0:00

Are you getting pressure from finance to justify or

0:02

reduce your cloud bill? Cloud Zero

0:04

is the only cloud cost platform loved by

0:06

engineers and trusted by finance. Cloud

0:09

Zero can identify unused, idle, or

0:11

over-provision resources, alert you to spend anomalies,

0:13

and organize 100% of your

0:15

spend into a framework that mirrors your business

0:18

structure, like cost per customer, product feature, or

0:20

team. It's the most

0:22

powerful platform ever built to provide

0:24

accurate, granular visibility into your total

0:26

cloud spend, without the typical pitfalls

0:28

of legacy cloud cost management tools,

0:30

like endless tagging or clunky Kubernetes

0:32

support. Manage cost, optimize

0:34

development, and maximize profit, all in

0:37

one platform. Join companies

0:39

like Rapid7, Drift, and

0:41

SeatGeek by visiting cloudzero.com/cloudcast

0:43

to get started. That's

0:46

cloudzero.com/cloudcast. Visit today

0:48

to experience immediate and ongoing savings on your

0:50

cloud bill. Cloudcast Media

0:53

presents from the Massive Studios in

0:55

Raleigh, North Carolina. This is the

0:57

Cloudcast with Aaron Delb and Brian

0:59

Gracely, bringing you the best of

1:01

cloud computing from around the world.

1:08

Good morning, good evening, wherever you are, and welcome back to

1:10

the Cloudcast. We are coming to you live from the Massive

1:12

Cloudcast Studios here in Raleigh, North Carolina. Hope

1:14

everybody is doing well. Another Sunday perspective show

1:16

as we creep towards the middle of February,

1:18

2024. Hope everybody is

1:21

doing well. I know a lot of folks are

1:23

going to be listening to this on a Sunday,

1:25

the day of the Super Bowl here in the

1:27

United States. Lots and lots of people watch that,

1:29

both here in the States and around the world.

1:31

Hopefully, whether you're tuning in for football or tuning

1:33

in for Taylor Swift, you'll enjoy the game. Whether

1:35

you are getting ready for potentially Valentine's Day

1:38

with your significant other, the

1:40

person you love, some sort of plans you have around

1:42

that. Hopefully, a little Sunday

1:44

perspective jumps in there somewhere. I

1:46

want to talk about one of

1:48

these Sunday perspectives where somewhere

1:51

during the week or the weeks ahead, some

1:53

sort of data point pops up and we try and

1:56

make a little bit of sense of it. As I

1:58

was digging into this more and more, there's The

2:00

sort of a lot of lot of

2:02

weirdness around this one. So where this

2:04

kind of sort of for me out

2:06

and will obviously would dig it isn't.

2:08

Second part of Shelves was listening to

2:10

a new podcast the popped up at

2:12

my tech feed as called the Bg

2:14

to podcast it is. Couple of very

2:16

well known Silicon Valley venture capitalists are

2:19

just starting of Podcast. The two of

2:21

them are talking about how much tops

2:23

in the industry and one of the

2:25

things that they were talking about was

2:27

the recent kind of trend of the

2:29

big cloud providers investing. Making very, very

2:31

large investments in. Things.

2:33

Like open A I in Anthropic and

2:35

some of very big A I modeled

2:37

building companies and their their point was

2:40

you know essentially this is this. Feels

2:42

a little different Eight they've essentially were

2:44

saying you know how is the Vc

2:46

world. Measuring or try to figure

2:48

out the valuations of Ai companies and

2:50

they were having a hard time doing

2:53

that because so much of the Ai

2:55

world right now feels a little bit

2:57

askew towards somebody is really big Models:

2:59

Are you the Anthropic snow bunny eyes?

3:02

Mistral since with his other. Arm

3:05

and. You know, so they

3:07

are trying to figure out how does that work in

3:09

than they were kind of bemoaning the idea that. While

3:12

he sees typically have to go out

3:14

and raise funding and in that funding

3:16

is is invested into the series companies

3:18

that the cloud riders were in essence

3:20

using ah, I'm less of a cash

3:22

investment and more have club credits. So

3:25

in essence ah you know we we

3:27

invested four billion dollars but really only

3:29

say a half billion dollars of as

3:31

in cash and the or three and

3:33

a half billion is in club credits.

