August 13, 2024 • 41 mins
Welcome back, for a scorching news episode! Alex Perkins from The Cables2Cloud Podcast joins William and Eyvonne as we dive into the inferno of outages, chaos, legal battles, and hacks that set the July news cycle ablaze. Buckle up for this high-temperature discussion.
Show Links
Switzerland Requires government software to be open source: https://www.zdnet.com/article/switzerland-now-requires-all-government-software-to-be-open-source/
AT&T Number hack: https://www.washingtonpost.com/business/2024/07/12/att-wireless-hacker-data-breach/
TikTok OpenAI Spend: https://www.datacenterdynamics.com/en/news/tiktok-spent-20m-a-month-on-microsofts-azure-openai-service-report/
CrowdStrike sued by Shareholders: https://www.bbc.com/news/articles/cy08ljxndr4o
Microsoft says Cyber Attack Triggered Latest Outage: https://www.bbc.com/news/articles/c903e793w74o
Microsoft lists OpenAI as Competitor: https://techcrunch.com/2024/08/01/microsoft-now-lists-openai-as-a-competitor-in-ai-and-search
Chip manufacturing continues to be a problem: https://www.theverge.com/2024/8/1/24210656/intel-is-laying-off-over-10000-employees-and-will-cut-10-billion-in-costs
Follow, Like, and Subscribe!
Podcast: https://www.thecloudgambit.com/
YouTube: https://www.youtube.com/@TheCloudGambit
LinkedIn: https://www.linkedin.com/company/thecloudgambit
Twitter: https://twitter.com/TheCloudGambit
TikTok: https://www.tiktok.com/@thecloudgambit
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Intro (00:00):
Welcome back for a scorching news episode.
Alex Perkins, from the Cablesto Cloud podcast, joins William
and Yvonne as we dive into theinferno of outages, chaos, legal
battles and hacks that set theJuly news cycle ablaze.
Buckle up for thishigh-temperature discussion.
William (00:29):
Welcome to another episode of the Cloud Gambit.
With me is my excellent co-host, yvonne Sharp.
Welcome, how are you doing thismorning, yvonne?
Eyvonne (00:41):
Hello, doing great.
It's a Monday morning.
While we're recording.
William (00:50):
I don't know what day it's going to be when you're
listening, but it's shaken up tobe a Monday and with us we have
a returning guest, alex Perkins.
How?
Alex (00:55):
are you doing Alex?
Yes, like Yvonne said, it isdefinitely a shaping up to be a
Monday with all the news andstuff.
William (01:02):
It's a case of the Mondays oh man, yeah it's got a
case of the Mondays.
Eyvonne (01:05):
Oh man, yeah, it's a collective case of the Mondays.
William (01:06):
Lots of news and a good Monday it is yeah.
What about when Mondays justare like every day is a Monday,
like Monday through Friday, it'slike, oh, that's called a week
and you're working Saturday?
Alex (01:17):
Yeah, yeah, it's called a week, this is true, or life
right.
William (01:24):
Yeah, the news has been a spinning month of July, Just
a crazy month and just kind ofcontext for the audience.
There is no way that Yvonne andI are going to be cranking out
a news episode every other weekon the Cloud Gambit like the
(01:47):
media production team over onCables to Cloud.
It's probably going to be aonce in a while when a lot of
interesting news is queued up,kind of like what's happened
this last July.
There's just too much to ignore.
Too many wild things havehappened.
Do you want to kick us off,Yvonne?
Eyvonne (02:09):
Yeah, and we're talking tech news.
We're not even talkingpolitical news, but July was
quite the month for headlines,but we're going to stick to tech
today, much to all of ourrelief.
So the first article we wantedto discuss is from ZDNet, and
the Switzerland federalgovernment has required the
(02:32):
releasing of software as opensource.
So this is an incrediblyinteresting topic to think about
what it would mean for anentire federal government to
start pushing their entireapparatus toward open source.
One of the things in thearticle, though, that it says it
(02:54):
says this new law requires allpublic bodies to disclose the
source of code softwaredeveloped by or for them, unless
third party rights or securityconcerns prevent it.
This public money, public codeapproach aims to enhance
government operations,transparency, security and
efficiency.
So you know, I think I firstread the headline and thought,
(03:18):
oh, how do you go?
100% open source?
But I think it's a little bitmore nuanced than that.
But I do appreciate that whatthey're saying is, if we have
code that we have written orthat is written for us, it must
be open source, and I believethis will also drive the entire
(03:39):
organization toward open sourcein a new way.
Thoughts from you two I got afew just quick, toward open
source in a new way Thoughtsfrom you two.
William (03:48):
I got a few just quick, I think.
First of all, I think it'sawesome, I think it's a great
proposition, especially on thesurface, you know, one that kind
of like think of Switzerland atthe forefront of digital
transparency and innovation.
Really know, at least on.
You know, again, on the surface, but in practice I just can't
(04:10):
even begin to fathom howchallenging this is going to be,
this is going to be.
The core definition of this isa journey.
This isn't a thing that we'regonna, you know, execute in a
few years, you know, I thinkthat you know it is in a few
years, you know.
I think that you know it is inthe realm of possibility, of
course, and especially when youthink of in comparison to like
the US.
You know, switzerland'sgovernment is much smaller and
(04:34):
less complex than ours and eventhinking about like European
countries in general, they tendto, you know, put strong
emphasis on digital sovereigntyis the word I'm looking for
which lines right up with opensource principles.
So I can't imagine ourgovernment ever agreeing to move
(04:56):
in this direction.
But smaller governments I guessit's finally the year of the
Linux desktop, alex.
What do you think?
Alex (05:05):
It definitely is the year of the Linux desktop, alex.
What do you think?
It definitely is the year ofthe Linux desktop, for sure.
To me, I think, one of the keythings that, yvonne, you already
pointed this out but it doessay unless the rights of third
parties or security-relatedreasons would exclude or
restrict this.
I think that's really key here,because a company can just claim
, like, how big is that umbrella?
(05:27):
A company can just say, well,this is security related, so we
can't, we can't open source thiscode.
You know, I mean it'sinteresting because will
companies they won't be able tohave that proprietary advantage
when, when writing this code.
So are there going to be lesscompanies that kind of make this
(05:49):
software for the governmentbecause they won't be able to
lock it down and get that moneylike forever from this?
There's like a lot ofinteresting ripple effects that
could happen from this, but Ithink what is most likely going
to happen is that a lot ofcompanies are going to claim
that it's security relatedreasons and they can't open
source.
And I would just I'd like tosee the restrictions on that and
(06:10):
what that actually means.
And this also took a decadejust to get this passed.
So it's been a long fight.
A couple of things that's a goodpoint.
Eyvonne (06:21):
Did I think about that.
So I think it would probably beup to the Swiss government, as
whether or not you know thatbeing written for them is is
meets security requirements notto be released publicly.
If they're using softwarethat's copyrighted by a vendor
(06:42):
Microsoft, VMware, whoever thenthey're not going to release
that Right.
But going forward, I wouldbelieve that they would write
into all their contracts thatany code that is being written
for them will be open sourced,and so I think it will change
the economic reality of thosecontracts, and I suspect if
(07:05):
they're contracting out code, itwill become more expensive to
contract that out to the Swissgovernment, because they will
own it and the individualswriting the code will have to
release it Right.
So it's not like we'll writethis for you, but then we're
also going to maintain therights to it and sell it to
(07:25):
other people, which might have adifferent economic model than
paying somebody to write opensource.
But I also think you know, ifmore governments adopted this
approach, maybe they could learnfrom one another and use one
another's code and systems toimprove.
