So I'm

686

00:17:39,433 --> 00:17:40,642

like who's going to turn down the speed.

687

00:17:40,642 --> 00:17:42,019

Right. Like right.

688

00:17:42,019 --> 00:17:42,436

Yes.

689

00:17:42,436 --> 00:17:43,103

No one is saying,

690

00:17:43,103 --> 00:17:44,605

hey, it'd be cool if it's slower,

691

00:17:44,605 --> 00:17:46,565

less data slower. Yes. Right.

692

00:17:46,565 --> 00:17:49,109

There's no there's no, sticker for that.

693

00:17:49,109 --> 00:17:51,403

No. No one's worry that slow is good.

694

00:17:51,403 --> 00:17:52,654

You know, a great sticker.

695

00:17:52,654 --> 00:17:54,823

So if a fortune 500 leader

696

00:17:54,823 --> 00:17:57,242

is watching this and what? What's neat?

697

00:17:57,242 --> 00:17:58,619

And if they're wondering,

698

00:17:58,619 --> 00:17:59,995

is my

699

00:17:59,995 --> 00:18:02,456

use case ready for AI infrastructure?

700

00:18:02,456 --> 00:18:04,708

What would you tell them now

701

00:18:04,708 --> 00:18:06,168

to flip that script?

702

00:18:06,168 --> 00:18:06,502

Yeah.

703

00:18:06,502 --> 00:18:07,586

Well, I'd say, number one

704

00:18:07,586 --> 00:18:09,421

that the status quo is terrorist.

705

00:18:09,421 --> 00:18:09,713

Right.

706

00:18:09,713 --> 00:18:11,006

Like there is

707

00:18:11,006 --> 00:18:11,965

things are changing

708

00:18:11,965 --> 00:18:13,133

at such a rapid rate

709

00:18:13,133 --> 00:18:13,884

and things,

710

00:18:13,884 --> 00:18:14,218

you know,

711

00:18:14,218 --> 00:18:15,928

just talking about computer graphics

712

00:18:15,928 --> 00:18:17,096

in particular,

713

00:18:17,096 --> 00:18:20,057

we've been so, so used to living in a

714

00:18:20,057 --> 00:18:21,308

like polygon.

715

00:18:21,308 --> 00:18:22,518

Yeah. Forever.

716

00:18:22,518 --> 00:18:24,144

And polygons will persist, right?

717

00:18:24,144 --> 00:18:25,687

Like they're not going anywhere.

718

00:18:25,687 --> 00:18:26,980

But things that way, they're

719

00:18:26,980 --> 00:18:29,274

we're seeing like the radiance field,

720

00:18:29,274 --> 00:18:30,859

Gaussian splat, etc.

721

00:18:30,859 --> 00:18:32,152

space or accelerating

722

00:18:32,152 --> 00:18:33,278

essentially created that.

723

00:18:33,278 --> 00:18:34,738

And what I would tell people

724

00:18:34,738 --> 00:18:35,697

leaders in the fortune

725

00:18:35,697 --> 00:18:36,782

500 companies are

726

00:18:36,782 --> 00:18:37,991

using that as an example.

727

00:18:37,991 --> 00:18:40,994

And it's like be mindful of like how fast

728

00:18:41,203 --> 00:18:42,496

this is changing

729

00:18:42,496 --> 00:18:44,581

and stay ahead of the adoption curve.

730

00:18:44,581 --> 00:18:45,082

Okay.

731

00:18:45,082 --> 00:18:47,918

So our goal is in effect, an exponential.

732

00:18:47,918 --> 00:18:48,210

Yes.

733

00:18:48,210 --> 00:18:49,169

Like maybe even beyond

734

00:18:49,169 --> 00:18:50,462

an exponential curve.

735

00:18:50,462 --> 00:18:52,548

And it's really like a good time

736

00:18:52,548 --> 00:18:52,923

every day.

737

00:18:52,923 --> 00:18:54,216

Like when I wake up in the morning

738

00:18:54,216 --> 00:18:55,884

I'm seeing the world

739

00:18:55,884 --> 00:18:57,386

subtly change, right?

740

00:18:57,386 --> 00:18:58,929

Sometimes right, in subtle ways.

741

00:18:58,929 --> 00:18:59,847

And so for me,

742

00:18:59,847 --> 00:19:01,515

like really staying very close

743

00:19:01,515 --> 00:19:02,808

to all the innovation

744

00:19:02,808 --> 00:19:03,976

that's happening

745

00:19:03,976 --> 00:19:06,979

in computer graphics and AI in general,

746

00:19:07,062 --> 00:19:07,896

and how you can, like,

747

00:19:07,896 --> 00:19:08,981

train your workforce

748

00:19:08,981 --> 00:19:10,482

to take advantage of that.