3:35

And so there is a lot of

3:37

discussion about what does that mean to

3:39

the industry. What does that mean in

3:41

terms. Of valuation these companies and

3:43

and valuation of the companies mostly

3:45

is only relevant to the company

3:48

themselves into feces is not necessarily

3:50

relevant to you and I have

3:52

the headlines. but they

3:54

did sort of you know kind of then

3:56

get into this question of he is it

3:58

a good thing for the cloud providers

4:00

to be giving away cloud credits.

4:02

And then as those credits are

4:05

renewed, claiming those credits as both

4:07

an investment, but also more importantly,

4:09

a set of revenue,

4:11

right? So a new set of revenue. So

4:13

that was kind of the beginning of me thinking

4:15

about this stuff. And then, you know, a whole

4:17

bunch of other things happened around cloud investments and

4:20

people looking for money around GPUs and around AI.

4:22

And so it kind of made me think about,

4:24

okay, how do we connect the dots on some

4:26

of these things? Because it does feel like as

4:29

much as everybody's talking about AI, and it

4:31

feels like it's overwhelming the conversation in

4:33

our industry right now, which it is

4:35

for good reasons and other

4:38

reasons. It almost

4:40

feels like the funding and

4:42

the cost of AI and

4:44

sort of some of the interesting dynamics happening

4:46

around it are maybe not getting as much play

4:48

as they should. So I thought what I would

4:51

do in the second part of the show is

4:53

sort of dig into, you know, kind of pull

4:55

at a bunch of threads that are starting to

4:57

emerge and starting to come together and

4:59

ultimately, you know, what feels like

5:01

the beginning of a bunch

5:04

of kind of rules and frameworks kind

5:06

of getting rejiggered, if you will, re-framed,

5:08

restructured, you know, whatever the right word

5:11

is. I kind of want to dig into that

5:13

after the break. Transform your business with

5:15

Microsoft Azure, the one-stop shop for everything

5:17

cloud and AI. With With Azure

5:19

Migrate and Modernize, move your existing workloads

5:21

to Azure with ease. With With Azure

5:24

Innovate, build new solutions with focused and

5:26

deep engagement. Their Their core offerings

5:28

provide comprehensive resources, expert

5:30

assistance, and cost optimization tools. Accelerate

5:33

your cloud journey and stay ahead of the

5:35

competition with Microsoft Azure. Visit aka.ms

5:38

slash Azure Hero Offerings

5:40

to learn more. That's

5:43

aka.ms slash Azure Hero

5:45

Offerings to learn more.

5:50

And we're back. And as I mentioned at

5:52

the top of the show, we want to dig into

5:54

kind of what feels like some very interesting different

5:57

threads, different data points that are kind of popping

5:59

up. And what

6:01

feels like the beginning of maybe some

6:03

remaking of where's money

6:05

coming from in our industry as it relates

6:07

to AI? How

6:10

is it being used and spent

6:12

in terms of model building versus

6:14

company building and so forth? And

6:17

what are some of the kind of

6:20

limiting factors or things

6:22

that are starting to emerge that

6:24

may have very, very large term

6:26

ramifications, right? So I'm going

6:29

to throw out a couple of data points or

6:31

a couple of kind of instances and we'll try

6:33

and connect the dots between them. And I think

6:35

quite honestly, this may be the first of what

6:39

will probably be a whole bunch of conversations that

6:42

we'll want to have around the way that

6:44

AI is being funded

6:47

in essence throughout the industry. Because ultimately,

6:49

one of the things that we've done

6:51

over the last 10 plus

6:54

years on this show is we've

6:56

really not only tried to highlight the

6:58

technologies that are emerging in cloud computing,

7:02

but just as importantly trying to paint

7:05

a pretty good picture or get a very

7:07

good understanding of where's the money coming from

7:09

in this industry. And I know for a lot of people talking

7:13

about money and revenue and profits and all

7:15

that stuff kind of turns them off and

7:17

that's fine, right? If you're kind of turning

7:19

to the show purely as a technologist, that's

7:21

fine. Aaron

7:23

and I have always sort of guided or

7:25

said, hey, as

7:28

much as you are trying to wrap

7:30

your head around technology, apply technology, learn your

7:32

skills, it does you at

7:34

least a certain amount of good to understand

7:36

the economics of your industry. Because if you

7:38

don't understand at least the basics of the

7:40

economics around the industry, especially

7:42

at that time, you potentially

7:45

leave yourself vulnerable to situations in which

7:47

you fall in love with the technology.

