Revolutionizing Decentralized Computing: Alessandro De Carli on Acurast's Mobile Device Network (Part 1)

Anthony Perl: 00:00:02

Revolutionizing decentralized computing Alexandro Dali

2

: 00:00:06

on Acura Mobile Device Network, part one.

3

: 00:00:09

In this episode, Dr.

4

: 00:00:10

Gemma Green sits down with Alexandro Dali,

co-founder of Acura, a groundbreaking

5

: 00:00:16

platform that's transforming the

decentralized computing space.

6

: 00:00:20

Alessandro shares how accuracy allows

people to onboard their smartphones

7

: 00:00:24

to a global network, creating

a powerful computing resource.

8

: 00:00:28

With over 106,000 devices and

800,000 CPU core already connected,

9

: 00:00:35

they explore how this technology

gives mobile devices a second life.

10

: 00:00:39

The security advantages of trusted

execution environments and the

11

: 00:00:43

innovative token economics that

make the system sustainable.

12

: 00:00:46

With his background in mobile

security, as former lead engineer at

13

: 00:00:50

Credit Suisse Alexandro explains how

accuracy is creating an unstoppable,

14

: 00:00:55

decentralized computing network

spanning 140 countries that rivals

15

: 00:01:00

traditional data centers in capacity.

16

: 00:01:03

Jacob Stanler from Power Ledger

also joins the discussion.

17

: 00:01:06

I'm your co-host, Anthony Perlin.

18

: 00:01:08

Whether you're an investor or

a startup looking for insights,

19

: 00:01:11

it's time to get unblocked.

20

: 00:01:16

Jemma Green: Alessandro,

welcome to Unblock.

21

: 00:01:19

You are a co-founder of aist, which

is really gaining attention in

22

: 00:01:25

the decentralized compute space.

23

: 00:01:28

For those that are new to that, could

you walk us through what AIST is

24

: 00:01:32

and the problem that it's solving?

25

: 00:01:35

Alessandro De Carli: Yeah, absolutely.

26

: 00:01:36

Hello Gemma, and thank you so much

for, uh, having me on your podcast.

27

: 00:01:41

I'm very excited to be here.

28

: 00:01:43

What actors does.

29

: 00:01:45

In a nutshell is it allows people to

onboard their mobile device, so their

30

: 00:01:52

smartphone and dedicate the compute

power that that device has in it to

31

: 00:01:57

the network so that somebody else is

capable of utilizing that resource as

32

: 00:02:02

a compute unit for their project to

host something to calculate something.

33

: 00:02:08

Or even to observe something.

34

: 00:02:10

That is what AC is doing in a nutshell.

35

: 00:02:13

Taking the billions of phones out there

that get churned through regularly,

36

: 00:02:19

and every year we have about 1.4

37

: 00:02:22

billion of new mobile devices

that reach your pockets, right?

38

: 00:02:27

But that also means that every year, 1.4.

39

: 00:02:31

Billion devices land on the secondary

market, and Accu gives them basically a

40

: 00:02:37

second life and allows people to really

participate in a super decentralized

41

: 00:02:42

network with their mobile device.

42

: 00:02:45

Jemma Green: Great.

43

: 00:02:46

And how many mobile phones are

connected to the Accu Network right now?

44

: 00:02:51

Alessandro De Carli: Well as we

speak right now, and this is a

45

: 00:02:54

number that you can look up live

because the network is already live.

46

: 00:02:59

We have 106,000 mobile devices connected

offering their compute to the network.

47

: 00:03:06

Jemma Green: Amazing.

48

: 00:03:07

And what does that translate

to in terms of like the amount

49

: 00:03:10

of storage that's available?

50

: 00:03:13

Alessandro De Carli: I think ultimately

storage is maybe the thing where mobile

51

: 00:03:18

devices don't shine the brightest.

52

: 00:03:22

However, where they are amazing is in

the compute density, and with these a

53

: 00:03:29

hundred thousand plus devices, you end

up having more than 800,000 CPU cores.

54

: 00:03:35

And many people think that these CPU cores

are less performant than let's say, the

55

: 00:03:40

ones in servers, but that's not the case.

56

: 00:03:43

There is a CPU benchmark score that

does not discriminate between server

57

: 00:03:48

CPUs and mobile CPUs, and you will quite

quickly see that, especially because of

58

: 00:03:54

the Apple chips, but also the Qualcomm

chips that are in Android phones.