(07:47):
I agree that this is nevergoing to happen in the US, just
because the lobbyists would keepit from happening.
And one more quick thought isthat you know we talk about how
scale breaks things.
You know, the population ofSwitzerland is 8.7 million
(08:08):
people, not over 300 millionlike the United States, right,
and so the scale of managing apopulation of 300 million is
orders of magnitude differentthan managing a population of
less than 9 million, and I thinkpart of that is what makes this
(08:29):
possible and makes a countrylike Switzerland a good place to
incubate this kind of idea.
Alex (08:38):
Yeah, definitely agree with all that.
William (08:40):
Yeah, A lot to think about.
It'll be interesting to see howthis shakes out and the level
of success.
And then also if other um, youknow, other countries uh,
European or otherwise, you knowlook at this and think it's
crazy or they try to follow suit, and how quickly.
(09:00):
Uh, it's definitely to me seemslike kind of a kind of a
science experiment, but it's ascience experiment that's got
some wheels.
So, yeah, I guess moving,moving along to the next thing,
do you want to?
I guess I'll pivot over to you,Alex, Do you want to talk about
(09:22):
the news that you brought tothis lovely roundtable?
Alex (09:34):
Yep, and this isn't really good news, right, but so Intel
just reported a loss of $1.6billion just in quarter two of
2024.
And they lost $437 million inQ1,.
It looks like so.
This was cost $437 million inQ1, it looks like so.
This was a very significantincrease of over a billion in
just one quarter and due to this, they have announced 15,000
workers are going to be laid off, so downsizing its workforce by
(09:56):
over 15%, as part of a $10billion cost savings plan for
2025.
Just crazy numbers.
And this just continues to keephappening in tech everywhere.
William (10:10):
Yeah, I wonder.
The key priorities in thatarticle, you know are
interesting.
They're kind of like the Idon't want to say stereotypical,
but just this thing that youalways tend to see when this
happens.
You know operational costs arehigh.
You got to reduce some, youknow, simplify portfolio,
eliminate that complexity, youknow, be more efficient when it
(10:35):
comes to the fabs and bringingsome net, you know absolutely
necessary to the.
You know the success of theUnited States back over to the
United States to diversify wheresome of these things come from.
(10:57):
So we're not reliant, so supplychain isn't so crazy, so
they're doing some, some veryimportant work and it's, you
know, it's.
This is part of, I guess,running a business.
You know again, when you,yvonne, was talking about scale,
when you scale to a certainpoint and then your business
really does a hard pivot.
So it's really challenging fora business the size of Intel,
(11:22):
you know, to pivot like that.
So it'll be interesting to seehow this plays out.
Eyvonne (11:29):
Well, and this I put this Intel news in a different
category than all the other bigtech layoffs that we've seen in
the last year or two, partlybecause what's going on at
NVIDIA and even Broadcom right,those two chip manufacturers are
doing incredibly well in themarketplace.
(11:50):
So I think there's more goingon here with NVIDIA than just
the standard tech.
Yeah, we overhired and weoverspent, and now we have to
come back.
The story's different here, andI believe I saw Keith Townsend
make some comments on Twitterand so I'm probably paraphrasing
some of his thoughts, but Ithink what's going on is that,
(12:13):
ultimately, intel needed to makea pivot a couple of years ago
and needed to make someinvestments that weren't made,
and now the market has shiftedand so they're having to make
those investments, and they'realso suffering from the losses
of not making those a couple ofyears ago, and so they're having
(12:33):
to make some really difficultdecisions.
Now we live in a reality where,at some point, companies have to
invest in the long term tocontinue to be successful, but
the market incentives don'tencourage that kind of
investment.
They're always for theshort-term payback, and unless
(12:57):
you have very clear-minded,strong leadership and folks who
really understand the market andwho are thinking long-term and
aren't thinking about theshort-term payout, then we're
going to continue to seeorganizations make these kinds
of difficult moves because theyover-optimize on the short-term
(13:18):
and not the long-term, and Ithink it's unusual in this
market for a chip manufacturernot to be doing well.
And I think what we're seeinghere is the results of some
short-term thinking.
That happened a few years ago,before Gelsinger was even in the
seat, and now they're payingthe piper.
Alex (13:42):
Yeah, I'm glad you pointed that out, that Gelsinger hasn't
been here that long because hewas with VMware and then just a
couple of years ago came backover to Intel.
He previously was at Intel fora long time I think before that.
But it takes a long time for acompany this big to pivot.
And a lot of what you said,Yvonne, is so true that
(14:05):
everybody's so focused on theshort term right now that
pivoting an entire company thatcan lose one point six billion
in a quarter, Like that's not aneasy task to do, but especially
in chip fabrication?
Eyvonne (14:19):
Right, because you have to engineer the thing, you have
to design the fabricationprocess and then you have to
fabricate it and it's you haveto do all of that before you
make the first dollar.
Right, and so it's.
It's a high stakes, very highcapital world and that's why we
see the dominance right now ofcompanies like NVIDIA, because
(14:42):
they were leading the charge onthose chips and they got those
chips to market.
All those GPUs had them fabbedand ready, and then just
opportunities in software reallyare taking advantage of those
capabilities now and it's hardto catch up.
It's not like we can hire somedevelopers and lock them in the
(15:03):
room for a year and be there.
William (15:06):
It's way more complex than that.
And one thing that's justinteresting about this too is
it's not like the chip fab game,is.
It's not like cybersecurity,where there's like a bazillion
companies doing it.
It's not that big and thatsaturated of a market.
So when one of these thingshappens, it's definitely
(15:28):
noteworthy news.
And well, I do got to say alex,you know you mentioned what is
it 1, 1.8 billion in q2, 1.6 yepor 1.6 um, I think cloud strike
uh walks in the room and sayshold my beer, because I think
they lost 25 billion in like twoweeks or less.
What has it been?
Not two weeks, like 12 days.
Alex (15:52):
Yeah, probably.
William (15:53):
Maybe a little more something like that.
So the next article I actuallywanted to talk about this is a
good pivot is a BBC article fromthe BBC CloudStrike, sued by
shareholders over global outage.
So if you're, basically ifyou're a cybersecurity startup
and you push a faulty softwareupdate that crashes like over
(16:15):
like 8 million computersglobally, which subsequently
leads to like over I don'tremember the percentage like
over 30% drop in share price andhowever long it's been, you
know it turns out.
You will hear from yourshareholders or their lawyers,
you know, I mean, you know 25billion loss.
That's pretty gigantic.
(16:36):
I don't think it can get anyworse than that, at least
historically.
Don't say that.
Eyvonne (16:43):
I really don't want you to tempt the universe to show
us otherwise.
William (16:47):
You saw July, wait till you see August.
Eyvonne (16:50):
That's right.
William (16:50):
I better knock on this wood.
So I guess to you both like anyideas what I mean, we could
talk about the semantics of thelike all the you know what could
have been done and what theydid wrong, and oh, the you know
cloud strike so bad or whatever.
But um, what happens next, likewhat do you all see is kind of
(17:13):
like, or just what are yourthoughts on it in general?
Alex (17:17):
I you know what, what I have a little different
perspective on this.
Um, I don't really this is hardto explain properly, but I don't
really love that shareholderscan come in and sue a company
for something like this.
The lawsuit accuses it says itright here the lawsuit accuses
(17:41):
the company of making false andmisleading statements about its
software testing.
It just sounds like is that?
Are the shareholders makingthis lawsuit truly educated
about the internal softwaredevelopment processes and
testing that's done by thiscompany?
It's just a little weird to methat like because I've read a
(18:02):
lot of articles about you knowwhy this happened and the stuff
that they're going to put inplace, and it just seems weird
that a bunch of shareholders canget together and say, oh well,
you guys did this on purpose.