7:49

You believe the technology is going to

7:51

be really impactful for whatever

7:54

you do. And if you don't

7:56

understand the economics of it, you can find yourself

7:58

very quickly kind of getting line-sided by

8:01

this love of the technology and not

8:03

understanding that the technology and

8:06

the economics don't line themselves up and that

8:08

one or the other may quickly

8:10

disappear or crash or not behave in the

8:12

way that you expected. So

8:14

anyways, I use that as caveat emptor, beware. And

8:20

anyway, so here's the things that I've

8:22

been noodling around in my brain. So

8:26

we have this huge amount of investment by

8:28

the three large cloud providers and probably more

8:30

so other cloud providers as well

8:33

in not just GPUs. We've all kind

8:36

of seen the numbers in terms of

8:38

how many GPUs that the

8:40

various cloud providers are buying. We've

8:43

got Facebook talking about having a 600,000 GPU

8:46

farm and the cloud providers having four,

8:49

five, 600,000 GPUs and making

8:52

these massive, massive investments. Obviously right now,

8:55

Nvidia is the biggest beneficiary

8:57

of all of that. We expect that'll evolve

8:59

over time. We started to see some

9:02

of the cloud providers talk about even building their own

9:04

GPUs. Azure is already talking about building

9:06

their own GPUs and so forth. But

9:08

we've got this huge investment by the

9:10

cloud providers, not only in GPUs,

9:13

but also because for

9:15

the most part, they don't necessarily

9:17

build their own models. And that's not completely

9:21

true. Google obviously has been in

9:23

the model building business for a long time, probably more

9:25

so than anybody else. Microsoft

9:27

is investing in open AI heavily, heavily. Amazon

9:31

has some of their own models, but they've

9:33

also kind of feels like they're,

9:35

they've sort of, I don't want to acknowledge, but

9:37

they sort of acknowledge they're a little bit behind.

9:40

So they're investing in people like Entropic and other stuff like that. So

9:43

we have both massive investment in GPUs,

9:47

mostly benefiting Nvidia.

9:50

We've got massive investment in

9:52

the models themselves. And

9:55

what's interesting is the investment in

9:57

GPUs obviously has to be driven.

