January 15, 2025 • 23 mins
Akamai’s journey to regain control of their cloud spending reveals the pitfalls of over-provisioning and the importance of optimized workloads. The episode navigates through significant updates like OpenAI’s transition to a Public Benefit Corporation and Microsoft’s substantial investments in AI data centers while pondering Google Cloud's rising competitive positioning against Azure.
• Akamai struggles with escalating cloud costs
• The ramifications of leveraging cloud as a crutch for poor workload design
• OpenAI’s shift to a Public Benefit Corporation for societal good
• Microsoft's ambitious $80 billion investment in AI infrastructure
• Analysts predict Google Cloud could surpass Azure by 2025
• The rising role of AI agents in tech and IT management
Purchase Chris and Tim's new book on AWS Cloud Networking: https://www.amazon.com/Certified-Advanced-Networking-Certification-certification/dp/1835080839/
Check out the Fortnightly Cloud Networking News
https://docs.google.com/document/d/1fkBWCGwXDUX9OfZ9_MvSVup8tJJzJeqrauaE6VPT2b0/
Visit our website and subscribe: https://www.cables2clouds.com/
Follow us on BlueSky: https://bsky.app/profile/cables2clouds.com
Follow us on YouTube: https://www.youtube.com/@cables2clouds/
Follow us on TikTok: https://www.tiktok.com/@cables2clouds
Merch Store: https://store.cables2clouds.com/
Join the Discord Study group: https://artofneteng.com/iaatj
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Tim (00:01):
Welcome to the Cables to Clouds podcast, your one-stop
shop for all things hybrid andmulti-cloud networking.
Now here are your hosts, tim,chris and Alex.
Hello, and welcome back toanother episode of the Cables to
Clouds podcast Fortnightly News.
I'm your host this week, tim,and of course, with me always
(00:23):
I've got my hetero life mate,chris Miles, at BGP Main on Blue
Sky, or no, yeah, you changedit.
Bgp Main yeah, I changed it.
Yeah, that's right.
And then, of course, myself atCarpe-DMVPN on Blue Sky as well.
So, this being a news episode,we're just going to launch right
into the interesting news thatwe rounded up from the last
(00:44):
couple of weeks.
And our first story actuallycomes to us from CIO Dive and
it's an article from, or about,should I say, akamai
Technologies, and specificallyit has to do with Akamai
realizing their cloud spend wascrazy out of control and how
they kind of regained control oftheir runaway spending.
(01:06):
So the article specificallypoints out that their
acquisition of Linode last yearwas a big trigger for them to
kind of realize that they lettheir cloud spending just get
completely out of control.
So I mean, we see this a lot,where a lot of organizations go
all in on cloud.
They start developing for it.
They just put pretty mucheverything they throw the
kitchen sink at the cloud andthen eventually the cost of
(01:30):
running that comes back to hauntthem.
So we've talked about this onthe podcast many, many times
about cloud spend and where arethe cloud savings spend and
where are the cloud savings?
So this is an interesting lookat just a major vendor who kind
(01:51):
of bought into the cloud thingand is finding that after they
acquired Linode that they justspent a ton of money in the
cloud.
It kind of triggered them to goback and look at the books and
in the article they're talkingabout moving some of the
repatriating some of thoseworkloads essentially to their
their private cloud workloadthat that they bought.
Right, linode basically is aprivate cloud uh provider for
(02:12):
those that don't know, and uh,yeah, it's pretty interesting.
Of course, not everybody hastheir own private cloud provider
.
Um, you know, I don't.
There's not much more to sayabout about this one.
It was just, it was just kindof interesting.
So, uh, you took a look at thisone too.
Uh, chris, you have anything toadd there?
Chris (02:27):
Yeah not much.
I mean, yeah, it's just, it'slike obviously funny to see that
someone was trying to controltheir cloud spend and
repatriation was a part of thepiece of it.
Um, so we've we've obviouslyhad many talks on this podcast
about repatriation and whetheror not it's.
I mean, yeah, people alwaystalk about repatriation.
They're like, oh, that meansthey're moving everything and
it's like, no, it's, they'removing certain things that are,
(02:49):
you know, a crucial piece of thebusiness.
That may not mean need to runin the cloud, right, so that's,
it's good to see that coming tofruition.
But obviously it's Akamai,right, they, they pretty much
have.
I mean, they do, they do CDNtype stuff as well.
Right, like there's a lot of um, yeah, so it's like obviously
they, they need such a wide,they need to cast such a wide
(03:09):
net across the globe for to dostuff like that.
They need the cloud in order todo that.
Um, so it's just good to, Iguess, kind of get proven right
in some aspects on this one.
So that's good to see.
Tim (03:20):
I forgot to mention it, but there's a gold, just a huge
nugget of gold, there in thearticle.
It's near the bottom and it's aquote, basically from the
person at Akamai.
I forgot the name, but I justwanted to bring it up because
this is such a true statement.
Specifically, it says we wereover-provisioned because we were
(03:46):
leveraging cloud to compensatefor deficiencies in the way our
workloads were developed andengineered Right.
And if that isn't a truestatement, you know any
networking person is well awareof the over provisioning and
gold plating of networks to kindof compensate for poor
application design and it's goodto see somebody come out and
just say like that and so andwe've said that also many times
(04:10):
on the podcast that like hey, ifyou take these workloads, these
unoptimized workloads right,that have no tolerance and that
are chewing up processes, andjust put a middleman in there by
throwing it in the cloud, likeof course, you're not going to
save money, it's going to ruinyou.
Yeah.
Chris (04:25):
Garbage in, garbage out, all right.
Next up we have this one, Ithink, just squeezed into our
news cycle, right?