59

: 00:03:58

They actually are pretty much and

very high on top there, so you really

60

: 00:04:05

do get a lot of compute and these

devices also have quite a lot of ramp.

61

: 00:04:11

So compute and ram is where they excel at.

62

: 00:04:14

Let's also not forget, especially now

in the era of AI and whatnot, the GU of

63

: 00:04:20

the mobile device, there is something

really special about it and it is

64

: 00:04:25

that mobile devices have a so-called

unified memory architecture, and what

65

: 00:04:30

that means is that your CPU and your

GPU are capable of utilizing the ram.

66

: 00:04:37

So if your CPU gets.

67

: 00:04:40

24 gigs of ram, which some glasses do

have, then your GPU also gets that,

68

: 00:04:46

which is, uh, interesting for running

things like LLMs and so on and so forth.

69

: 00:04:51

Jemma Green: Got it.

70

: 00:04:52

Alessandro De Carli: But

it's not necessarily the

71

: 00:04:53

reason why we started this.

72

: 00:04:55

Oh, because my background

is in mobile security.

73

: 00:04:58

I used to be the lead mobile

security engineer at, uh,

74

: 00:05:02

credit Suisse back in the days.

75

: 00:05:04

And basically, I.

76

: 00:05:06

Built a very deep relationship with the

security aspect of mobile devices because

77

: 00:05:12

about 14 years ago, I then decided to

leave the bank and create an own company

78

: 00:05:18

that was offering mobile security to very

large enterprises all around the globe.

79

: 00:05:23

We have four out of the five

largest defense companies

80

: 00:05:27

in the US as our clients.

81

: 00:05:30

And that's also why I ultimately started

Accurus because if you want to create

82

: 00:05:37

a decentralized compute network, there

is no other way than having it secure

83

: 00:05:43

because there's two aspects that are

really, really important for a real world

84

: 00:05:48

adoption of this technology, and one is

that nobody is capable of faking hardware.

85

: 00:05:53

That's where distrust of execution

environment comes in Really, really

86

: 00:05:57

handy for mobile devices, and you

can basically see the proof of that.

87

: 00:06:02

If you go on social media channels

like Reddit or even YouTube and the

88

: 00:06:07

type accuracy, you will see a ton

of people showcasing their setup.

89

: 00:06:13

But the other aspect is also

the fact that you're able to

90

: 00:06:17

run the compute confidentially.

91

: 00:06:20

So let's say you have iPhone that

you're now onboarding onto accuracy.

92

: 00:06:26

If I need to compute something like

running an LLM and want to give you that.

93

: 00:06:31

It's really important that you're not

capable of inspecting the data that I

94

: 00:06:35

share with at LLA because otherwise,

yeah, basically confidentiality really

95

: 00:06:40

matters in this case, because otherwise

nobody's going to use a more public

96

: 00:06:45

kind of version of a cloud, right?

97

: 00:06:48

Mm-hmm.

98

: 00:06:49

So that's ultimately the

reason why I started with aist,

99

: 00:06:52

but then realized the scale.

100

: 00:06:56

Jemma Green: Got it.

101

: 00:06:57

I mean, I think it would be helpful

to get like contextualize the scale.

102

: 00:07:01

You said you talked about

like 800,000 set use.

103

: 00:07:05

What does that translate to in the

way that, like how does that compare

104

: 00:07:08

to, say, a data center or some other

metric that they could ground that in?

105

: 00:07:13

Alessandro De Carli: I mean,

there's many ways to compare this.

106

: 00:07:15

The standard mobile device has about

eight core and eight gigabytes software.

107

: 00:07:20

This is a median modern device.

108

: 00:07:23

Today, if you take a look

at something like an AWS.

109

: 00:07:29

They offer you also servers that have

that, those specs, but those servers will

110

: 00:07:36

cost you per month around 200 to $300

with those exact same specs in terms of

111

: 00:07:43

compute and run that you get out of it.

112

: 00:07:45

Mm-hmm.

113

: 00:07:46

So now ultimately you are

able to run a hundred thousand

114

: 00:07:51

compute units on the network.

115

: 00:07:55

Which will translate to a small to medium

sized data center here in Switzerland.

116

: 00:08:03

So basically, if you take a look

at Green, which is a data center,

117

: 00:08:08

which is posting one of the, uh,

hyperscalers is not official, but I do

118

: 00:08:14

know that, and they have about 50,000

compute units in their data center.