10:00

by cash and cash reserves and probably

10:02

going through a certain amount of what

10:05

the CPU vendors went through during

10:08

the last 10 plus years where the cloud

10:10

providers were placing orders and they were rapidly

10:12

consuming them, not necessarily paying for

10:14

them right away and probably getting into different ways

10:17

of paying for them. It'll

10:21

be interesting to see how they're paying for the GPUs. Obviously,

10:24

Nvidia would probably like them to be all we

10:26

paid for upfront. We'll

10:28

see. The reason I mention that is

10:31

because we are starting to see

10:33

some reports come out from various places

10:36

that the GPUs, while they're available and

10:39

people are reserving them,

10:42

they're not necessarily being utilized

10:44

at high rates. To

10:47

a certain extent, this is probably to be

10:49

expected because a lot of this

10:51

stuff is still somewhat new. It's

10:56

interesting because, again, we would think of the GPUs

10:58

as being such

11:00

incredibly valuable resources

11:04

that people would try to utilize them to

11:06

the greatest extent. There are

11:08

a bunch of links in the show notes to various reports that

11:10

are starting to say, well, maybe the

11:12

GPUs aren't necessarily being as highly

11:14

utilized. The reason that's being

11:16

said is because if they were as utilized

11:18

as people would expect them to be, the

11:21

cloud revenues would be higher than they would be. Anyway,

11:23

another data point about the cloud revenues. Now,

11:26

what's interesting and what got me starting with

11:28

this was, as I mentioned, this

11:31

podcast is BG2 podcast, which was

11:34

a bunch of VCs bemoaning

11:36

the idea that they were being cut

11:38

out of the loop to a certain

11:40

extent in terms of investing in these

11:43

high-flying, high-gravity companies,

11:45

the Anthropics, the OpenAI, the

11:47

big model builders. The

11:51

way that things were being funded was serving

11:53

cloud credits from the cloud providers. Obviously,

11:57

that gets interesting because the cloud providers

12:01

being part of very, very large

12:03

companies, Google, Amazon, Microsoft, have

12:05

access to a certain extent to huge,

12:08

huge amounts of cash. They've been hoarding

12:10

cash for many, many years. They've got

12:12

huge cash reserves. But

12:15

instead, they are essentially making

12:18

these investments with cloud

12:20

credits. So I'm giving

12:22

you $4 billion, but essentially, $3.5 billion of

12:25

that, for example, is in you

12:27

directly utilizing my service. And

12:29

it's a really interesting way of going about

12:32

funding it because not only does

12:34

it allow them probably to do a

12:36

couple of things. Number one, they

12:39

get to take an equity position many

12:42

times in these companies. So they're investing

12:44

in the future. They're

12:46

getting near-term revenues back for

12:49

that pseudo investment. The money didn't really

12:51

actually leave them. And they

12:55

already have this investment, which

12:57

is depreciating, and it's helping them accelerate

12:59

the depreciation of it. So it

13:02

gets into some really interesting economics

13:04

and math and accounting and so forth as to what

13:07

do these investments look like? And

13:09

they look very different than the risk that

13:12

venture capitalists take in which they're

13:15

not necessarily getting their money back right away.

13:18

So that part was very interesting. I

13:20

think it has some interesting ramifications as we do

13:23

our normal cloud news the week, cloud news the month

13:25

stuff, or what will be cloud news of the quarter

13:27

as we look at the cloud revenues. We're

13:30

already starting to see the cloud

13:32

providers talk about the revenue bump

13:34

that they're getting from AI. But

13:37

it does, for example,

13:39

so Microsoft's Q2 2024 earnings report

13:41

states that AI services contributed

13:43

6 percentage points of growth to Azure revenue.

13:45

This is an increase from 3 percentage points

13:48

in the previous quarter. So this is a

13:50

quote from the recent Microsoft earnings. And

13:52

it does make you sort of wonder, when

13:55

we start talking about AI growth in the cloud,

13:57

are we going to be talking about... growth

14:01

from customers, so meaning like

14:04

not open AI training their models,

14:06

which technically is a customer of

14:08

Azure. But you know, it's

14:10

not JP Morgan, it's not Ford, it's not

14:12

Boeing, it's not, you know,

14:14

Barclays or, you know, HSBC or,

14:17

you know, anybody along those lines. So

14:19

it'd be interesting to start and watch

14:21

and see, do we see these

14:24

numbers popping out in the earnings

14:26

calls from cloud providers, but

14:28

yet maybe not clarification of who

14:30

exactly is utilizing all these

14:32

GPUs, who's doing all this quote unquote

14:35

AI, that's propping up the numbers

14:37

for the cloud providers. So I think that's gonna

14:39

be interesting sort of thing to watch is, you

14:41

know, will we be able

14:43

to figure out the granularity between

14:46

essentially the renewal of these investment

14:48

credits cranking out, trying

14:50

to build bigger models, faster models, and so

14:52

on and so forth, more capable models? Or

14:55

are they actually sort of

14:57

second order downstream, actual end

14:59

customers utilizing the AI services or,

15:01

you know, building applications that then touch these

15:03

models and so forth. So I think

15:06

that's gonna be really interesting to watch. The other

15:08

thing that's second

15:11

thing that's interesting is, you

15:13

know, what happens to VC investment, because

15:15

we all know, there

15:17

is huge pent up demand to

15:20

be involved in the AI space to be investing

15:22

in the AI space, right? Sort of, you know,

15:24

we've mentioned this many times, it's, it feels like

15:27

the next computing era, it feels like sort of

15:29

the next gold rush in computing. But,

15:32

you know, will, will we see the

15:35

VC community be able to participate at the level

15:37

that they want to? Will we start to see

15:40

the cloud providers, you know, sort of through

15:42

their VC arms, if you will, their own

15:44

internal VC arms, basically

15:46

saying, Hey, this is a great model.

15:49

Let's start cutting out the VCs,

15:51

right? Let's, let's go directly to

15:53

these technologies. Let's offer cloud credits.

15:56

Let's sort of build this flywheel in which,

15:58

you know, you're not

16:00

having to chase money, we're providing this thing, we're

16:02

able to utilize our resources

16:04

more, we're able to depreciate them. So

16:06

it does feel a little bit like

16:09

we might be at the beginning of

16:12

sort of a new VC framework.