I think we actually debated oncovering this on the last
episode but with the way it fellwe pushed it to this one.
So there's been an announcementfrom OpenAI that they're
basically doing anotherrestructure to the organization
(04:48):
and they're going to turn thepublic arm, uh, of the company
Uh, so if you, if you recallthey, you know kind of like a
split these things across.
Where there was a, it startedholistically as a nonprofit and
then there was a for-profit armand now they're kind of
operating as two entities andwe've obviously been a bit
cynical about how we think thosethings operate on this podcast
(05:11):
before.
But now they're changing theprofit arm to be what is called
a PBC or a public benefitcorporation.
And basically, I'll be honestwith you, I had not heard of
what a PPC was before I saw thisarticle.
So I did a little bit ofresearch into it and basically
just sounds like thisdifferentiates from a typical
(05:32):
public company or a Ccorporation, kind of like, just
in terms of where theorientation of the mission for
that company is like.
It's like legally obligated toum file, follow um, you know
kind of whatever their, theirmission statement is, um and
balance their financialprofitability with that specific
(05:54):
purpose.
Um don't know how that'sgoverned, but that's that's kind
of what I've uh taken from it.
And you know, there's obviouslygreater accountability and
flexibility in decision making.
It sounds like Like they willbe able to prioritize like kind
of societal or environmentalgoals over you know, direct
profits for shareholders andthings like that.
(06:16):
I'll be honest with you, man,when I saw this, I was naturally
cynical about it again, becausewe've've um.
You know, we've seen what'shappening with open ai and and
there's been a lot of kerfuffles.
Obviously sam altman's also inhot water right now for um some
potential non-technical,non-technical related uh aspects
(06:36):
of his life.
But I'll be honest with you,man, I'm actually struggling to
see what is bad about this.
This actually seems like itcould be a good thing, Assuming,
like I said before, though,with the PBCs and things like
that, I don't know how that'sgoverned, I don't know how
that's enforced, so I'm surethere's loopholes and things
(06:56):
like that that I may not beaware of, but overall this does
seem like a good thing.
It seems like they're going tothe way they they're framing it
is that this aligns to theirmission to reach the AGI right,
the artificial generalintelligence for for the public,
um, which is seen as their, Iguess their societal good
contribute, uh, uh, contributionto the to the world Right, um,
(07:20):
but yeah, I don't know.
Um, how do you feel, tim, doyou?
I'm?
Tim (07:28):
sure you've got some uh, uh , hopefully a hot take where I
don't, no, um.
So this is interesting though.
So, the public benefitscorporation uh, we talked about
this a little bit last year whenwe had Joe on a sick on, if you
remember.
We were talking about the techlayoffs and like what, basically
what people can do, and likewhat the answer is.
And he had had, he had beentalking about the, the public
benefits corporations.
Uh, I forget it, it's b, Iforget what, the, the, the
(07:51):
specific uh 802.1c or 802.1b orsomething like that.
I'm not thinking of likewireless, but uh, the, um, but,
but whatever you call it,there's a corporation, a, a
filing, that is Benefits Corp,and I have the same question,
which is who determines whetherit's like what is the public
good and all of that in there,and maybe it's the board.
(08:13):
But what's interesting aboutwhat I found interesting back
then, and what I still findinteresting here, is that we
have this legal vehicle by whichthe corporation can essentially
become unsuitable by investors.
You know cause like the.
You know part of the reason whycapitalism is just, in my
opinion, gone off the, you knowthe, driven straight off of the
(08:35):
cliff, is this, this thing wherecorporations can be sued if
they don't do everythingpossible to maximize their
shareholder value.
They can be sued byshareholders for not doing
everything possible, whatever,whether that could be good, bad
or ugly or whatever that lookslike.
Anything possible to maximizeshare value is becomes doable
like you can.
The shareholder can sue themand get money, basically.
(08:56):
So I think that this the bitwith the benefits corporation is
is essentially structured legalprotection for that,
essentially.
So you know, if you invest inour corporation, you understand
that you can't sue us if we dosomething that's, you know, for
the public good, which is whatkind of world do we live in,
where that's the thing we haveto do, but whatever.
Chris (09:18):
That's a good point because, like, obviously money
and shareholders, that's atangible asset.
Right, you can immediately, youcan immediately say if, if that
was met or not met.
Um, whereas I guess with a PBClike, I don't know how you, I
don't know how you measure, thatyeah Right.
Yeah.
So it's probably going to bemuch easier, or much harder, I
should say, to build a caseagainst, like you saying like oh
(09:40):
, they did not fulfill theirduty as a corp towards goal A, b
or C, right, yeah, see, that'swhy we have you here, man, give
me a check.
Tim (09:51):
I'm glad to help.
Okay, let's see, let's move onto the next one here.
So our next article comes fromoh hold on, I just got to, so
sorry.
Our next article comes from CIODive, actually, and this one is
talking about Microsoft's plansthis year to pour 80 billion
dollars into their cloud datacenters.
(10:12):
Holy crap, yeah, basically tobuild out of all the framework
and stuff for AI workloads.
So I mean it doesn'tspecifically say, hey, this is
like we're going to build Azure,all the Azure regions.
It doesn't go that far andexplain exactly what these $80
billion is earmarked for.
But I mean, if we're assumingthat their plans to upgrade and
(10:37):
increase the amount of AIworkloads you know that are
available, you know, part ofthat probably is just, I don't
know if they're going to behanding over tech to open AI I
guess part of this or if they'rejust going to, like, build out
more Azure infrastructure tomake, you know, ai services
available for customers, orlittle column, a little column B
, it's not clear from thearticle, but I mean $80 billion
(10:59):
is well, it's just a shitload ofmoney.