Like it almost seems like anegligence thing, like they're
kind of representing it, likeCrowdStrike was being negligent
(18:22):
and just pushing things out, butthey had a lot of stuff in
place to account for this and itwas a bug in.
Like testing software is, Ithink, what it really boiled
down to.
So to me it's just a weird turnof events.
William (18:38):
It's all about the Benjamins, really.
Eyvonne (18:42):
A couple thoughts.
I mean I think for me thebiggest like as I look at
CrowdStrike, the thing I findinexcusable is that they weren't
doing canary deployments.
You know that there should havebeen a smaller subset of
devices that got this updatebefore the whole world did Like
(19:04):
and I do find that fairlyinexcuscusable considering the
industry they're in and thescale that they were operating
at I agree with that Right.
At the same time, I agree that,like wait, can shareholders just
lose money and get ticked offand go sue.
You know that creates anotherdynamic which makes it hard for
(19:26):
companies to grow and innovateand change.
At the same time, you know,this article also mentions that
Delta is reporting that theylost $500 million in revenue due
to this outage.
When they talk aboutcompensating passengers extra
overtime for repair lost service.
(19:50):
So this whole event, as we allknow, is so significant.
It's incredibly non-trivial, Ithink, ultimately, what we're
going to see from this, andmaybe it's a bit of wishful
thinking, but I am hopeful thatnow we have an event significant
enough to make us rethink howwe do end user computing with
(20:30):
which they can do anything, andthen requiring a stack of agents
and security tools on thosedevices to keep them safe is
suboptimal and that's probablythe understatement of the year.
You know we all you knowbillions of people on the planet
at this point walk around withcell phones and we've never seen
(20:50):
anything like this in the cellphone world and I think that may
become a model going forwardfor how end user devices really
ought to operate.
And yes, there are going to beexceptions, there are going to
be developers.
There are going to beexceptions.
They're going to be developers,they're going to be power users
.
But I think at the end of theday, we now have enough data and
(21:14):
there's been a painful enoughevent to drive some change in
the industry and I thinkultimately that will be good.
But it's going to be hard toget there because there's still
layers of systems that are goingto be some inertia.
So we'll see whether or not thisis significant enough to force
(21:36):
change or whether or not folksjust move on to the next
catastrophe.
William (21:42):
Yeah, now you're, you're absolutely right like
some of the laptops I've workedon in the past and I know you
both have been in this positiontoo.
But like you have so many likethe procurement process just for
all the agents you have, likean agent for per security
engineer at the company.
Practically you go to turn abrand new top of the line laptop
(22:06):
on with all these things onthere and it's like like I
remember you know windows 2000it's just so slow and then all
the updates and it's just.
It almost reminds me of likespyware in a sense, but I
understand why it's there andwhy it's necessary and the
compliance.
There's so many differentthings.
(22:26):
But you're right, there's gotto be a better way to do it.
Eyvonne (22:32):
There just has to be.
I've seen large organizationstake a month to deliver a device
to an end user.
In other words, you've got avery well-paid engineer sitting
on their hands readingdocumentation and not able to
work because the procurement,deployment, security process is
(22:59):
so overwhelming.
And I think, again, that'sanother kind of malfeasance all
on its own.
That's another kind ofmalfeasance all on its own.
That is a hidden cost rightthat I'm hopeful that an event
like this might disrupt.
William (23:17):
I guess we can pivot over to your next article that
you brought Yvonne.
Eyvonne (23:24):
Yeah, again, staying in the security space.
I believe this happened beforeCrowdStrike, so it almost seems
minimal and I hate to say that,but AT&T says hackers stole call
records of nearly all wirelesscustomers and so if you are, or
(23:44):
have ever been an AT&T wirelesscustomer, somebody has the
metadata of all your calls andthere could be names, text
messages at least.
Maybe not the content of thattext message, but the fact that
a text message happened, that atext message happened.
(24:13):
It could reveal the people youfrequently call, allow people to
impersonate numbers.
It's just yet another exampleof data loss.
And how?
Yeah, no matter what you do, ifyou operate in the modern world
, some of that is always likelyto be disclosed, and I'm going
(24:34):
to say likely now as opposed tocould be possible to be
disclosed.
I think it's likely for it tobe disclosed, so careful who you
call.
Alex (24:47):
I have been an AT&T customer since I became an adult
, so it's funny because when Igot the email from AT&T about
this, I brought my wife does notfollow anything tech related
and I told her about it and shewas.
The look on her face of justthis was already obvious that
(25:09):
something like this happens somuch that it she just didn't
even care.
It was like it's just par forthe course.
This is so normal thatsomething like this happened and
it's just it continues tohappen and and, like you said,
it got buried in the news.
Is anything even gonna happenfrom this?
Like I just feel like there's.
It happens so often and no onecares or says anything and
(25:32):
there's no consequence for it.
William (25:35):
I think this one is very this is almost like a
different kind.
I mean, it's just massivelysignificant.
You know, based on the scale.
But you know, unlike the, youknow things we see where you
lose, like your PII, likepersonal identifiable, you know
socials, maybe, things like that.
You know, like Yvonne said,this is, you know this targeted
(25:57):
metadata which can really revealsome you know crazy detailed
patterns of communication andjust movement.
Even you know the ability to.
You know, take this metadata andand like, maybe analyze it at
some sort of scale.
You're thinking every, everycustomer in that time window.
I mean to me that's almost likea national security issue.
(26:20):
I mean how many critical youknow officials are using at&&T
out there, like, and you thinkabout their patterns, their
movements and all you knowmetadata is.
So there's just so much there,like every time you take a
picture and you upload it to thecloud.
It's like longitude, latitude,where it was taken, location,
(26:41):
there's everything you can thinkof.
So a lot of the metadata that'stied to stuff.
It's extremely detailed, itjust seems detailed.
Alex (26:50):
It's a really good point and, especially with all the AI
stuff going on, I mean, what'sstopping someone from getting
this massive data and feeding itinto a learning model?
And just there you go, you havea profile on every single
customer that has used AT&T.
So yeah, that's a very scarything to think about.
Eyvonne (27:11):
And if this article is to be believed, it says AT&T
said the attack began withillicit access to one of its
accounts with a major but lowprofile cloud data storage
company called Snowflake, which,first of all, that description
makes me chuckle.
Second, that description seemsto imply that it was just a
(27:35):
credential attack, that itwasn't an attack of
infrastructure, it wasn'tnecessarily a failure of the
security of the snowflakeservice.
It was a credential attack, ishow I'm reading this right.
And so again, two-factor folks,all day, every day, two-factor
(28:02):
at all.
Please, pretty please, for thesake of all of our data.
And and again, I'm not asecurity expert, so I don't
there may be more to this thanthan the Washington post is
revealing.
At the same time, it's, it's ajungle out there.
William (28:24):
It is definitely a jungle.
So next on the list is TikTokspent $20 million a month on
Microsoft Azure's OpenAI service.
That's a lot of money that Ithink are worth mentioning here
(28:47):
from my perspective is you know,azure's opening eye service is
obviously generating giganticamounts of revenue.
I think the expectations werelike for this year is like over
a billion or something, and akey contributor to that revenue,
as it happens, it was TikTok.
I don't think it actually tellsthe exact amount that TikTok
(29:08):
spends overall for the service,but this is honestly a testament
in my opinion.
Of course, it says that TikTok'suh parent company, bytedance,
(29:29):
is also developing its own ummodels and process for this
stuff.