749

00:19:10,482 --> 00:19:12,568

Really take a second, third, fourth

750

00:19:12,568 --> 00:19:14,278

look at how you can move

751

00:19:14,278 --> 00:19:16,029

faster or scale

752

00:19:16,029 --> 00:19:17,573

to like millions of users,

753

00:19:17,573 --> 00:19:17,865

like what

754

00:19:17,865 --> 00:19:19,408

would that potentially look like?

755

00:19:19,408 --> 00:19:20,659

It's just all changing.

756

00:19:21,869 --> 00:19:22,744

Quickly.

757

00:19:22,744 --> 00:19:23,537

And it's yeah, it's

758

00:19:23,537 --> 00:19:24,746

an exciting time to be a tech.

759

00:19:24,746 --> 00:19:26,790

It is such an exciting time because,

760

00:19:26,790 --> 00:19:27,291

I mean,

761

00:19:27,291 --> 00:19:28,250

I can't even imagine

762

00:19:28,250 --> 00:19:28,834

what we're going to be

763

00:19:28,834 --> 00:19:29,835

talking about next.

764

00:19:29,835 --> 00:19:32,337

GTC oh, is it fair to say, though, that

765

00:19:32,337 --> 00:19:33,213

that mirrors

766

00:19:33,213 --> 00:19:36,216

where we are, you enabling your customers

767

00:19:36,383 --> 00:19:38,969

to get ahead of that of that

768

00:19:38,969 --> 00:19:40,637

curve because

769

00:19:40,637 --> 00:19:41,722

so many folks are behind

770

00:19:41,722 --> 00:19:42,848

how what you're saying

771

00:19:42,848 --> 00:19:44,099

it's it's realistic

772

00:19:44,099 --> 00:19:45,058

with these two companies.

773

00:19:45,058 --> 00:19:46,476

You can get ahead of this.

774

00:19:46,476 --> 00:19:47,519

Yeah, absolutely.

775

00:19:47,519 --> 00:19:47,936

I get like

776

00:19:47,936 --> 00:19:49,021

this is where the CoreWeave

777

00:19:49,021 --> 00:19:51,064

partnership has been so vital

778

00:19:51,064 --> 00:19:52,566

because they have this same vision

779

00:19:52,566 --> 00:19:53,400

of like

780

00:19:53,400 --> 00:19:55,861

computer graphics is having this shift,

781

00:19:55,861 --> 00:19:57,029

that moment.

782

00:19:57,029 --> 00:19:57,613

And so,

783

00:19:57,613 --> 00:19:58,655

you know, going back

784

00:19:58,655 --> 00:20:01,450

to like my ten year old son like, yeah,

785

00:20:01,450 --> 00:20:02,784

we're seeing just like this

786

00:20:02,784 --> 00:20:05,204

the views of 3D content.

787

00:20:05,204 --> 00:20:07,039

How do we provide a platform together

788

00:20:07,039 --> 00:20:08,665

that can truly scale to that?

789

00:20:08,665 --> 00:20:10,292

How can we get content

790

00:20:10,292 --> 00:20:11,168

where it needs to be?

791

00:20:11,168 --> 00:20:12,002

How can we take

792

00:20:12,002 --> 00:20:14,171

even like really rich physical

793

00:20:14,171 --> 00:20:15,255

AI, multi

794

00:20:15,255 --> 00:20:17,090

gigabyte ranging into terabytes

795

00:20:17,090 --> 00:20:18,050

of content?

796

00:20:18,050 --> 00:20:18,759

How can we get that

797

00:20:18,759 --> 00:20:19,676

to where it needs to be,

798

00:20:19,676 --> 00:20:21,178

and how can we get it to a place

799

00:20:21,178 --> 00:20:23,305

where upwards of billions of people

800

00:20:23,305 --> 00:20:25,349

can all see that at the same time?

801

00:20:25,349 --> 00:20:26,475

Like that?

802

00:20:26,475 --> 00:20:27,601

All those things

803

00:20:27,601 --> 00:20:28,977

like we talk about

804

00:20:28,977 --> 00:20:32,105

scale, fidelity, cost and speed

805

00:20:32,481 --> 00:20:33,523

when you're able to like, hit

806

00:20:33,523 --> 00:20:35,275

all those four things in parallel.