I mean it's just, it's anabsolutely mind boggling amount
of money.
So this is interesting becauseup until now they've pretty much
had, like he was saying beforewe record, chris, that OpenAI
has been like their poster child, like their baby.
So I don't know if thisindicates a doubling down on
(11:21):
that and I think we're going tohave to wait and see or if it
indicates that, hey, they'rediversifying, or maybe just
straight up, that they're addingthe ability for customers to
consume and hoping that thatwill bring the customers right.
Chris (11:33):
Yeah, my initial thought about this is that it's probably
the latter, Because itspecifically calls out them
putting $80 billion into thepurpose of training and
developing AI within these datacenters, and to me that seems
like I mean, obviously you couldsay like oh, we're just giving
infrastructure to open AI toperform that, but if they're
calling out specifically to dotraining and things like that,
(11:53):
it sounds like they're trying tocome up either with something
on their own or at least youknow, kind of maybe taking a bit
out of AWS's book and sayinglike hey, we're just going to
have a kind of a market for thiskind of stuff, or maybe they're
developing their own thing.
I don't know.
But I mean, like you said,they've always been obviously
keeping OpenAI a little bit atarm's length.
(12:14):
They haven't really completelyembraced them, I think, just
because probably Ndela doesn'twant to pull in Sam Altman just
yet, just by way of there'salways been controversy and
things like that.
Tim (12:27):
A little bit of a hot potato.
Chris (12:29):
Yeah, for sure.
So I don't know, we'll see, butthat's my prediction anyway.
Tim (12:35):
Yeah, I think that's pretty good, I mean truthfully, we're
just gonna have to wait and see,but it does point out training.
So yeah, is Microsoft going tocreate their own model, I mean,
and if so is, does it take 80billion dollars to create your
own model these days?
Chris (12:49):
Like that's a good point.
Tim (12:50):
That's a good point.
Chris (12:51):
The person who puts the most money in wins, wins.
Yeah, absolutely All right.
Next up, we have an articlehere from Fierce Network.
I feel like every time we havesomething on here from this
publication, it's always like adrastically hot take Like a hot
take, yeah, yeah.
This one is of no exception,right?
So they have an article herethat in 2025, it looks like an
(13:15):
analyst of some sort, I forgetwho exactly the name was is
predicting that Google Cloud mayovertake Azure Microsoft Azure
within 2025 in terms of themarket cap for public cloud,
which I think is I would nothave said that.
We've been on here talking fora couple of years now.
(13:37):
We're like, oh, microsoft mighttake over as number one soon,
let alone getting you know, putin the third spot.
I mean, they specifically callout the fact that Google does
very well in the SMB space,which is totally, totally valid.
I see that we just had Erica onhere and you know she was
(13:58):
talking about working at astartup and you know they went
into Google cause, um, you know,people like Google and it was
probably easy to get started andthey do a lot of credits and
things like that, so it makesmakes perfect sense.
Um, but like I think the theymentioned that the the market
saturation was like AWS was like31 percent, azure 20 percent,
google 13 percent, which, I'llbe honest, that sounds higher
(14:20):
than I thought for Google, ifI'm being completely transparent
, but nonetheless, yeah, likeyou know, they call out
specifically that Microsoft doesvery well with enterprises and
is working in arenas with likethings like ibm and sap and
things like that.
Um, but google is you know whatwe said last, uh, last news
(14:40):
episode they're making this bigpush about.
Tim (14:42):
You know, gemini is the main focus for 2029 or 2029,
2025, I mean, maybe it'll be2029 as well but I cannot
foresee that happening within a12 month span for them.
That's a hell of a growthperiod, yeah.
In 12 months.
That's a very large month overmonth.
Chris (15:04):
Unless there's a tank sprout that, like it's, one's
got to go up and that one'sgoing to go way down Right.
Tim (15:09):
Yeah Right, I mean, there's not an infinite amount of
market share, like I thinkthey're definitely jockeying
there.
So, yeah, it's a.
I mean, it's a radioactive take.
I don't know what to say to it.
I I do, having read the article.
The article gets into thebasically why we think this and
they actually have at least a, athesis.
It's not just a let's, let'sget clicks by saying weird shit
(15:31):
like they do nothing right.
I don't know if I'm still likeI'm with you, like I don't know
that I agree, still like I'mwith you, like I don't know that
I agree, but you know, I mean,the truth is we don't ever know.
One thing this thing has provenis that we truly don't know.
Like you said, we thought Azure, by this point, might be like
really really dogging AWS'sheels, and I think the AI thing
(15:53):
has definitely really shaken thewhole damn market up and yeah,
so that's a big one.
But yeah, I didn't thinkGoogle's share was, I would not
have thought 13% either, and ofcourse, you do the math on this
right.
So you have what?
31% for Google, 20% for Azure,less for, uh, azure and what was
(16:17):
it 13?
yeah, exactly.
So we're saying are we seeingbasically, the?
The other side is other.
So how much does oci have then?
How much is?
Is all the other private andpublic cloud, you know, startup
or ibm or whatever we're callingpublic cloud now, alibaba and
all those right.
So that's, that seems like apretty yeah.
Anyway, yeah, not much to say,I guess.
(16:39):
Well, we will see if Googlereally does catch up.
I will point out that, ofcourse, we've struggled with
this too.
All the analysts do isdecoupling Microsoft's cloud
Azure offering from 365 and allof their other stuff that's
hosted on Azure.
That's always been an extremelyhard thing to decouple, which I
(17:01):
think they actually use.