Of course, um and and thatwould lead, you know, I believe
they originally tried to buildtheir own models and then they
pivoted to a partnership withOpenAI, you know, in order to
(29:55):
probably move things along untilsuch a time that they've got
their own, you know, homegrownsolution available, and then
they will probably go that way.
It was kind of surprising,honestly, to see apple do that
just because they're so they'reall like about building
internally and um, you know,they're an engineering company
(30:17):
at the end of the day.
So, um, any any thoughts fromthe peanut gallery on this one
yeah, I mean 20 million.
Alex (30:29):
It crazy, these numbers that just keep getting thrown
around for all these newsarticles, man.
But 20 million a month is a lotfor TikTok to spend, um.
But at the same time it's likeopen AI, like could I might even
actually lose?
I was trying to find somespecific numbers for this, but I
(30:50):
saw something that said theycould lose as much as $5 billion
this year just from how much itcosts for them to train their
models and run on Azure and I'msure they have deals with
Microsoft to run on Azure.
I was trying to find the datafor that and I couldn't find it.
William (31:07):
That makes TikTok multi-cloud.
Don't they run their core stuffon OCI?
I think yeah, they do.
I'd have to look it up.
I'm pretty sure it's OCI it's.
Alex (31:17):
OCI.
I think I remember reading thatfor sure.
There you go.
There's a good example of acompany that's multi-cloud.
Go ahead, alex.
No, go ahead.
Eyvonne (31:30):
I'm trying to get my thoughts together, it's an
interesting topic and I think,ultimately, when we see this
kind of spend, I don't know, I'm, I'm, I'm having a hard time
too.
William, you're just going tohave to cut us out.
William (31:53):
Yeah, I guess we can, I'll, I'll do it.
So we're on a sort of a youknow a home run, just uh, with
Microsoft, I think my out of allthe companies that had a bad
month, I think Microsoft, youknow, I didn't even include the
central US outage on here, butthey had a tough it was a tough
tough month for.
Microsoft.
Eyvonne (32:11):
It's been a rough go, yeah, yeah, I mean even even
though the CrowdStrike outagewasn't officially their fault,
they're taking a ton of heat forit.
But I also mentioned thisbecause we all know how folks
like to blame the network thatI've seen several articles that
spoke about the global Internetoutage that happened the day of
the CrowdStrike attack, and soI'm just here as a source of
(32:36):
clarity.
I'm fighting the misinformationthat that Friday was not an
Internet outage and if youweren't using a Windows device
running CrowdStrike, theinternet works just fine.
William (32:49):
Coincidence, Interesting Last article from me
that I wanted to just throw outhere, since we're on the theme
of Microsoft-y rough July.
On BBCcom, Microsoft says cyberattack triggered latest outage.
Six days ago this was publishedand again Microsoft is having a
(33:15):
rough month Cyber securityattack causing outages, Central
US going down, the CrowdStrikedebacle and then also Microsoft.
To throw just one more in there, Microsoft is now listing
OpenAI as a competitor in AI andsearch.
And that's significant as well,because Microsoft you know, we
(33:40):
know Microsoft invested I don'teven know how much capital in
OpenAI billions and billions,and you know it's integrated
OpenAI pretty heavily.
You know it's also OpenAI'sexclusive cloud provider, I
think you know.
But despite this beautiful,like notebook-esque love story,
you know we're now seeingMicrosoft classifying OpenAI as
(34:01):
a competitor, and I mean my ownthoughts here.
There's always some narrativebehind the scenes, but I think
the the article references theftc's ongoing antitrust
investigations, you know,basically between um, you know,
ai, startups and cloud providers.
Alex (34:17):
So an interesting nugget there yeah, I mean, I people had
to think that this was going tohappen.
I know a lot of this came aboutwhen they just announced their
new like search GPT beta feature, because Google I saw in
another article Google also saidthat they now see them as a
(34:40):
competitor, or listed them as acompetitor.
So of course, as soon as theyannounce a search tool, right,
they're going to becomecompetitors for everyone else.
It is a very strange turn ofevents because they're they run
on Azure, so that is weird, and,of course, microsoft owns like
(35:02):
a very large portion of of openAI as well.
Um, but yeah, I, it's.
It's funny seeing them publiclyacknowledge that their
competitor, after they've spentso much money, invested so much
money into them, for sure.
Eyvonne (35:17):
Well, I think you know, in any time you you work on the
vendor side in in technology,you find that there's often a
frenemy situation that comes upthat in some ways we work
together and in some ways wecompete with one another and
trying to find that sweet spot.
For how do we leverage partnersand other vendors and work
(35:43):
together to create a solutionthat's mutually beneficial for
everybody, but not let thebalance of power get so out of
whack that one person isbenefiting more, one company is
benefiting more than another oneor ends up being in a predatory
kind of a situation.
It is a really difficult lineto walk.
(36:04):
I think Microsoft'srelationship with OpenAI has
always been a little different.
You know, we had that incidentseveral months ago where OpenAI
the board fired Sam Altmanwithout Microsoft's input and
then Satya jumped in andreversed that decision, but then
we found Microsoft ended upwith just a non-voting position
(36:30):
on the board and now they're nolonger on the board.
I mean, I think you know OpenAIis on the ascendancy and
they're using the relationshipsthat they can, at whatever stage
that they're in, to moveforward, and then they're going
to move on, and so I think itmakes sense that Microsoft says
(36:53):
that they're a competitor.
Google definitely sees them as acompetitor and it's going to be
interesting as AI shakes out,especially as we head into the
trough of disillusionment.
I think we're pretty much,we're almost there, right?
What happens as we emerge fromthat into the slope of
enlightenment, you know, andthat whole hype cycle and AI
(37:13):
actually gets implemented and isused more broadly?
Or is it going to be asituation where you have a few
core companies that really useit and they become the control
centers for AI?
I don't think we know howthat's going to shake out yet,
but it's an interesting.
(37:35):
It will be interesting to watchover the next few years.
William (37:39):
That's a good point.
I think what you're saying Imean what you're saying is kind
of a feather in the cap for howfast.
I think a lot of these thingsare results of how fast this
technology is moving.
You know, it is just changingfaster than anything I've ever
seen in my life.
Eyvonne (37:59):
Yeah, and I think usually it's years to go through
that whole process for anytechnology.
And I think what that means isthat you know, like, as we've
compressed the ends of theheight cycle, we've increased
the frequency or increased thepeaks of the curve Right, and so
you know, and I think you know,it's also going to be
(38:20):
interesting to watch as themarket responds to that.
So, you know, I think we've gotinteresting days ahead.
As far as, like the Microsoftoutage that you first started
with William it, you know, thisarticle indicates it was a DDoS
attack.
Ddos attacks are incrediblydifficult.
(38:43):
I think you know we saw, youknow, google Cloud defend
successfully against a 398million packet per second attack
several years ago, and it'sgoing to be an increasing
challenge.
I know Russ White, over on thehedge, has been on a
centralization kick.
I think we're you know, and Ithink we're going to see some
(39:04):
move to edge.
I think we're you know and Ithink we're going to see some
move to edge.
I think we're always going tohave centralization, though, and
we're going to have to figureout how to better defend and,
yes, it's been a really roughmonth for Microsoft.
Alex (39:20):
Yeah, that article.
It's funny because it says ittook down Minecraft as well.
Eyvonne (39:26):
So just want to point that out Elementary age children
were distressed everywhere.
William (39:32):
Many, many tears were shed in this outage by many age
groups.
Alex (39:37):
Yeah, but also real quick.
There's in this article.
It says Professor Alan Woodwardsaid you'd expect Microsoft's
network infrastructure to bebomb proof.
That's I.
Just why would you say thatthat's such a?
I don't know who the professoris, but that's a very weird
(39:57):
statement to make.