119

: 00:08:21

So Switzerland obviously is not a

huge country, and the compute units

120

: 00:08:26

tend to get way more if you go in a

country such as the us but for Swiss

121

: 00:08:32

scale, and this is basically where

we are located, we already would have

122

: 00:08:37

the crowdsourced capacity of two data

centers ultimately in terms of compute?

123

: 00:08:43

Yeah.

124

: 00:08:44

Jemma Green: Mm-hmm.

125

: 00:08:45

Gotcha.

126

: 00:08:46

That's useful.

127

: 00:08:47

How do you guys make money?

128

: 00:08:48

Like is it the same model as Amazon

Web Service who you mentioned or

129

: 00:08:52

other cloud service providers?

130

: 00:08:53

Like how do people pay for the

service and how do the participants

131

: 00:08:57

in the network benefit and how

does that curious make money?

132

: 00:09:01

Alessandro De Carli: So it's

important to say that the.

133

: 00:09:04

Network.

134

: 00:09:05

The accuracy network and protocol is set

up in a very, very decentralized way.

135

: 00:09:10

This means that there is an entity that

bootstrapped the whole thing, and this

136

: 00:09:15

entity is called the Accuracy Association,

and it's a nonprofit association which is

137

: 00:09:22

located in Switzerland and basically is

there to be kind of legal entity that has.

138

: 00:09:31

And Kick started this network, but it

is a completely decentralized network.

139

: 00:09:36

So now that it's out and it started,

it's not like the Actors Association

140

: 00:09:41

would be capable of stopping it anymore.

141

: 00:09:45

And that's.

142

: 00:09:46

One aspect that me personally I find

very interesting because ultimately

143

: 00:09:51

similar to how Bitcoin operates,

and no single entity is capable of

144

: 00:09:55

shutting that down for accuracy.

145

: 00:09:57

Now, instead of the service of having the

ability to transfer value, you have the

146

: 00:10:04

ability to provide and get the compute

like decentralized compute, right?

147

: 00:10:11

This all around the globe,

no matter where you live.

148

: 00:10:14

We have now more than 140 countries

and that we're covering with people

149

: 00:10:19

offering computes in those countries.

150

: 00:10:22

That's something would never have

been possible with a centralized

151

: 00:10:26

setup because all of a sudden

you need to create all these

152

: 00:10:30

subsidiaries and maybe some countries

are more interesting than others.

153

: 00:10:34

So you wouldn't think of it

that way when Yeah, calculating

154

: 00:10:39

the business case around it.

155

: 00:10:40

So now we have this very, very accessible,

unstoppable network all around the globe.

156

: 00:10:46

The question is, how does value

get generated and the value

157

: 00:10:52

that is being generated comes

from your real mobile device.

158

: 00:10:56

So there is the security

aspect that I mentioned before.

159

: 00:11:00

We have a couple of companies

that heavily leveraging that one.

160

: 00:11:05

There is one company that is using

this to wrap and unwrap bitcoins.

161

: 00:11:10

They're securing about a hundred

million dollars worth of bitcoins

162

: 00:11:15

just with accuracy alone.

163

: 00:11:17

That service is obviously a paid

service that people are then paying

164

: 00:11:23

in accurate tokens and how the.

165

: 00:11:26

Model around the network works.

166

: 00:11:29

Is that similar to how Bitcoin

has a fixed inflation that gets

167

: 00:11:34

distributed among whoever is

participating in lock production?

168

: 00:11:38

With ura, you have a fixed inflation

that is distributed among whoever

169

: 00:11:42

is participating in providing

compute and is also advertising

170

: 00:11:47

their compute availability.

171

: 00:11:50

Now whenever someone utilizes compute,

what you end up doing is that you

172

: 00:11:55

completely just burn as kind of compute

fee Acura tokens, so that ultimately you

173

: 00:12:04

do have a way to generate that flywheel

wheel right, and counter this inflation

174

: 00:12:11

that I mentioned in the beginning.

175

: 00:12:12

So ultimately, if you're able

to reach a usage capacity.

176

: 00:12:17

Of about 20% of the network.

177

: 00:12:19

The whole, yeah.

178

: 00:12:21

Economics is designed so that the

supply starts to become deflationary,

179

: 00:12:27

and that's also where the, the

value ultimately can be captured.