They point out in the articleas one of the reasons why
Microsoft has always seemed sofar ahead, but might not really
be so yeah we're going to say,yeah, it's weird.
Chris (17:12):
It's funny how many customers I talk to on a daily
basis, or you know, I should sayprobably prospects, and I just
you know kind of line ofquestioning like hey, what
clouds are you in?
What are you typically using?
They're like oh well, a bit ofAzure, because you know we use a
lot of O365.
And I'm like even the responseto that question is like in that
(17:33):
camp, I'm like that's not, Inot.
I mean, technically, I guessyou're probably using azure, but
I wouldn't say that's, that'sjust pure sass right, like
that's right.
I'm consuming services that arehosted in azure, but I'm not
like an azure customer right soit's like, but that's I mean but
that that could also speak towhy microsoft has such success,
that there's cohesion betweenthat like, even the mindset, is
(17:53):
that like if I, using O365, Iconsider myself an Azure
customer, and therefore, it'seasier for me to get started in
Azure, because I'm quote-unquotealready there, even though
you're not really right.
Tim (18:03):
But yeah, I get that.
That's interesting.
Okay, and we'll close out thisweek with kind of a funny one.
Honestly, I think it'sabsolutely ridiculous.
So NVIDIA had a conferencerecently, uh, where uh jensen
huang of uh nvidia basicallysaid that he did, he did a bunch
of stuff.
(18:24):
That was hilarious.
I don't know if you saw it.
There's a video where he comesout with a giant uh uh nvidia
gpu on like what looks like ashield.
Basically, no, I did not didnot see that.
Oh, you got to find it Because Iwas joking with somebody.
I was like it's way too flashy,You're not going to be able to
deflect any and it looks weak.
You can't deflect any.
Sword strikes or maces withthat thing or anything.
(18:45):
It looks like a buckler, like abig old shield.
Anyway, so in this article fromAOLcom of all places, so I you
know, if you right, I mean justgo get one of the CDs.
If you want to read this, go getone of those free CDs the
thousand hours so you can readthe article, so anyway.
(19:06):
So Jensen Huang says basically,that AI agents are becoming a
thing and, in this new era of AIagents, that IT service
organizations will become thequote, unquote, hr of AI agents.
So yeah, a lot of heavy liftingwith a lot of words there.
So the idea, of course, is thatthese agents are essentially
(19:36):
not quite autonomous but, like,purpose built to perform
whatever task the agent has beencreated for, and that,
therefore, the IT would becomethe shepherd, I guess, of these
agents.
And I don't know, I feel likeHR is a bit of a, I don't know.
I don't want to say hot, take,it's a bit of a stretch or
something, because I mean, can Iput the AI agent on a pip if
it's not doing well, you know,can we let it go?
Chris (19:57):
or, you know, I don't know, give it a pay cut.
This just has me feeling sorryfor all the people that are
specifically individualcontributors because they don't
want to be a manager.
It's like, well, well,according to jensen, eventually
you will be a fucking manager.
(20:18):
Yep over something that could.
It won't talk back to you, Iguess.
So, um, you know, maybe, maybeyou get.
You get the, you know, thatpower complex that you've always
wanted.
Tim (20:22):
You can yeah you really, uh , you know, push down hard on
your, on your, uh, your workersor you can build an ai agent
whose, uh, whose job is to agreewith everything that you say.
Perhaps that's you know.
Give it a prompt, load up theagent with a prompt that says
it'll uh, yeah, like you're mymanager and I love you and I'll
do anything you say who knows?
(20:42):
Uh, I don't know what to say tothis.
It's an absolute.
It's kind of a ridiculousarticle.
I mean, I get the feeling likeit was made for clicks, but not
the article, sorry.
Chris (20:54):
The wording or whatever, if you will, the quotation.
Tim (20:57):
It was meant to turn heads and get headlines.
But I mean, ai agents is athing.
It's the idea of we're kind ofpreloading prompts and then kind
of packaging up, almost like aDocker container, an AI agent
that will, you know, follow theprompt and and and carry out the
task for which it has beenprepared.
But uh, yeah, I don't, I stilldon't know if it becomes the hr
(21:21):
of ai agents or just kind oflike the you know, life cycle
manager of an ai agent.
Chris (21:26):
It's probably more accurate yeah, I mean like you
said.
I mean the.
He can have the quote asridiculous as he wants, but
really what it means is justthat you're changing how you
interface with their technology,right?
So it's like, sure, maybe Imean 2025, I have actually heard
multiple times that 2025 willbe the year of the AI agents.
Tim (21:47):
Yeah, so that does sound like a thing.
2025 will be the year of the AIagents.
Chris (21:48):
Yeah, so that does sound like a thing.
But you know, we potentiallyhave some talks coming up with
some guests, maybe about some ofthese agents, and we'll get
into that.
But yeah, I feel like thenumber one phrase we say on this
Fortnum Huntley news show iswe'll see.
This is no exception right, weshall see.
Tim (22:11):
Yeah, yeah, absolutely, it's a.
It's a new way to consume ai by, by kind of making it more
autonomous, and I think that'sthe goal and that's where he's
getting to.
Is that that you know we'regoing to allow ai to be more
autonomous and do more work onits own so it can secretly
create skynet under our notices?
But, uh, you know, yeah, 2025,2025, you're the AI agent.
You heard it here first.
Alright, and with that I guesswe can go ahead and wrap it up.
(22:36):
I hope everybody had a goodtime.
Of course, we always have somestories that we don't
necessarily cover on the news.
They're in the Fortnightly Newsdocument that we put in the
show notes of every episode.
We encourage you to take a lookat those stories.