Eyvonne (39:58):
Show me someone who has never run an operational
network.
Exactly Without showing mesomeone who's never run an
operational network.
Yeah.
William (40:09):
The thing we work in called technology is apparently
fickle, and we're sitting atopprotocols that were invented in
the 1980s 80s really and 90sthat we rely on predominantly
for all of this stuff.
All right Well I guess it'sabout time um, I guess everybody
(40:31):
knows where to find you.
Alex, do you want to just throwyour social handles anyway?
Alex (40:38):
yeah, uh, just at bumps in the wire.
Um, I don't really post toomuch on twitter x, whatever you
want to call it but veryoccasionally I'll just write
some random, something random,linkedin I'm always on LinkedIn
or X.
So if you want to reach out tome, feel free and I'm sure
you'll get a response veryquickly.
Welcome back for a scorching news episode.
Alex Perkins, from the Cablesto Cloud podcast, joins William
and Yvonne as we dive into theinferno of outages, chaos, legal
battles and hacks that set theJuly news cycle ablaze.
Buckle up for thishigh-temperature discussion.
William (00:29):
Welcome to another episode of the Cloud Gambit.
With me is my excellent co-host, yvonne Sharp.
Welcome, how are you doing thismorning, yvonne?
Eyvonne (00:41):
Hello, doing great.
It's a Monday morning.
While we're recording.
William (00:50):
I don't know what day it's going to be when you're
listening, but it's shaken up tobe a Monday and with us we have
a returning guest, alex Perkins.
How?
Alex (00:55):
are you doing Alex?
Yes, like Yvonne said, it isdefinitely a shaping up to be a
Monday with all the news andstuff.
William (01:02):
It's a case of the Mondays oh man, yeah it's got a
case of the Mondays.
Eyvonne (01:05):
Oh man, yeah, it's a collective case of the Mondays.
William (01:06):
Lots of news and a good Monday it is yeah.
What about when Mondays justare like every day is a Monday,
like Monday through Friday, it'slike, oh, that's called a week
and you're working Saturday?
Alex (01:17):
Yeah, yeah, it's called a week, this is true, or life
right.
William (01:24):
Yeah, the news has been a spinning month of July, Just
a crazy month and just kind ofcontext for the audience.
There is no way that Yvonne andI are going to be cranking out
a news episode every other weekon the Cloud Gambit like the
(01:47):
media production team over onCables to Cloud.
It's probably going to be aonce in a while when a lot of
interesting news is queued up,kind of like what's happened
this last July.
There's just too much to ignore.
Too many wild things havehappened.
Do you want to kick us off,Yvonne?
Eyvonne (02:09):
Yeah, and we're talking tech news.
We're not even talkingpolitical news, but July was
quite the month for headlines,but we're going to stick to tech
today, much to all of ourrelief.
So the first article we wantedto discuss is from ZDNet, and
the Switzerland federalgovernment has required the
(02:32):
releasing of software as opensource.
So this is an incrediblyinteresting topic to think about
what it would mean for anentire federal government to
start pushing their entireapparatus toward open source.
One of the things in thearticle, though, that it says it
(02:54):
says this new law requires allpublic bodies to disclose the
source of code softwaredeveloped by or for them, unless
third party rights or securityconcerns prevent it.
This public money, public codeapproach aims to enhance
government operations,transparency, security and
efficiency.
So you know, I think I firstread the headline and thought,
(03:18):
oh, how do you go?
100% open source?
But I think it's a little bitmore nuanced than that.
But I do appreciate that whatthey're saying is, if we have
code that we have written orthat is written for us, it must
be open source, and I believethis will also drive the entire
(03:39):
organization toward open sourcein a new way.
Thoughts from you two I got afew just quick, toward open
source in a new way Thoughtsfrom you two.
William (03:48):
I got a few just quick, I think.
First of all, I think it'sawesome, I think it's a great
proposition, especially on thesurface, you know, one that kind
of like think of Switzerland atthe forefront of digital
transparency and innovation.
Really know, at least on.
You know, again, on the surface, but in practice I just can't
(04:10):
even begin to fathom howchallenging this is going to be,
this is going to be.
The core definition of this isa journey.
This isn't a thing that we'regonna, you know, execute in a
few years, you know, I thinkthat you know it is in a few
years, you know.
I think that you know it is inthe realm of possibility, of
course, and especially when youthink of in comparison to like
the US.
You know, switzerland'sgovernment is much smaller and
(04:34):
less complex than ours and eventhinking about like European
countries in general, they tendto, you know, put strong
emphasis on digital sovereigntyis the word I'm looking for
which lines right up with opensource principles.
So I can't imagine ourgovernment ever agreeing to move
(04:56):
in this direction.
But smaller governments I guessit's finally the year of the
Linux desktop, alex.
What do you think?
Alex (05:05):
It definitely is the year of the Linux desktop, alex.
What do you think?
It definitely is the year ofthe Linux desktop, for sure.
To me, I think, one of the keythings that, yvonne, you already
pointed this out but it doessay unless the rights of third
parties or security-relatedreasons would exclude or
restrict this.
I think that's really key here,because a company can just claim
, like, how big is that umbrella?
(05:27):
A company can just say, well,this is security related, so we
can't, we can't open source thiscode.
You know, I mean it'sinteresting because will
companies they won't be able tohave that proprietary advantage
when, when writing this code.
So are there going to be lesscompanies that kind of make this
(05:49):
software for the governmentbecause they won't be able to
lock it down and get that moneylike forever from this?
There's like a lot ofinteresting ripple effects that
could happen from this, but Ithink what is most likely going
to happen is that a lot ofcompanies are going to claim
that it's security relatedreasons and they can't open
source.
And I would just I'd like tosee the restrictions on that and
(06:10):
what that actually means.
And this also took a decadejust to get this passed.
So it's been a long fight.
A couple of things that's a goodpoint.
Eyvonne (06:21):
Did I think about that.
So I think it would probably beup to the Swiss government, as
whether or not you know thatbeing written for them is is
meets security requirements notto be released publicly.
If they're using softwarethat's copyrighted by a vendor
(06:42):
Microsoft, VMware, whoever thenthey're not going to release
that Right.
But going forward, I wouldbelieve that they would write
into all their contracts thatany code that is being written
for them will be open sourced,and so I think it will change
the economic reality of thosecontracts, and I suspect if
(07:05):
they're contracting out code, itwill become more expensive to
contract that out to the Swissgovernment, because they will
own it and the individualswriting the code will have to
release it Right.
So it's not like we'll writethis for you, but then we're
also going to maintain therights to it and sell it to
(07:25):
other people, which might have adifferent economic model than
paying somebody to write opensource.
But I also think you know, ifmore governments adopted this
approach, maybe they could learnfrom one another and use one
another's code and systems toimprove.
(07:47):
I agree that this is nevergoing to happen in the US, just
because the lobbyists would keepit from happening.
And one more quick thought isthat you know we talk about how
scale breaks things.
You know, the population ofSwitzerland is 8.7 million
(08:08):
people, not over 300 millionlike the United States, right,
and so the scale of managing apopulation of 300 million is
orders of magnitude differentthan managing a population of
less than 9 million, and I thinkpart of that is what makes this
(08:29):
possible and makes a countrylike Switzerland a good place to
incubate this kind of idea.
Alex (08:38):
Yeah, definitely agree with all that.
William (08:40):
Yeah, A lot to think about.
It'll be interesting to see howthis shakes out and the level
of success.
And then also if other um, youknow, other countries uh,
European or otherwise, you knowlook at this and think it's
crazy or they try to follow suit, and how quickly.