807

00:20:35,275 --> 00:20:37,402

Again, that's where the magic is.

808

00:20:37,402 --> 00:20:38,612

That's Nirvana.

809

00:20:38,612 --> 00:20:39,363

It is. Right.

810

00:20:39,363 --> 00:20:40,280

Like that

811

00:20:40,280 --> 00:20:41,240

kind of going all the way back

812

00:20:41,240 --> 00:20:43,158

to where we started, like in my career.

813

00:20:43,158 --> 00:20:43,492

Yeah,

814

00:20:43,492 --> 00:20:44,743

like being able to build

815

00:20:44,743 --> 00:20:45,410

a platform

816

00:20:45,410 --> 00:20:46,370

that allows you

817

00:20:46,370 --> 00:20:49,373

to kind of unleash 3D content at scale

818

00:20:49,414 --> 00:20:50,540

to get it to everyone.

819

00:20:50,540 --> 00:20:52,042

Like, that's kind of a holy grail for me,

820

00:20:52,042 --> 00:20:54,211

is able to help creators and developers.

821

00:20:54,211 --> 00:20:55,420

I'm sure that you from 10

822

00:20:55,420 --> 00:20:56,463

or 15 years ago couldn't

823

00:20:56,463 --> 00:20:59,466

even imagine where we are in 2026.

824

00:20:59,466 --> 00:20:59,967

Oh no.

825

00:20:59,967 --> 00:21:01,510

I mean, I was

826

00:21:01,510 --> 00:21:02,844

in my early 3D,

827

00:21:02,844 --> 00:21:04,554

you know, my early 3D years.

828

00:21:04,554 --> 00:21:04,763

You know,

829

00:21:04,763 --> 00:21:06,014

I was used to like

830

00:21:06,014 --> 00:21:07,516

downloading off the network

831

00:21:07,516 --> 00:21:09,393

like a bytes of data

832

00:21:09,393 --> 00:21:10,769

coming back the next day.

833

00:21:10,769 --> 00:21:13,272

Right, like rendering massive frames

834

00:21:13,272 --> 00:21:14,690

that could take weeks, right?

835

00:21:14,690 --> 00:21:15,816

Like, you know,

836

00:21:15,816 --> 00:21:17,859

some of these things will persist by

837

00:21:17,859 --> 00:21:18,235

the way,

838

00:21:18,235 --> 00:21:19,278

I just

839

00:21:19,278 --> 00:21:19,778

again, like,

840

00:21:19,778 --> 00:21:21,321

things are changing at such a rate,

841

00:21:21,321 --> 00:21:23,282

like the idea that I can provide,

842

00:21:23,282 --> 00:21:25,701

I can take a ten gigabyte asset

843

00:21:25,701 --> 00:21:26,493

and get that to

844

00:21:26,493 --> 00:21:28,745

someone in almost an instant.

845

00:21:28,745 --> 00:21:29,037

You know,

846

00:21:29,037 --> 00:21:30,998

if you told me that 20 years ago,

847

00:21:30,998 --> 00:21:31,373

like, oh,

848

00:21:31,373 --> 00:21:33,208

you saw the speed of light right now.

849

00:21:33,208 --> 00:21:34,584

Yeah. Not by the way,

850

00:21:35,961 --> 00:21:36,545

is it?

851

00:21:36,545 --> 00:21:36,795

Yeah.

852

00:21:36,795 --> 00:21:38,380

That's a concept

853

00:21:38,380 --> 00:21:40,048

that has radically changed.

854

00:21:40,048 --> 00:21:41,091

And again,

855

00:21:41,091 --> 00:21:43,385

that's the last question for you.

856

00:21:43,385 --> 00:21:46,388

You've been to 4 or 5 gigs before.

857

00:21:46,430 --> 00:21:48,890

What is different this year

858

00:21:48,890 --> 00:21:50,225

even compared to last year.

859

00:21:50,225 --> 00:21:51,768

Oh I mean look at it down here.

860

00:21:51,768 --> 00:21:53,270

It's just like wow.

861

00:21:53,270 --> 00:21:54,813

Right. Like it is wild.

862

00:21:54,813 --> 00:21:55,063

You know,

863

00:21:55,063 --> 00:21:57,816

I've been to so many conferences, right?