There's just stuff that wedidn't have time or or need to
get into very, very deeply onthe show, uh, but we think are
(23:01):
interesting and that ourlisteners, viewers, whatever,
will want to take a look at.
So, uh, take a look at that.
And, of course, always, uh,like, subscribe, buy the
breakfast cereal, play the homegame and we'll see you next time
.
Peace, hi, everyone.
It's Tim, and this has been theCables to Clouds podcast.
Thanks for tuning in today.
(23:21):
If you enjoyed our show, pleasesubscribe to us in your
favorite podcast catcher as wellas subscribe and turn on
notifications for our YouTubechannel to be notified of all
our new episodes.
Follow us on socials atCables2Clouds.
You can also visit our websitefor all the show notes at
Cables2Cloudscom.
Thanks again for listening andsee you next time.
Welcome to the Cables to Clouds podcast, your one-stop
shop for all things hybrid andmulti-cloud networking.
Now here are your hosts, tim,chris and Alex.
Hello, and welcome back toanother episode of the Cables to
Clouds podcast Fortnightly News.
I'm your host this week, tim,and of course, with me always
(00:23):
I've got my hetero life mate,chris Miles, at BGP Main on Blue
Sky, or no, yeah, you changedit.
Bgp Main yeah, I changed it.
Yeah, that's right.
And then, of course, myself atCarpe-DMVPN on Blue Sky as well.
So, this being a news episode,we're just going to launch right
into the interesting news thatwe rounded up from the last
(00:44):
couple of weeks.
And our first story actuallycomes to us from CIO Dive and
it's an article from, or about,should I say, akamai
Technologies, and specificallyit has to do with Akamai
realizing their cloud spend wascrazy out of control and how
they kind of regained control oftheir runaway spending.
(01:06):
So the article specificallypoints out that their
acquisition of Linode last yearwas a big trigger for them to
kind of realize that they lettheir cloud spending just get
completely out of control.
So I mean, we see this a lot,where a lot of organizations go
all in on cloud.
They start developing for it.
They just put pretty mucheverything they throw the
kitchen sink at the cloud andthen eventually the cost of
(01:30):
running that comes back to hauntthem.
So we've talked about this onthe podcast many, many times
about cloud spend and where arethe cloud savings spend and
where are the cloud savings?
So this is an interesting lookat just a major vendor who kind
(01:51):
of bought into the cloud thingand is finding that after they
acquired Linode that they justspent a ton of money in the
cloud.
It kind of triggered them to goback and look at the books and
in the article they're talkingabout moving some of the
repatriating some of thoseworkloads essentially to their
their private cloud workloadthat that they bought.
Right, linode basically is aprivate cloud uh provider for
(02:12):
those that don't know, and uh,yeah, it's pretty interesting.
Of course, not everybody hastheir own private cloud provider
.
Um, you know, I don't.
There's not much more to sayabout about this one.
It was just, it was just kindof interesting.
So, uh, you took a look at thisone too.
Uh, chris, you have anything toadd there?
Chris (02:27):
Yeah not much.
I mean, yeah, it's just, it'slike obviously funny to see that
someone was trying to controltheir cloud spend and
repatriation was a part of thepiece of it.
Um, so we've we've obviouslyhad many talks on this podcast
about repatriation and whetheror not it's.
I mean, yeah, people alwaystalk about repatriation.
They're like, oh, that meansthey're moving everything and
it's like, no, it's, they'removing certain things that are,
(02:49):
you know, a crucial piece of thebusiness.
That may not mean need to runin the cloud, right, so that's,
it's good to see that coming tofruition.
But obviously it's Akamai,right, they, they pretty much
have.
I mean, they do, they do CDNtype stuff as well.
Right, like there's a lot of um, yeah, so it's like obviously
they, they need such a wide,they need to cast such a wide
(03:09):
net across the globe for to dostuff like that.
They need the cloud in order todo that.
Um, so it's just good to, Iguess, kind of get proven right
in some aspects on this one.
So that's good to see.
Tim (03:20):
I forgot to mention it, but there's a gold, just a huge
nugget of gold, there in thearticle.
It's near the bottom and it's aquote, basically from the
person at Akamai.
I forgot the name, but I justwanted to bring it up because
this is such a true statement.
Specifically, it says we wereover-provisioned because we were
(03:46):
leveraging cloud to compensatefor deficiencies in the way our
workloads were developed andengineered Right.
And if that isn't a truestatement, you know any
networking person is well awareof the over provisioning and
gold plating of networks to kindof compensate for poor
application design and it's goodto see somebody come out and
just say like that and so andwe've said that also many times
(04:10):
on the podcast that like hey, ifyou take these workloads, these
unoptimized workloads right,that have no tolerance and that
are chewing up processes, andjust put a middleman in there by
throwing it in the cloud, likeof course, you're not going to
save money, it's going to ruinyou.
Yeah.
Chris (04:25):
Garbage in, garbage out, all right.
Next up we have this one, Ithink, just squeezed into our
news cycle, right?
I think we actually debated oncovering this on the last
episode but with the way it fellwe pushed it to this one.
So there's been an announcementfrom OpenAI that they're
basically doing anotherrestructure to the organization
(04:48):
and they're going to turn thepublic arm, uh, of the company
Uh, so if you, if you recallthey, you know kind of like a
split these things across.
Where there was a, it startedholistically as a nonprofit and
then there was a for-profit armand now they're kind of
operating as two entities andwe've obviously been a bit
cynical about how we think thosethings operate on this podcast
(05:11):
before.
But now they're changing theprofit arm to be what is called
a PBC or a public benefitcorporation.
And basically, I'll be honestwith you, I had not heard of
what a PPC was before I saw thisarticle.