(09:00):
Uh, it's definitely to me seemslike kind of a kind of a
science experiment, but it's ascience experiment that's got
some wheels.
So, yeah, I guess moving,moving along to the next thing,
do you want to?
I guess I'll pivot over to you,Alex, Do you want to talk about
(09:22):
the news that you brought tothis lovely roundtable?
Alex (09:34):
Yep, and this isn't really good news, right, but so Intel
just reported a loss of $1.6billion just in quarter two of
2024.
And they lost $437 million inQ1,.
It looks like so.
This was cost $437 million inQ1, it looks like so.
This was a very significantincrease of over a billion in
just one quarter and due to this, they have announced 15,000
workers are going to be laid off, so downsizing its workforce by
(09:56):
over 15%, as part of a $10billion cost savings plan for
2025.
Just crazy numbers.
And this just continues to keephappening in tech everywhere.
William (10:10):
Yeah, I wonder.
The key priorities in thatarticle, you know are
interesting.
They're kind of like the Idon't want to say stereotypical,
but just this thing that youalways tend to see when this
happens.
You know operational costs arehigh.
You got to reduce some, youknow, simplify portfolio,
eliminate that complexity, youknow, be more efficient when it
(10:35):
comes to the fabs and bringingsome net, you know absolutely
necessary to the.
You know the success of theUnited States back over to the
United States to diversify wheresome of these things come from.
(10:57):
So we're not reliant, so supplychain isn't so crazy, so
they're doing some, some veryimportant work and it's, you
know, it's.
This is part of, I guess,running a business.
You know again, when you,yvonne, was talking about scale,
when you scale to a certainpoint and then your business
really does a hard pivot.
So it's really challenging fora business the size of Intel,
(11:22):
you know, to pivot like that.
So it'll be interesting to seehow this plays out.
Eyvonne (11:29):
Well, and this I put this Intel news in a different
category than all the other bigtech layoffs that we've seen in
the last year or two, partlybecause what's going on at
NVIDIA and even Broadcom right,those two chip manufacturers are
doing incredibly well in themarketplace.
(11:50):
So I think there's more goingon here with NVIDIA than just
the standard tech.
Yeah, we overhired and weoverspent, and now we have to
come back.
The story's different here, andI believe I saw Keith Townsend
make some comments on Twitterand so I'm probably paraphrasing
some of his thoughts, but Ithink what's going on is that,
(12:13):
ultimately, intel needed to makea pivot a couple of years ago
and needed to make someinvestments that weren't made,
and now the market has shiftedand so they're having to make
those investments, and they'realso suffering from the losses
of not making those a couple ofyears ago, and so they're having
(12:33):
to make some really difficultdecisions.
Now we live in a reality where,at some point, companies have to
invest in the long term tocontinue to be successful, but
the market incentives don'tencourage that kind of
investment.
They're always for theshort-term payback, and unless
(12:57):
you have very clear-minded,strong leadership and folks who
really understand the market andwho are thinking long-term and
aren't thinking about theshort-term payout, then we're
going to continue to seeorganizations make these kinds
of difficult moves because theyover-optimize on the short-term
(13:18):
and not the long-term, and Ithink it's unusual in this
market for a chip manufacturernot to be doing well.
And I think what we're seeinghere is the results of some
short-term thinking.
That happened a few years ago,before Gelsinger was even in the
seat, and now they're payingthe piper.
Alex (13:42):
Yeah, I'm glad you pointed that out, that Gelsinger hasn't
been here that long because hewas with VMware and then just a
couple of years ago came backover to Intel.
He previously was at Intel fora long time I think before that.
But it takes a long time for acompany this big to pivot.
And a lot of what you said,Yvonne, is so true that
(14:05):
everybody's so focused on theshort term right now that
pivoting an entire company thatcan lose one point six billion
in a quarter, Like that's not aneasy task to do, but especially
in chip fabrication?
Eyvonne (14:19):
Right, because you have to engineer the thing, you have
to design the fabricationprocess and then you have to
fabricate it and it's you haveto do all of that before you
make the first dollar.
Right, and so it's.
It's a high stakes, very highcapital world and that's why we
see the dominance right now ofcompanies like NVIDIA, because
(14:42):
they were leading the charge onthose chips and they got those
chips to market.
All those GPUs had them fabbedand ready, and then just
opportunities in software reallyare taking advantage of those
capabilities now and it's hardto catch up.
It's not like we can hire somedevelopers and lock them in the
(15:03):
room for a year and be there.
William (15:06):
It's way more complex than that.
And one thing that's justinteresting about this too is
it's not like the chip fab game,is.
It's not like cybersecurity,where there's like a bazillion
companies doing it.
It's not that big and thatsaturated of a market.
So when one of these thingshappens, it's definitely
(15:28):
noteworthy news.
And well, I do got to say alex,you know you mentioned what is
it 1, 1.8 billion in q2, 1.6 yepor 1.6 um, I think cloud strike
uh walks in the room and sayshold my beer, because I think
they lost 25 billion in like twoweeks or less.
What has it been?
Not two weeks, like 12 days.
Alex (15:52):
Yeah, probably.
William (15:53):
Maybe a little more something like that.
So the next article I actuallywanted to talk about this is a
good pivot is a BBC article fromthe BBC CloudStrike, sued by
shareholders over global outage.
So if you're, basically ifyou're a cybersecurity startup
and you push a faulty softwareupdate that crashes like over
(16:15):
like 8 million computersglobally, which subsequently
leads to like over I don'tremember the percentage like
over 30% drop in share price andhowever long it's been, you
know it turns out.
You will hear from yourshareholders or their lawyers,
you know, I mean, you know 25billion loss.
That's pretty gigantic.
(16:36):
I don't think it can get anyworse than that, at least
historically.
Don't say that.
Eyvonne (16:43):
I really don't want you to tempt the universe to show
us otherwise.
William (16:47):
You saw July, wait till you see August.
Eyvonne (16:50):
That's right.
William (16:50):
I better knock on this wood.
So I guess to you both like anyideas what I mean, we could
talk about the semantics of thelike all the you know what could
have been done and what theydid wrong, and oh, the you know
cloud strike so bad or whatever.
But um, what happens next, likewhat do you all see is kind of
(17:13):
like, or just what are yourthoughts on it in general?
Alex (17:17):
I you know what, what I have a little different
perspective on this.
Um, I don't really this is hardto explain properly, but I don't
really love that shareholderscan come in and sue a company
for something like this.
The lawsuit accuses it says itright here the lawsuit accuses
(17:41):
the company of making false andmisleading statements about its
software testing.
It just sounds like is that?
Are the shareholders makingthis lawsuit truly educated
about the internal softwaredevelopment processes and
testing that's done by thiscompany?
It's just a little weird to methat like because I've read a
(18:02):
lot of articles about you knowwhy this happened and the stuff
that they're going to put inplace, and it just seems weird
that a bunch of shareholders canget together and say, oh well,
you guys did this on purpose.
Like it almost seems like anegligence thing, like they're
kind of representing it, likeCrowdStrike was being negligent
(18:22):
and just pushing things out, butthey had a lot of stuff in
place to account for this and itwas a bug in.
Like testing software is, Ithink, what it really boiled
down to.
So to me it's just a weird turnof events.
William (18:38):
It's all about the Benjamins, really.
Eyvonne (18:42):
A couple thoughts.
I mean I think for me thebiggest like as I look at
CrowdStrike, the thing I findinexcusable is that they weren't
doing canary deployments.
You know that there should havebeen a smaller subset of
devices that got this updatebefore the whole world did Like
(19:04):
and I do find that fairlyinexcuscusable considering the
industry they're in and thescale that they were operating
at I agree with that Right.