864

00:21:57,816 --> 00:21:59,067

GDC and

865

00:21:59,067 --> 00:22:01,403

what I'm like noticing about GDC,

866

00:22:01,403 --> 00:22:03,030

when you talk about like an exponential,

867

00:22:03,030 --> 00:22:04,114

like the energy

868

00:22:04,114 --> 00:22:05,365

and cheating in GTC

869

00:22:05,365 --> 00:22:07,117

is like on this exponential curve,

870

00:22:07,117 --> 00:22:10,203

like it's really, really high this year.

871

00:22:10,203 --> 00:22:13,290

I was noticing even yesterday, like just

872

00:22:13,290 --> 00:22:13,707

the sheer

873

00:22:13,707 --> 00:22:14,875

number of people here

874

00:22:14,875 --> 00:22:17,085

are just, like, incredible.

875

00:22:17,085 --> 00:22:17,878

It's not even,

876

00:22:17,878 --> 00:22:18,337

you know, like, you

877

00:22:18,337 --> 00:22:19,504

go to a lot of these problems.

878

00:22:19,504 --> 00:22:20,756

It's it's like, you know,

879

00:22:20,756 --> 00:22:21,631

the experience is like

880

00:22:21,631 --> 00:22:22,632

within the exhibit,

881

00:22:22,632 --> 00:22:24,634

all GDC has gotten to the scale

882

00:22:24,634 --> 00:22:26,261

where it's like spilled out. Oh, yeah.

883

00:22:26,261 --> 00:22:26,678

And there's like

884

00:22:26,678 --> 00:22:28,055

a whole festival outside.

885

00:22:28,055 --> 00:22:30,682

Yeah, it is like a festival. Yeah. Yes.

886

00:22:30,682 --> 00:22:31,808

You know, I was like,

887

00:22:31,808 --> 00:22:33,143

I was walking in the other day,

888

00:22:33,143 --> 00:22:34,102

I heard like

889

00:22:34,102 --> 00:22:36,521

a live band with a singing country

890

00:22:36,521 --> 00:22:37,647

song about fiscal.

891

00:22:37,647 --> 00:22:39,316

I know, I'm like, wow.

892

00:22:39,316 --> 00:22:39,775

Like,

893

00:22:39,775 --> 00:22:40,650

I don't know if this is, like,

894

00:22:40,650 --> 00:22:42,110

the sound of a cold or what,

895

00:22:42,110 --> 00:22:43,403

but whatever it is like,

896

00:22:43,403 --> 00:22:44,237

it's pretty incredible.

897

00:22:44,237 --> 00:22:45,739

I love I yeah, I love it

898

00:22:45,739 --> 00:22:47,366

and and my wrist has a beta program.

899

00:22:47,366 --> 00:22:47,949

Tell the audience

900

00:22:47,949 --> 00:22:49,284

a little bit more about the beta program

901

00:22:49,284 --> 00:22:50,077

and where they can access

902

00:22:50,077 --> 00:22:50,952

the beta program.

903

00:22:50,952 --> 00:22:53,580

Starting on, March 24th and next week.

904

00:22:53,580 --> 00:22:54,414

We are hard at work

905

00:22:54,414 --> 00:22:56,416

and making it a great point.

906

00:22:56,416 --> 00:22:58,126

Yeah, we're signing people up now.

907

00:22:58,126 --> 00:22:59,503

We're going to start

908

00:22:59,503 --> 00:23:01,004

with when on the 24th.

909

00:23:01,004 --> 00:23:03,924

And yeah, we're really excited about it.

910

00:23:03,924 --> 00:23:05,258

Very interested

911

00:23:05,258 --> 00:23:06,510

in seeing what people create

912

00:23:06,510 --> 00:23:07,719

and distribute on this platform.

913

00:23:07,719 --> 00:23:08,720

What interest theta.

914

00:23:08,720 --> 00:23:09,596

What will they have

915

00:23:09,596 --> 00:23:11,681

unique access to from Miris?

916

00:23:11,681 --> 00:23:12,182

Oh, sure.

917

00:23:12,182 --> 00:23:13,767

They'll have access to the full faith.

918

00:23:13,767 --> 00:23:14,101

Right.

919

00:23:14,101 --> 00:23:17,229

So you'll be able to upload, 3D content.

920

00:23:17,229 --> 00:23:19,606

So we support open USD,

921

00:23:19,606 --> 00:23:21,775

similar to our friends over at Nvidia.

922

00:23:21,775 --> 00:23:23,318

So that's why we call it a video player.

923

00:23:23,318 --> 00:23:24,569

And also,

924

00:23:24,569 --> 00:23:27,072

you'll be able to open, or upload open

925

00:23:27,072 --> 00:23:29,574

USD content. We take it from there.