So I did a little bit ofresearch into it and basically
just sounds like thisdifferentiates from a typical
(05:32):
public company or a Ccorporation, kind of like, just
in terms of where theorientation of the mission for
that company is like.
It's like legally obligated toum file, follow um, you know
kind of whatever their, theirmission statement is, um and
balance their financialprofitability with that specific
(05:54):
purpose.
Um don't know how that'sgoverned, but that's that's kind
of what I've uh taken from it.
And you know, there's obviouslygreater accountability and
flexibility in decision making.
It sounds like Like they willbe able to prioritize like kind
of societal or environmentalgoals over you know, direct
profits for shareholders andthings like that.
(06:16):
I'll be honest with you, man,when I saw this, I was naturally
cynical about it again, becausewe've've um.
You know, we've seen what'shappening with open ai and and
there's been a lot of kerfuffles.
Obviously sam altman's also inhot water right now for um some
potential non-technical,non-technical related uh aspects
(06:36):
of his life.
But I'll be honest with you,man, I'm actually struggling to
see what is bad about this.
This actually seems like itcould be a good thing, Assuming,
like I said before, though,with the PBCs and things like
that, I don't know how that'sgoverned, I don't know how
that's enforced, so I'm surethere's loopholes and things
(06:56):
like that that I may not beaware of, but overall this does
seem like a good thing.
It seems like they're going tothe way they they're framing it
is that this aligns to theirmission to reach the AGI right,
the artificial generalintelligence for for the public,
um, which is seen as their, Iguess their societal good
contribute, uh, uh, contributionto the to the world Right, um,
(07:20):
but yeah, I don't know.
Um, how do you feel, tim, doyou?
I'm?
Tim (07:28):
sure you've got some uh, uh , hopefully a hot take where I
don't, no, um.
So this is interesting though.
So, the public benefitscorporation uh, we talked about
this a little bit last year whenwe had Joe on a sick on, if you
remember.
We were talking about the techlayoffs and like what, basically
what people can do, and likewhat the answer is.
And he had had, he had beentalking about the, the public
benefits corporations.
Uh, I forget it, it's b, Iforget what, the, the, the
(07:51):
specific uh 802.1c or 802.1b orsomething like that.
I'm not thinking of likewireless, but uh, the, um, but,
but whatever you call it,there's a corporation, a, a
filing, that is Benefits Corp,and I have the same question,
which is who determines whetherit's like what is the public
good and all of that in there,and maybe it's the board.
(08:13):
But what's interesting aboutwhat I found interesting back
then, and what I still findinteresting here, is that we
have this legal vehicle by whichthe corporation can essentially
become unsuitable by investors.
You know cause like the.
You know part of the reason whycapitalism is just, in my
opinion, gone off the, you knowthe, driven straight off of the
(08:35):
cliff, is this, this thing wherecorporations can be sued if
they don't do everythingpossible to maximize their
shareholder value.
They can be sued byshareholders for not doing
everything possible, whatever,whether that could be good, bad
or ugly or whatever that lookslike.
Anything possible to maximizeshare value is becomes doable
like you can.
The shareholder can sue themand get money, basically.
(08:56):
So I think that this the bitwith the benefits corporation is
is essentially structured legalprotection for that,
essentially.
So you know, if you invest inour corporation, you understand
that you can't sue us if we dosomething that's, you know, for
the public good, which is whatkind of world do we live in,
where that's the thing we haveto do, but whatever.
Chris (09:18):
That's a good point because, like, obviously money
and shareholders, that's atangible asset.
Right, you can immediately, youcan immediately say if, if that
was met or not met.
Um, whereas I guess with a PBClike, I don't know how you, I
don't know how you measure, thatyeah Right.
Yeah.
So it's probably going to bemuch easier, or much harder, I
should say, to build a caseagainst, like you saying like oh
(09:40):
, they did not fulfill theirduty as a corp towards goal A, b
or C, right, yeah, see, that'swhy we have you here, man, give
me a check.
Tim (09:51):
I'm glad to help.
Okay, let's see, let's move onto the next one here.
So our next article comes fromoh hold on, I just got to, so
sorry.
Our next article comes from CIODive, actually, and this one is
talking about Microsoft's plansthis year to pour 80 billion
dollars into their cloud datacenters.
(10:12):
Holy crap, yeah, basically tobuild out of all the framework
and stuff for AI workloads.
So I mean it doesn'tspecifically say, hey, this is
like we're going to build Azure,all the Azure regions.
It doesn't go that far andexplain exactly what these $80
billion is earmarked for.
But I mean, if we're assumingthat their plans to upgrade and
(10:37):
increase the amount of AIworkloads you know that are
available, you know, part ofthat probably is just, I don't
know if they're going to behanding over tech to open AI I
guess part of this or if they'rejust going to, like, build out
more Azure infrastructure tomake, you know, ai services
available for customers, orlittle column, a little column B
, it's not clear from thearticle, but I mean $80 billion
(10:59):
is well, it's just a shitload ofmoney.
I mean it's just, it's anabsolutely mind boggling amount
of money.
So this is interesting becauseup until now they've pretty much
had, like he was saying beforewe record, chris, that OpenAI
has been like their poster child, like their baby.
So I don't know if thisindicates a doubling down on
(11:21):
that and I think we're going tohave to wait and see or if it
indicates that, hey, they'rediversifying, or maybe just
straight up, that they're addingthe ability for customers to
consume and hoping that thatwill bring the customers right.