At the same time, I agree that,like wait, can shareholders just
lose money and get ticked offand go sue.
You know that creates anotherdynamic which makes it hard for
(19:26):
companies to grow and innovateand change.
At the same time, you know,this article also mentions that
Delta is reporting that theylost $500 million in revenue due
to this outage.
When they talk aboutcompensating passengers extra
overtime for repair lost service.
(19:50):
So this whole event, as we allknow, is so significant.
It's incredibly non-trivial, Ithink, ultimately, what we're
going to see from this, andmaybe it's a bit of wishful
thinking, but I am hopeful thatnow we have an event significant
enough to make us rethink howwe do end user computing with
(20:30):
which they can do anything, andthen requiring a stack of agents
and security tools on thosedevices to keep them safe is
suboptimal and that's probablythe understatement of the year.
You know we all you knowbillions of people on the planet
at this point walk around withcell phones and we've never seen
(20:50):
anything like this in the cellphone world and I think that may
become a model going forwardfor how end user devices really
ought to operate.
And yes, there are going to beexceptions, there are going to
be developers.
There are going to beexceptions.
They're going to be developers,they're going to be power users
.
But I think at the end of theday, we now have enough data and
(21:14):
there's been a painful enoughevent to drive some change in
the industry and I thinkultimately that will be good.
But it's going to be hard toget there because there's still
layers of systems that are goingto be some inertia.
So we'll see whether or not thisis significant enough to force
(21:36):
change or whether or not folksjust move on to the next
catastrophe.
William (21:42):
Yeah, now you're, you're absolutely right like
some of the laptops I've workedon in the past and I know you
both have been in this positiontoo.
But like you have so many likethe procurement process just for
all the agents you have, likean agent for per security
engineer at the company.
Practically you go to turn abrand new top of the line laptop
(22:06):
on with all these things onthere and it's like like I
remember you know windows 2000it's just so slow and then all
the updates and it's just.
It almost reminds me of likespyware in a sense, but I
understand why it's there andwhy it's necessary and the
compliance.
There's so many differentthings.
(22:26):
But you're right, there's gotto be a better way to do it.
Eyvonne (22:32):
There just has to be.
I've seen large organizationstake a month to deliver a device
to an end user.
In other words, you've got avery well-paid engineer sitting
on their hands readingdocumentation and not able to
work because the procurement,deployment, security process is
(22:59):
so overwhelming.
And I think, again, that'sanother kind of malfeasance all
on its own.
That's another kind ofmalfeasance all on its own.
That is a hidden cost rightthat I'm hopeful that an event
like this might disrupt.
William (23:17):
I guess we can pivot over to your next article that
you brought Yvonne.
Eyvonne (23:24):
Yeah, again, staying in the security space.
I believe this happened beforeCrowdStrike, so it almost seems
minimal and I hate to say that,but AT&T says hackers stole call
records of nearly all wirelesscustomers and so if you are, or
(23:44):
have ever been an AT&T wirelesscustomer, somebody has the
metadata of all your calls andthere could be names, text
messages at least.
Maybe not the content of thattext message, but the fact that
a text message happened, that atext message happened.
(24:13):
It could reveal the people youfrequently call, allow people to
impersonate numbers.
It's just yet another exampleof data loss.
And how?
Yeah, no matter what you do, ifyou operate in the modern world
, some of that is always likelyto be disclosed, and I'm going
(24:34):
to say likely now as opposed tocould be possible to be
disclosed.
I think it's likely for it tobe disclosed, so careful who you
call.
Alex (24:47):
I have been an AT&T customer since I became an adult
, so it's funny because when Igot the email from AT&T about
this, I brought my wife does notfollow anything tech related
and I told her about it and shewas.
The look on her face of justthis was already obvious that
(25:09):
something like this happens somuch that it she just didn't
even care.
It was like it's just par forthe course.
This is so normal thatsomething like this happened and
it's just it continues tohappen and and, like you said,
it got buried in the news.
Is anything even gonna happenfrom this?
Like I just feel like there's.
It happens so often and no onecares or says anything and
(25:32):
there's no consequence for it.
William (25:35):
I think this one is very this is almost like a
different kind.
I mean, it's just massivelysignificant.
You know, based on the scale.
But you know, unlike the, youknow things we see where you
lose, like your PII, likepersonal identifiable, you know
socials, maybe, things like that.
You know, like Yvonne said,this is, you know this targeted
(25:57):
metadata which can really revealsome you know crazy detailed
patterns of communication andjust movement.
Even you know the ability to.
You know, take this metadata andand like, maybe analyze it at
some sort of scale.
You're thinking every, everycustomer in that time window.
I mean to me that's almost likea national security issue.
(26:20):
I mean how many critical youknow officials are using at&&T
out there, like, and you thinkabout their patterns, their
movements and all you knowmetadata is.
So there's just so much there,like every time you take a
picture and you upload it to thecloud.
It's like longitude, latitude,where it was taken, location,
(26:41):
there's everything you can thinkof.
So a lot of the metadata that'stied to stuff.
It's extremely detailed, itjust seems detailed.
Alex (26:50):
It's a really good point and, especially with all the AI
stuff going on, I mean, what'sstopping someone from getting
this massive data and feeding itinto a learning model?
And just there you go, you havea profile on every single
customer that has used AT&T.
So yeah, that's a very scarything to think about.
Eyvonne (27:11):
And if this article is to be believed, it says AT&T
said the attack began withillicit access to one of its
accounts with a major but lowprofile cloud data storage
company called Snowflake, which,first of all, that description
makes me chuckle.
Second, that description seemsto imply that it was just a
(27:35):
credential attack, that itwasn't an attack of
infrastructure, it wasn'tnecessarily a failure of the
security of the snowflakeservice.
It was a credential attack, ishow I'm reading this right.
And so again, two-factor folks,all day, every day, two-factor
(28:02):
at all.
Please, pretty please, for thesake of all of our data.
And and again, I'm not asecurity expert, so I don't
there may be more to this thanthan the Washington post is
revealing.
At the same time, it's, it's ajungle out there.
William (28:24):
It is definitely a jungle.
So next on the list is TikTokspent $20 million a month on
Microsoft Azure's OpenAI service.
That's a lot of money that Ithink are worth mentioning here
(28:47):
from my perspective is you know,azure's opening eye service is
obviously generating giganticamounts of revenue.
I think the expectations werelike for this year is like over
a billion or something, and akey contributor to that revenue,
as it happens, it was TikTok.
I don't think it actually tellsthe exact amount that TikTok
(29:08):
spends overall for the service,but this is honestly a testament
in my opinion.
Of course, it says that TikTok'suh parent company, bytedance,
(29:29):
is also developing its own ummodels and process for this
stuff.
Of course, um and and thatwould lead, you know, I believe
they originally tried to buildtheir own models and then they
pivoted to a partnership withOpenAI, you know, in order to
(29:55):
probably move things along untilsuch a time that they've got
their own, you know, homegrownsolution available, and then
they will probably go that way.
It was kind of surprising,honestly, to see apple do that
just because they're so they'reall like about building
internally and um, you know,they're an engineering company
(30:17):
at the end of the day.
So, um, any any thoughts fromthe peanut gallery on this one
yeah, I mean 20 million.
Alex (30:29):
It crazy, these numbers that just keep getting thrown
around for all these newsarticles, man.
But 20 million a month is a lotfor TikTok to spend, um.
But at the same time it's likeopen AI, like could I might even
actually lose?
I was trying to find somespecific numbers for this, but I
(30:50):
saw something that said theycould lose as much as $5 billion
this year just from how much itcosts for them to train their
models and run on Azure and I'msure they have deals with
Microsoft to run on Azure.