16:16

And it'll be interesting to sort of dig into, you know,

16:18

what does that look like from a risk perspective? What

16:21

does that look like from a partnership

16:24

perspective with the

16:26

cloud providers? Because obviously, you know,

16:28

companies and startups have a certain relationship with VCs,

16:30

sometimes it's good, sometimes it's bad. What

16:33

does that look like in terms of, you

16:36

know, working with the cloud providers? Is that a good thing? Because

16:38

not only are they providing you

16:40

credits, is it a bad thing because

16:42

they're not providing you as much cash, so maybe you're more

16:45

constrained in terms of the number of people you can hire, you

16:48

know, or what you can pay your data scientists,

16:50

that'll be interesting to watch. Obviously,

16:54

the VCs don't necessarily have any go

16:56

to market path to help you with

16:58

that. The cloud providers, you know, obviously

17:00

have cloud, you know, partner go to

17:03

market paths and partner joint programs and

17:05

marketplaces and stuff. So that seems like

17:07

it would be sort of beneficial more

17:09

for the cloud provider than a traditional VC.

17:13

But I think that space is going to be really, really interesting to watch

17:15

is, you know, how

17:17

much investment do we see from

17:19

them? And then how do, you know,

17:21

how do the cloud providers, if they go down this path, how

17:24

do they manage sort of the conflicts of

17:26

interest that might arise or the sort of

17:29

prioritization that might arise from

17:32

getting access to these GPUs, their

17:34

go to market paths and all those sort of things. So

17:36

that becomes sort of interesting. Now,

17:39

the third thing that I thought was sort of interesting, and

17:41

this, this again popped up here in this last week, and

17:44

this is one of those articles that you read

17:46

and you think like, seems

17:48

a little bit crazy. You

17:51

know, and so your red flags sort of go up your

17:53

Spidey sense sort of goes up. But

17:55

there's an article out there floating around that

17:57

some Sam Altman, who is at least wearing

18:00

one hat, the CEO

18:02

of OpenAI. We

18:05

talked about Sam for the year

18:07

end and he had a little bit of bumpiness at OpenAI. It

18:09

seems to be back at OpenAI. And

18:11

the articles that are being written

18:13

and the things that are floating around is that

18:16

he is seeking up to $7 trillion

18:19

– T trillion – $7

18:21

trillion for this sort of new

18:23

chip concept of

18:25

whatever he's trying to do. So he's

18:27

basically saying, look, in order for

18:30

us to achieve these massive, massive AI goals,

18:32

we're going to need far greater GPU

18:35

capacity than the world provides

18:37

today and is out

18:40

going to all sorts of investment vehicles

18:43

from nation states and sovereign wealth funds

18:45

to VCs and so forth and

18:47

allegedly is looking for up to $7 trillion.

18:49

And the reason

18:52

I bring this up is it's interesting sort of

18:54

headline fodder and it's a gigantic number. I

18:57

mean, it's just an insane number. There's

18:59

actually a tweet that's floating around that I put

19:01

a link to in the show notes that

19:04

sort of says, these are the

19:06

companies that if

19:09

this amount of money was

19:11

raised, who they could buy?

19:13

NVIDIA, TSMC, Broadcom, ASML, Samsung,

19:15

AMD, Intel, Qualcomm, Applied Materials,

19:17

Texas Instruments, Lamb Research,

19:19

ARM, Anilong Crisis, Micron, KLA, Foxconn,

19:21

Marvell, and MediaTake, all these companies

19:23

in the chip sector and you'd

19:26

still have $1.5 trillion

19:28

left over. So buying up

19:30

all those companies, essentially the

19:33

entire chip market is

19:35

about $5.5 billion or $5. something billion and he's

19:40

kicking around these ideas of trying to raise nearly

19:43

$7 trillion. So

19:46

anyway, the reason this was sort of interesting is

19:49

not from the sort of eye candy

19:51

or clickbait-y-ness of how big it is or

19:54

the title, but it does make you

19:56

wonder, the

19:58

last time we've seen these

20:00

kind of crazy, crazy ambitions

20:02

about raising insane amounts of money

20:05

were, you know, in the

20:07

sort of we work or Uber sort

20:10

of things where somebody who is

20:13

believed to be sort of the, you know,

20:15

this visionary of this segment of the industry,

20:18

essentially saying, there

20:20

are no rules in this new game that I'm going

20:22

to create, there are no laws, there are no national

20:25

priorities, there are no nothing, you

20:29

know, but I'm going to need so much money to go about

20:31

doing this. And it's interesting

20:33

in that so much of this money is

20:36

being raised, you know, is

20:38

it being raised aligned to open AI?