Chris (11:33):
Yeah, my initial thought about this is that it's probably
the latter, Because itspecifically calls out them
putting $80 billion into thepurpose of training and
developing AI within these datacenters, and to me that seems
like I mean, obviously you couldsay like oh, we're just giving
infrastructure to open AI toperform that, but if they're
calling out specifically to dotraining and things like that,
(11:53):
it sounds like they're trying tocome up either with something
on their own or at least youknow, kind of maybe taking a bit
out of AWS's book and sayinglike hey, we're just going to
have a kind of a market for thiskind of stuff, or maybe they're
developing their own thing.
I don't know.
But I mean, like you said,they've always been obviously
keeping OpenAI a little bit atarm's length.
(12:14):
They haven't really completelyembraced them, I think, just
because probably Ndela doesn'twant to pull in Sam Altman just
yet, just by way of there'salways been controversy and
things like that.
Tim (12:27):
A little bit of a hot potato.
Chris (12:29):
Yeah, for sure.
So I don't know, we'll see, butthat's my prediction anyway.
Tim (12:35):
Yeah, I think that's pretty good, I mean truthfully, we're
just gonna have to wait and see,but it does point out training.
So yeah, is Microsoft going tocreate their own model, I mean,
and if so is, does it take 80billion dollars to create your
own model these days?
Chris (12:49):
Like that's a good point.
Tim (12:50):
That's a good point.
Chris (12:51):
The person who puts the most money in wins, wins.
Yeah, absolutely All right.
Next up, we have an articlehere from Fierce Network.
I feel like every time we havesomething on here from this
publication, it's always like adrastically hot take Like a hot
take, yeah, yeah.
This one is of no exception,right?
So they have an article herethat in 2025, it looks like an
(13:15):
analyst of some sort, I forgetwho exactly the name was is
predicting that Google Cloud mayovertake Azure Microsoft Azure
within 2025 in terms of themarket cap for public cloud,
which I think is I would nothave said that.
We've been on here talking fora couple of years now.
(13:37):
We're like, oh, microsoft mighttake over as number one soon,
let alone getting you know, putin the third spot.
I mean, they specifically callout the fact that Google does
very well in the SMB space,which is totally, totally valid.
I see that we just had Erica onhere and you know she was
(13:58):
talking about working at astartup and you know they went
into Google cause, um, you know,people like Google and it was
probably easy to get started andthey do a lot of credits and
things like that, so it makesmakes perfect sense.
Um, but like I think the theymentioned that the the market
saturation was like AWS was like31 percent, azure 20 percent,
google 13 percent, which, I'llbe honest, that sounds higher
(14:20):
than I thought for Google, ifI'm being completely transparent
, but nonetheless, yeah, likeyou know, they call out
specifically that Microsoft doesvery well with enterprises and
is working in arenas with likethings like ibm and sap and
things like that.
Um, but google is you know whatwe said last, uh, last news
(14:40):
episode they're making this bigpush about.
Tim (14:42):
You know, gemini is the main focus for 2029 or 2029,
2025, I mean, maybe it'll be2029 as well but I cannot
foresee that happening within a12 month span for them.
That's a hell of a growthperiod, yeah.
In 12 months.
That's a very large month overmonth.
Chris (15:04):
Unless there's a tank sprout that, like it's, one's
got to go up and that one'sgoing to go way down Right.
Tim (15:09):
Yeah Right, I mean, there's not an infinite amount of
market share, like I thinkthey're definitely jockeying
there.
So, yeah, it's a.
I mean, it's a radioactive take.
I don't know what to say to it.
I I do, having read the article.
The article gets into thebasically why we think this and
they actually have at least a, athesis.
It's not just a let's, let'sget clicks by saying weird shit
(15:31):
like they do nothing right.
I don't know if I'm still likeI'm with you, like I don't know
that I agree, still like I'mwith you, like I don't know that
I agree, but you know, I mean,the truth is we don't ever know.
One thing this thing has provenis that we truly don't know.
Like you said, we thought Azure, by this point, might be like
really really dogging AWS'sheels, and I think the AI thing
(15:53):
has definitely really shaken thewhole damn market up and yeah,
so that's a big one.
But yeah, I didn't thinkGoogle's share was, I would not
have thought 13% either, and ofcourse, you do the math on this
right.
So you have what?
31% for Google, 20% for Azure,less for, uh, azure and what was
(16:17):
it 13?
yeah, exactly.
So we're saying are we seeingbasically, the?
The other side is other.
So how much does oci have then?
How much is?
Is all the other private andpublic cloud, you know, startup
or ibm or whatever we're callingpublic cloud now, alibaba and
all those right.
So that's, that seems like apretty yeah.
Anyway, yeah, not much to say,I guess.
(16:39):
Well, we will see if Googlereally does catch up.
I will point out that, ofcourse, we've struggled with
this too.
All the analysts do isdecoupling Microsoft's cloud
Azure offering from 365 and allof their other stuff that's
hosted on Azure.
That's always been an extremelyhard thing to decouple, which I
(17:01):
think they actually use.
They point out in the articleas one of the reasons why
Microsoft has always seemed sofar ahead, but might not really
be so yeah we're going to say,yeah, it's weird.
Chris (17:12):
It's funny how many customers I talk to on a daily
basis, or you know, I should sayprobably prospects, and I just
you know kind of line ofquestioning like hey, what
clouds are you in?
What are you typically using?
They're like oh well, a bit ofAzure, because you know we use a
lot of O365.
And I'm like even the responseto that question is like in that
(17:33):
camp, I'm like that's not, Inot.
I mean, technically, I guessyou're probably using azure, but
I wouldn't say that's, that'sjust pure sass right, like
that's right.