180

: 00:12:32

Jemma Green: Got it.

181

: 00:12:32

I mean, I think people that are deep

in the blockchain space will totally

182

: 00:12:35

understand what you've just described.

183

: 00:12:37

But for people that are perhaps new to

the topic and are listening to this,

184

: 00:12:41

could you just explain in, in a bit more

layman terms, like the commercial model

185

: 00:12:45

and the business model for people to

get paid and how accurate makes money?

186

: 00:12:50

Alessandro De Carli: Yeah, absolutely.

187

: 00:12:51

So if we take a look at a network

that maybe people are more

188

: 00:12:55

familiar with, the Bitcoin network.

189

: 00:12:58

The way the protocol works is that it

is a blockchain and the most important

190

: 00:13:06

thing is to create blocks so that

it can offer its services to people.

191

: 00:13:12

Now, in the context of Bitcoin,

that is the ability to transfer

192

: 00:13:16

value and hold value, right?

193

: 00:13:18

The this payment idea behind it

Perus, the idea is that you basically.

194

: 00:13:26

Purchase compute and consume

compute similar to a Bitcoin miner.

195

: 00:13:31

This is the entity that

generates those blocks.

196

: 00:13:34

Also, entities on the Acura network

get rewarded with the accuracy token

197

: 00:13:41

whenever they are kind of advertising and

onboarding their compute onto the network

198

: 00:13:47

and whenever that compute is being used.

199

: 00:13:50

So what this ultimately translates

to is that you're able to have

200

: 00:13:56

or get rewarded with a token.

201

: 00:13:59

This is the accurate token that

you then ultimately will be able to

202

: 00:14:04

convert into whichever currency that

you want through a exchange, be it a

203

: 00:14:09

decentralized or centralized exchange.

204

: 00:14:12

It's very similar to how Bitcoin

mining works in those regards.

205

: 00:14:16

Great.

206

: 00:14:18

Jemma Green: And then for Accurus

itself, how does it make money?

207

: 00:14:21

Does it take a percentage or

is there some other model?

208

: 00:14:25

Alessandro De Carli: So accuracy itself

only requires money to be able to

209

: 00:14:29

fund and foster and promote and make

visible the whole protocol and network.

210

: 00:14:36

It has not the purpose, like the

Accuracy association just to be here.

211

: 00:14:41

Here, it has not the purpose to

pursue any profit making activities.

212

: 00:14:48

However, obviously there are

tokens around and how this.

213

: 00:14:54

Mission can be fulfilled long-term

is through a treasury, which is

214

: 00:15:01

then on chain, completely on chain

controlled by the token holders,

215

: 00:15:06

and that treasury is constantly

being replenished by the inflation.

216

: 00:15:12

So that ultimately you

then as a token holder.

217

: 00:15:17

Will be able to decide which activity

makes the most sense at that specific

218

: 00:15:23

moment to further grow the network.

219

: 00:15:26

So it's very much aligned

with its community.

220

: 00:15:29

Once the whole system is completely,

let's say at some point, maybe you will

221

: 00:15:35

not even require an accurate association

anymore, similar to the fact that there is

222

: 00:15:40

no Bitcoin kind of entity controlling it.

223

: 00:15:46

This is also the power behind these

decentralized networks, right?

224

: 00:15:50

It's, it's all kind of designed and

set up with some crypto incentives and

225

: 00:15:58

that works around around the token.

226

: 00:16:00

Now, you may ask, okay,

how are early contributors.

227

: 00:16:06

Being rewarded and whatnot.

228

: 00:16:08

Obviously early contributors also

do have token allocations and are

229

: 00:16:13

incentivized with token locations

such as the team behind hackers.

230

: 00:16:17

So there is a very much an

alignment for the token holders

231

: 00:16:21

and the early contributors there.

232

: 00:16:25

Jemma Green: Great.

233

: 00:16:26

Thanks for unpacking that really

simply as well as in the more

234

: 00:16:29

complex terms, Alessandro.

235

: 00:16:31

Jakob Stammler: Absolutely, and I think.

236

: 00:16:33

Just kind of going off of what you've

already said in this regard, I think

237

: 00:16:37

you are obviously not just part of

this Urist association and a co-founder

238

: 00:16:41

of this accuracy protocol, but you're

also one of the founding members of

239

: 00:16:46

the Deep End Association and with Deep

N Networks where accuracy obviously

240

: 00:16:51

is part of the deep end space.