926

00:23:29,574 --> 00:23:32,661

We put that into our streaming content

927

00:23:32,661 --> 00:23:33,578

and out the other end

928

00:23:33,578 --> 00:23:35,664

you'll be able to see it in our software.

929

00:23:35,664 --> 00:23:36,706

You'll be able to take that

930

00:23:36,706 --> 00:23:38,417

with a few lines of code

931

00:23:38,417 --> 00:23:39,000

and integrate

932

00:23:39,000 --> 00:23:41,044

that into a web application.

933

00:23:41,044 --> 00:23:43,338

So three.js as an example,

934

00:23:43,338 --> 00:23:44,798

you don't have to do everything.

935

00:23:44,798 --> 00:23:47,384

It just seamlessly works, with bridges.

936

00:23:47,384 --> 00:23:48,301

And it's accessible

937

00:23:48,301 --> 00:23:49,302

from the virus website.

938

00:23:49,302 --> 00:23:50,429

That's right. Yeah.

939

00:23:50,429 --> 00:23:51,513

That's all you need to do this thing.

940

00:23:51,513 --> 00:23:54,015

Yes. I did have a little bit of code.

941

00:23:54,015 --> 00:23:54,683

We got plenty of,

942

00:23:54,683 --> 00:23:57,269

tutorials and documentation galore,

943

00:23:57,269 --> 00:23:58,687

and, yeah,

944

00:23:58,687 --> 00:24:00,105

you get up and running super fast.

945

00:24:00,105 --> 00:24:00,439

Okay.

946

00:24:00,439 --> 00:24:01,273

I bet you cannot wait

947

00:24:01,273 --> 00:24:02,190

to see the use cases

948

00:24:02,190 --> 00:24:03,066

that explode from this.

949

00:24:03,066 --> 00:24:05,152

Yeah, I'm very a kid in a candy store.

950

00:24:05,152 --> 00:24:06,903

I imagine a few early customers

951

00:24:06,903 --> 00:24:08,321

and just seeing

952

00:24:08,321 --> 00:24:10,073

not only like the visual quality,

953

00:24:10,073 --> 00:24:11,450

but also just like the scale

954

00:24:11,450 --> 00:24:12,534

that they're talking about.

955

00:24:12,534 --> 00:24:12,909

Yeah.

956

00:24:12,909 --> 00:24:14,453

You know, again, like talking

957

00:24:14,453 --> 00:24:15,412

to like the 20 years,

958

00:24:15,412 --> 00:24:17,164

like 20 years in the past first to me.

959

00:24:17,164 --> 00:24:18,582

Yeah. Right.

960

00:24:18,582 --> 00:24:19,916

Oh no.

961

00:24:19,916 --> 00:24:20,709

So I can't even

962

00:24:20,709 --> 00:24:22,586

I can't even like think out to next year

963

00:24:22,586 --> 00:24:23,753

because,

964

00:24:23,753 --> 00:24:25,380

you know, generative AI is so fast

965

00:24:25,380 --> 00:24:26,840

a genetic and everything. But.

966

00:24:26,840 --> 00:24:29,009

Well, thank you for sharing your story.

967

00:24:29,009 --> 00:24:31,136

The common thread of what you bring in,

968

00:24:31,136 --> 00:24:33,430

but also the massive impact

969

00:24:33,430 --> 00:24:34,431

that Miris and CoreWeave

970

00:24:34,431 --> 00:24:36,266

are making on other industries

971

00:24:36,266 --> 00:24:39,144

besides VFX, retail, commerce, physical.

972

00:24:39,144 --> 00:24:41,188

I cannot wait to hear

973

00:24:41,188 --> 00:24:41,938

what's next

974

00:24:41,938 --> 00:24:42,522

because I'm sure it's

975

00:24:42,522 --> 00:24:43,398

going to be fantastic.

976

00:24:43,398 --> 00:24:44,399

But thank you for stopping by

977

00:24:44,399 --> 00:24:45,650

and sharing your story.

978

00:24:45,650 --> 00:24:47,110

Yeah, for well, McDonald.

979

00:24:47,110 --> 00:24:48,320

I'm Lisa Martin, coming to you

980

00:24:48,320 --> 00:24:51,323

live from CoreWeave booth at GTC day two.

981

00:24:51,364 --> 00:24:52,574

Thanks for watching guys, and we'll

982

00:24:52,574 --> 00:24:53,867

see you on the next episode.

‍