I'm consuming services that arehosted in azure, but I'm not
like an azure customer right soit's like, but that's I mean but
that that could also speak towhy microsoft has such success,
that there's cohesion betweenthat like, even the mindset, is
(17:53):
that like if I, using O365, Iconsider myself an Azure
customer, and therefore, it'seasier for me to get started in
Azure, because I'm quote-unquotealready there, even though
you're not really right.
Tim (18:03):
But yeah, I get that.
That's interesting.
Okay, and we'll close out thisweek with kind of a funny one.
Honestly, I think it'sabsolutely ridiculous.
So NVIDIA had a conferencerecently, uh, where uh jensen
huang of uh nvidia basicallysaid that he did, he did a bunch
of stuff.
(18:24):
That was hilarious.
I don't know if you saw it.
There's a video where he comesout with a giant uh uh nvidia
gpu on like what looks like ashield.
Basically, no, I did not didnot see that.
Oh, you got to find it Because Iwas joking with somebody.
I was like it's way too flashy,You're not going to be able to
deflect any and it looks weak.
You can't deflect any.
Sword strikes or maces withthat thing or anything.
(18:45):
It looks like a buckler, like abig old shield.
Anyway, so in this article fromAOLcom of all places, so I you
know, if you right, I mean justgo get one of the CDs.
If you want to read this, go getone of those free CDs the
thousand hours so you can readthe article, so anyway.
(19:06):
So Jensen Huang says basically,that AI agents are becoming a
thing and, in this new era of AIagents, that IT service
organizations will become thequote, unquote, hr of AI agents.
So yeah, a lot of heavy liftingwith a lot of words there.
So the idea, of course, is thatthese agents are essentially
(19:36):
not quite autonomous but, like,purpose built to perform
whatever task the agent has beencreated for, and that,
therefore, the IT would becomethe shepherd, I guess, of these
agents.
And I don't know, I feel likeHR is a bit of a, I don't know.
I don't want to say hot, take,it's a bit of a stretch or
something, because I mean, can Iput the AI agent on a pip if
it's not doing well, you know,can we let it go?
Chris (19:57):
or, you know, I don't know, give it a pay cut.
This just has me feeling sorryfor all the people that are
specifically individualcontributors because they don't
want to be a manager.
It's like, well, well,according to jensen, eventually
you will be a fucking manager.
(20:18):
Yep over something that could.
It won't talk back to you, Iguess.
So, um, you know, maybe, maybeyou get.
You get the, you know, thatpower complex that you've always
wanted.
Tim (20:22):
You can yeah you really, uh , you know, push down hard on
your, on your, uh, your workersor you can build an ai agent
whose, uh, whose job is to agreewith everything that you say.
Perhaps that's you know.
Give it a prompt, load up theagent with a prompt that says
it'll uh, yeah, like you're mymanager and I love you and I'll
do anything you say who knows?
(20:42):
Uh, I don't know what to say tothis.
It's an absolute.
It's kind of a ridiculousarticle.
I mean, I get the feeling likeit was made for clicks, but not
the article, sorry.
Chris (20:54):
The wording or whatever, if you will, the quotation.
Tim (20:57):
It was meant to turn heads and get headlines.
But I mean, ai agents is athing.
It's the idea of we're kind ofpreloading prompts and then kind
of packaging up, almost like aDocker container, an AI agent
that will, you know, follow theprompt and and and carry out the
task for which it has beenprepared.
But uh, yeah, I don't, I stilldon't know if it becomes the hr
(21:21):
of ai agents or just kind oflike the you know, life cycle
manager of an ai agent.
Chris (21:26):
It's probably more accurate yeah, I mean like you
said.
I mean the.
He can have the quote asridiculous as he wants, but
really what it means is justthat you're changing how you
interface with their technology,right?
So it's like, sure, maybe Imean 2025, I have actually heard
multiple times that 2025 willbe the year of the AI agents.
Tim (21:47):
Yeah, so that does sound like a thing.
2025 will be the year of the AIagents.
Chris (21:48):
Yeah, so that does sound like a thing.
But you know, we potentiallyhave some talks coming up with
some guests, maybe about some ofthese agents, and we'll get
into that.
But yeah, I feel like thenumber one phrase we say on this
Fortnum Huntley news show iswe'll see.
This is no exception right, weshall see.
Tim (22:11):
Yeah, yeah, absolutely, it's a.
It's a new way to consume ai by, by kind of making it more
autonomous, and I think that'sthe goal and that's where he's
getting to.
Is that that you know we'regoing to allow ai to be more
autonomous and do more work onits own so it can secretly
create skynet under our notices?
But, uh, you know, yeah, 2025,2025, you're the AI agent.
You heard it here first.
Alright, and with that I guesswe can go ahead and wrap it up.
(22:36):
I hope everybody had a goodtime.
Of course, we always have somestories that we don't
necessarily cover on the news.
They're in the Fortnightly Newsdocument that we put in the
show notes of every episode.
We encourage you to take a lookat those stories.
There's just stuff that wedidn't have time or or need to
get into very, very deeply onthe show, uh, but we think are
(23:01):
interesting and that ourlisteners, viewers, whatever,
will want to take a look at.
So, uh, take a look at that.
And, of course, always, uh,like, subscribe, buy the
breakfast cereal, play the homegame and we'll see you next time
.
Peace, hi, everyone.
It's Tim, and this has been theCables to Clouds podcast.
Thanks for tuning in today.
(23:21):
If you enjoyed our show, pleasesubscribe to us in your
favorite podcast catcher as wellas subscribe and turn on
notifications for our YouTubechannel to be notified of all
our new episodes.
Follow us on socials atCables2Clouds.
You can also visit our websitefor all the show notes at
Cables2Cloudscom.
Thanks again for listening andsee you next time.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!