241

: 00:16:53

There's always a question of token

ons, and you've mentioned that people.

242

: 00:16:58

Pay for some of the services.

243

: 00:16:59

So the demand side of accuracy services

are paid for in the A CU token.

244

: 00:17:05

Of course.

245

: 00:17:06

Then there's a question about when

this token rises, doesn't the service

246

: 00:17:11

become more expensive and not as.

247

: 00:17:13

Available anymore to your partners.

248

: 00:17:16

So how do you kind of bridge between

this supply side, hope for the token

249

: 00:17:21

to increase so that you as a processor

have a return on your token that

250

: 00:17:25

you receive, and the demand side

that uses the token to pay for it.

251

: 00:17:29

I know you have a dual token system.

252

: 00:17:31

Maybe you can elaborate on

how this factors into this

253

: 00:17:34

whole, um, process as well.

254

: 00:17:37

Alessandro De Carli: Yeah, absolutely.

255

: 00:17:38

I think it's a very clever question

because it's also a question about the

256

: 00:17:43

long-term sustainability of the project.

257

: 00:17:46

As we all know with crypto projects, the

value of the PO is ultimately what the

258

: 00:17:53

market decides that that value should be.

259

: 00:17:56

Right?

260

: 00:17:57

And these can be, especially

for early projects.

261

: 00:18:01

Quite a lot of swing and volatility

in the token price, which obviously

262

: 00:18:07

would immediately hurt a system where

people cannot really plan ahead on

263

: 00:18:13

how much their compute costs would be.

264

: 00:18:15

Right?

265

: 00:18:16

Ultimately.

266

: 00:18:17

But this is where we apply a system.

267

: 00:18:21

And it's just to maybe

close off that bracket.

268

: 00:18:24

It's also a bit a vicious circle, right?

269

: 00:18:26

Because the, the moment you are

really, really successful and people

270

: 00:18:31

are using it would then probably

also be the moment where it becomes

271

: 00:18:35

very expensive for people to use it.

272

: 00:18:37

So that is a negative feedback

loop that you absolutely need

273

: 00:18:40

to take care of and think about.

274

: 00:18:43

And I thought about that for a really,

really long time, and with the very first

275

: 00:18:50

version of how we thought about billing

and whatnot on the Testnet, we did have

276

: 00:18:57

an issue with exactly this topic that

you're mentioning, and I will quickly

277

: 00:19:01

explain how the first version worked and

why that version had that issue versus

278

: 00:19:07

then why the new version that that we

just launched does not have that issue.

279

: 00:19:13

So for the first version, it

was very, very kind of naive and

280

: 00:19:16

natural to say, look, whenever you

have computer availability, the

281

: 00:19:20

network will pay for your heartbeat.

282

: 00:19:23

And whenever someone consumes

your resources, then you should

283

: 00:19:28

basically be paid by whoever

is consuming those resources.

284

: 00:19:31

Right?

285

: 00:19:32

That's a very naive way, non

crypto way of thinking about this.

286

: 00:19:38

But after user surveys, what we

found out is that people providing

287

: 00:19:43

compute, they don't necessarily

want to be part of this lottery.

288

: 00:19:46

They don't want to today earn this

much, and tomorrow earn another amount.

289

: 00:19:52

It's way more interesting

if this is somewhat a stable

290

: 00:19:56

amount that they're earning.

291

: 00:19:58

This allows them to calculate

their ROI and whatnot.

292

: 00:20:01

And this played basically in our hands.

293

: 00:20:04

So now the new system has a fixed

inflation that is distributed

294

: 00:20:10

among all the participants.

295

: 00:20:12

If it is a single participant offering

a single CPU, then that person

296

: 00:20:17

would get all of the inflation.

297

: 00:20:19

If it is many, then you just receive

a fraction of that inflation.

298

: 00:20:24

This means if you have a lot of compute.

299

: 00:20:27

You receive relatively

little amount of tokens.

300

: 00:20:31

If you have not so much compute available,

then you receive a lot of the tokens.

301

: 00:20:37

So far, so good.

302

: 00:20:38

Right now, how do we do the pricing then?

303

: 00:20:41

Well, the pricing is basically

a multiplier on the heart

304

: 00:20:45

beating reward that you receive.