20:40

Is it being aligned to something outside

20:42

of open AI such that Sam is,

20:45

you know, kind of trying to get outside

20:47

of that vacuum and saying, hey, I don't

20:49

necessarily want to own the greatest model, I

20:51

want to own the chipsets, you know, I

20:53

want to own the pickaxes and the shovels.

20:57

So anyways, I thought that was interesting as

20:59

well. And so, you know, when

21:01

we put all those sort of three things together in

21:03

terms of the VC industry feels

21:05

like it's being disrupted a little bit in terms

21:08

of sort of their traditional model, the

21:10

cloud providers are starting

21:13

to try and control their own destiny, but

21:15

they're doing it in sort of interesting, novel

21:18

and maybe maybe not good financial

21:20

ways, we'll see. And then, you

21:23

know, you have some overblown

21:26

kind of ambitions, you

21:28

know, grandiose, grandiose, grandiose ambitions

21:31

to completely change the market. And

21:34

the reason that one is interesting is because, you

21:37

know, what Uber

21:39

and eventually Lyft and so forth did to,

21:42

you know, things like ride sharing, like,

21:44

is there a correlation there between that

21:46

and this gigantic ask

21:49

of, you know, funding for something that's,

21:51

you know, sort of chips, essentially infrastructure

21:55

or, you know, huge, huge swaths of

21:57

what could become AI? Like, is it

21:59

too much? and so forth. So anyways,

22:02

I just thought those three things were sort of

22:04

interesting. They do feel like, you know, people trying

22:06

to look at the big map of things and

22:08

going, I don't know if I like the map

22:11

anymore. I don't know if I like the game,

22:13

I have these ambitions for what

22:15

the world could be. And in order to get

22:17

there, we're essentially going to have to

22:20

spend more time with the eraser than we will with

22:22

the with the pen or the Sharpie erasing

22:24

sort of the lines of the old games,

22:27

the old expectations, the old models, and

22:29

basically redrawing them with something new. So

22:32

anyways, I think it's been very interesting to sort

22:34

of start to think about how

22:36

this all comes together. Obviously, on this

22:38

show, we will do a bunch to do our best

22:41

to to cover the technology as we've always done. But

22:43

I think we will, you know, do

22:46

just as much in terms of

22:48

trying to make sure we understand the economics and get

22:50

ahead of it. Because I think, you know,

22:52

we're beginning to watch AI blossom.

22:56

You know, we kind of fully believe in the in

22:59

the thesis that this is the beginning of the

23:01

AI era. But again, every

23:03

era is is dictated, both

23:05

by technology, but also by economics

23:08

and debt and interest rates and all those sort of things.

23:10

And so it'll be very interesting for us to kind

23:12

of watch all that unfold. So anyways,

23:15

just not necessarily have a final

23:17

conclusion for today's show, but just some things that

23:19

we're really thinking about from a person, you know,

23:21

Sunday perspective, as opposed to a Sunday conclusion. So

23:23

anyways, thank you all for listening. Hope your

23:25

February is going well. Hope whoever you are

23:28

rooting for in the Super Bowl, whether it's

23:30

the football game or Taylor Swift, or something

23:32

else goes well for you. And hopefully, your

23:34

February is going well as the the days begin

23:36

to be a little bit longer in terms of daylight, start

23:38

to warm up a little bit. And

23:41

hopefully everybody's starting to get outside a little bit. So

23:43

anyways, with that, we'll wrap it up. Thanks

23:45

for listening. Thanks for telling a friend. Thanks for

23:47

clicking on the subscribe button. Thanks for, you know,

23:49

downloading the show and listening to it regularly. So

23:51

it stays up to date in the charts and

23:53

continues to get downloaded for you. So with that,

23:55

we'll wrap it up. We'll talk to you next

23:57

week. Thank you for listening to the cloud. Please

24:00

visit thecloudcast.net I

24:03

will show show notes, videos and

24:05

everything social media