305

: 00:20:48

Let's say that you're part of this

pool and there is 50 million tokens

306

: 00:20:53

that are being distributed annually.

307

: 00:20:55

There is 50 million participants,

which is not the numbers at the

308

: 00:20:59

moment, but just to have an example.

309

: 00:21:02

Then everyone would receive a

single token per year, right?

310

: 00:21:06

If now someone would.

311

: 00:21:08

Go there and want to purchase compute

from this pool of 50 million participants,

312

: 00:21:13

they would be paying five times that.

313

: 00:21:16

So basically five Acura tokens.

314

: 00:21:19

If you do this, you basically break

that cycle because let's assume that

315

: 00:21:26

the price for a single Acura token.

316

: 00:21:29

Really skyrockets becomes

very, very valuable.

317

: 00:21:32

Now all of a sudden, whoever is providing

compute has a lot of incentive to provide

318

: 00:21:38

more compute, and this would then dilute

everybody that is inside this pool of.

319

: 00:21:46

People, and that would flash the

price of the actual compute like this.

320

: 00:21:52

Although we have everything designed

around the Acura token, it does not

321

: 00:21:57

really matter that much what the

ultimate price of the Acura token is

322

: 00:22:01

because it's completely self-regulating.

323

: 00:22:04

Right?

324

: 00:22:05

So whenever there is, uh, very

little compute, then that would

325

: 00:22:09

probably mean that the Acura

token is really, really cheap.

326

: 00:22:13

That would mean however,

that you would need to.

327

: 00:22:15

Pay more accurate tokens to

get the compute out, hence

328

: 00:22:19

burning more accurate tokens.

329

: 00:22:21

That is how that system works.

330

: 00:22:23

But I agree with you that for a

deepen network, this is one of the

331

: 00:22:30

most crucial things to think about

because you don't want to be basically.

332

: 00:22:37

Deer in the headlights when all of a

sudden you're really successful and

333

: 00:22:42

people really want your token and

it becomes prohibitively expensive.

334

: 00:22:46

This was the big issue from

Chainlink, by the way, that they

335

: 00:22:50

still haven't really solved, okay?

336

: 00:22:52

They solved it with just bilateral

contracts, but it's not really

337

: 00:22:55

a very decentralized way, right?

338

: 00:22:59

Anthony Perl: I was gonna ask

as well, dumb it down for us.

339

: 00:23:02

You've got a smartphone,

how do you get involved?

340

: 00:23:04

And also I'm interested because

the smartphone technology isn't

341

: 00:23:08

kind of built to last that long.

342

: 00:23:11

So how do you deal with that kind of

lifespan of technology and the ability of

343

: 00:23:15

it to be working well when you need it?

344

: 00:23:19

Alessandro De Carli: Those are.

345

: 00:23:20

Two really interesting questions.

346

: 00:23:22

Let me first start with how

to onboard compute onboarding.

347

: 00:23:27

Compute onto Acura is really as

simple as downloading an app from the

348

: 00:23:31

stores and then launching that app.

349

: 00:23:34

It's really important that Acura never

runs in the background, so if you do

350

: 00:23:39

install that application, like you're

the one controlling when the computer

351

: 00:23:44

is being provided and you cannot even.

352

: 00:23:47

Do this by mistake, because the app

needs to be running in the foreground.

353

: 00:23:52

This means that you cannot use it

for anything else on that side.

354

: 00:23:56

And this was a conscious design

decision because we didn't want

355

: 00:24:01

people to be impacted negatively

in the turns that they started.

356

: 00:24:06

And then all of a sudden,

because a big deployment comes

357

: 00:24:09

in, that uses a ton of compute.

358

: 00:24:11

After three hours, the battery is dead.

359

: 00:24:14

Right?

360

: 00:24:15

So basically you go there,

you install the app.

361

: 00:24:18

You launch the app, and that's it.

362

: 00:24:21

You don't need to be a Linux leak, you

don't need to have any PhD in computer

363

: 00:24:26

science to be able to onboard compute.

364

: 00:24:29

So that's really easy.

365

: 00:24:31

And if you go on hub.acura.com,

366

: 00:24:33

you will also find out that it's

very easy to deploy stuff on Acura.

367

: 00:24:37

So both sites are quite easy, and

you can then manage your fleet of

368

: 00:24:42

devices directly from the Acura hub.

369

: 00:24:45

Now the other question is how do you

deal with the hardware itself, right?

370

: 00:24:52

Because you're saying that this

hardware is not necessarily

371

: 00:24:56

made to last for so long.

372

: 00:24:59

The first thing to be said

there is that this hardware is

373

: 00:25:03

absolutely made to run 24 7.

374

: 00:25:07

Okay?

375

: 00:25:07

Because if someone utilizes their

smartphone as their alarm clock to

376

: 00:25:12

wake up, they never turn off their

phone, so that's already clear.

377

: 00:25:17

The device runs 24 7.

378

: 00:25:19

It is designed to do so now, if you

want to let it run for a very long time.

379

: 00:25:26

The main issue that you're going to

have is not necessarily the aging of

380

: 00:25:31

the compute unit or the RAM or the GPU

or the storage, but is the battery.

381

: 00:25:38

The battery is the, the main issue that

really where the device ages and where

382

: 00:25:46

you can feel that the device ages, right?

383

: 00:25:49

I mean, think about it

also as a consumer, right?

384

: 00:25:51

If you have a, uh, five-year-old iPhone.

385

: 00:25:54

You will probably not want to just

change it because it's not fast enough,

386

: 00:25:59

because I personally cannot really

notice the performance difference from

387

: 00:26:03

an iPhone that is five years old towards

the one that is now the latest version.

388

: 00:26:08

But you will absolutely notice if

your battery is dead after half a

389

: 00:26:12

day, and that's typically where people

then start changing their phone.

390

: 00:26:17

Now for aist, we have the phones

always twenty four seven plugged in.

391

: 00:26:24

There's no people that have

that just running on battery.

392

: 00:26:29

This means we don't really

care that much about the actual

393

: 00:26:34

battery capacity on that side.

394

: 00:26:37

The cool thing about this spur of mobile

devices is that in computer science you

395

: 00:26:42

have this thing called Moore's Law, right?

396

: 00:26:45

Which basically says that transistors

over time double and the capacity doubles,

397

: 00:26:51

the performance doubles and whatnot.

398

: 00:26:53

If you take a look at the server,

it's usually being used for 10

399

: 00:26:58

years plus, depending on which

kind of use case for a phone.

400

: 00:27:03

There's people that change

them after one year, right?

401

: 00:27:06

So you have something like one

and a half, two years pops.

402

: 00:27:10

In those eight years difference you

made quantum leaps in in technology.

403

: 00:27:17

This hardware is also the most

designed piece of hardware

404

: 00:27:20

consumers can get their hands on.

405

: 00:27:23

Apple invests about $33 billion a

year on designing and crafting the

406

: 00:27:29

best possible piece of hardware,

and they make more than half of

407

: 00:27:33

their revenue with the iPhone alone.

408

: 00:27:35

So ultimately, you're able to just profit

from all of this technological advancement

409

: 00:27:41

and very high density in compute.

410

: 00:27:45

Because of that.

411

: 00:27:46

So ultimately it's something that

plays more in our favor, but than not.

412

: 00:27:52

Anthony Perl: And that's

all we have time for.

413

: 00:27:53

In part one of the discussion with

Alessandra Ali, we've explored how URA

414

: 00:27:57

has built a decentralized computing

network using everyday smartphones.

415

: 00:28:02

Or part two delves deeper into

the practical user experience

416

: 00:28:06

and real world applications.

417

: 00:28:08

Join us as we hear about the

experience of running accuracy,

418

: 00:28:11

including its onboarding process.

419

: 00:28:13

We'll explore how the system's popularity

created unexpected market effects with

420

: 00:28:18

Google Pixel phones selling out across

Europe, the advantages of constantly

421

: 00:28:22

upgrading hardware through normal

phone replacement cycles, and how this

422

: 00:28:26

technology is making decentralized

computing accessible to everyday users.

423

: 00:28:31

Don't miss this thought provoking

continuation of our discussion on

424

: 00:28:35

how mobile devices are creating a new

paradigm in decentralized infrastructure.

425

: 00:28:40

Please check out the show notes

for information on Power Ledger

426

: 00:28:43

and other contact information.

427

: 00:28:45

We welcome your comments and

feedback and please hit subscribe

428

: 00:28:49

wherever you are listening.

429

: 00:28:50

This podcast was produced

by podcast done for You.

430

: 00:28:53

We look forward to your

company next time on Unblocked.