I was trying to find the datafor that and I couldn't find it.
William (31:07):
That makes TikTok multi-cloud.
Don't they run their core stuffon OCI?
I think yeah, they do.
I'd have to look it up.
I'm pretty sure it's OCI it's.
Alex (31:17):
OCI.
I think I remember reading thatfor sure.
There you go.
There's a good example of acompany that's multi-cloud.
Go ahead, alex.
No, go ahead.
Eyvonne (31:30):
I'm trying to get my thoughts together, it's an
interesting topic and I think,ultimately, when we see this
kind of spend, I don't know, I'm, I'm, I'm having a hard time
too.
William, you're just going tohave to cut us out.
William (31:53):
Yeah, I guess we can, I'll, I'll do it.
So we're on a sort of a youknow a home run, just uh, with
Microsoft, I think my out of allthe companies that had a bad
month, I think Microsoft, youknow, I didn't even include the
central US outage on here, butthey had a tough it was a tough
tough month for.
Microsoft.
Eyvonne (32:11):
It's been a rough go, yeah, yeah, I mean even even
though the CrowdStrike outagewasn't officially their fault,
they're taking a ton of heat forit.
But I also mentioned thisbecause we all know how folks
like to blame the network thatI've seen several articles that
spoke about the global Internetoutage that happened the day of
the CrowdStrike attack, and soI'm just here as a source of
(32:36):
clarity.
I'm fighting the misinformationthat that Friday was not an
Internet outage and if youweren't using a Windows device
running CrowdStrike, theinternet works just fine.
William (32:49):
Coincidence, Interesting Last article from me
that I wanted to just throw outhere, since we're on the theme
of Microsoft-y rough July.
On BBCcom, Microsoft says cyberattack triggered latest outage.
Six days ago this was publishedand again Microsoft is having a
(33:15):
rough month Cyber securityattack causing outages, Central
US going down, the CrowdStrikedebacle and then also Microsoft.
To throw just one more in there, Microsoft is now listing
OpenAI as a competitor in AI andsearch.
And that's significant as well,because Microsoft you know, we
(33:40):
know Microsoft invested I don'teven know how much capital in
OpenAI billions and billions,and you know it's integrated
OpenAI pretty heavily.
You know it's also OpenAI'sexclusive cloud provider, I
think you know.
But despite this beautiful,like notebook-esque love story,
you know we're now seeingMicrosoft classifying OpenAI as
(34:01):
a competitor, and I mean my ownthoughts here.
There's always some narrativebehind the scenes, but I think
the the article references theftc's ongoing antitrust
investigations, you know,basically between um, you know,
ai, startups and cloud providers.
Alex (34:17):
So an interesting nugget there yeah, I mean, I people had
to think that this was going tohappen.
I know a lot of this came aboutwhen they just announced their
new like search GPT beta feature, because Google I saw in
another article Google also saidthat they now see them as a
(34:40):
competitor, or listed them as acompetitor.
So of course, as soon as theyannounce a search tool, right,
they're going to becomecompetitors for everyone else.
It is a very strange turn ofevents because they're they run
on Azure, so that is weird, and,of course, microsoft owns like
(35:02):
a very large portion of of openAI as well.
Um, but yeah, I, it's.
It's funny seeing them publiclyacknowledge that their
competitor, after they've spentso much money, invested so much
money into them, for sure.
Eyvonne (35:17):
Well, I think you know, in any time you you work on the
vendor side in in technology,you find that there's often a
frenemy situation that comes upthat in some ways we work
together and in some ways wecompete with one another and
trying to find that sweet spot.
For how do we leverage partnersand other vendors and work
(35:43):
together to create a solutionthat's mutually beneficial for
everybody, but not let thebalance of power get so out of
whack that one person isbenefiting more, one company is
benefiting more than another oneor ends up being in a predatory
kind of a situation.
It is a really difficult lineto walk.
(36:04):
I think Microsoft'srelationship with OpenAI has
always been a little different.
You know, we had that incidentseveral months ago where OpenAI
the board fired Sam Altmanwithout Microsoft's input and
then Satya jumped in andreversed that decision, but then
we found Microsoft ended upwith just a non-voting position
(36:30):
on the board and now they're nolonger on the board.
I mean, I think you know OpenAIis on the ascendancy and
they're using the relationshipsthat they can, at whatever stage
that they're in, to moveforward, and then they're going
to move on, and so I think itmakes sense that Microsoft says
(36:53):
that they're a competitor.
Google definitely sees them as acompetitor and it's going to be
interesting as AI shakes out,especially as we head into the
trough of disillusionment.
I think we're pretty much,we're almost there, right?
What happens as we emerge fromthat into the slope of
enlightenment, you know, andthat whole hype cycle and AI
(37:13):
actually gets implemented and isused more broadly?
Or is it going to be asituation where you have a few
core companies that really useit and they become the control
centers for AI?
I don't think we know howthat's going to shake out yet,
but it's an interesting.
(37:35):
It will be interesting to watchover the next few years.
William (37:39):
That's a good point.
I think what you're saying Imean what you're saying is kind
of a feather in the cap for howfast.
I think a lot of these thingsare results of how fast this
technology is moving.
You know, it is just changingfaster than anything I've ever
seen in my life.
Eyvonne (37:59):
Yeah, and I think usually it's years to go through
that whole process for anytechnology.
And I think what that means isthat you know, like, as we've
compressed the ends of theheight cycle, we've increased
the frequency or increased thepeaks of the curve Right, and so
you know, and I think you know,it's also going to be
(38:20):
interesting to watch as themarket responds to that.
So, you know, I think we've gotinteresting days ahead.
As far as, like the Microsoftoutage that you first started
with William it, you know, thisarticle indicates it was a DDoS
attack.
Ddos attacks are incrediblydifficult.
(38:43):
I think you know we saw, youknow, google Cloud defend
successfully against a 398million packet per second attack
several years ago, and it'sgoing to be an increasing
challenge.
I know Russ White, over on thehedge, has been on a
centralization kick.
I think we're you know, and Ithink we're going to see some
(39:04):
move to edge.
I think we're you know and Ithink we're going to see some
move to edge.
I think we're always going tohave centralization, though, and
we're going to have to figureout how to better defend and,
yes, it's been a really roughmonth for Microsoft.
Alex (39:20):
Yeah, that article.
It's funny because it says ittook down Minecraft as well.
Eyvonne (39:26):
So just want to point that out Elementary age children
were distressed everywhere.
William (39:32):
Many, many tears were shed in this outage by many age
groups.
Alex (39:37):
Yeah, but also real quick.
There's in this article.
It says Professor Alan Woodwardsaid you'd expect Microsoft's
network infrastructure to bebomb proof.
That's I.
Just why would you say thatthat's such a?
I don't know who the professoris, but that's a very weird
(39:57):
statement to make.
Eyvonne (39:58):
Show me someone who has never run an operational
network.
Exactly Without showing mesomeone who's never run an
operational network.
Yeah.
William (40:09):
The thing we work in called technology is apparently
fickle, and we're sitting atopprotocols that were invented in
the 1980s 80s really and 90sthat we rely on predominantly
for all of this stuff.
All right Well I guess it'sabout time um, I guess everybody
(40:31):
knows where to find you.
Alex, do you want to just throwyour social handles anyway?
Alex (40:38):
yeah, uh, just at bumps in the wire.
Um, I don't really post toomuch on twitter x, whatever you
want to call it but veryoccasionally I'll just write
some random, something random,linkedin I'm always on LinkedIn
or X.
So if you want to reach out tome, feel free and I'm sure
you'll get a response veryquickly.
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. Special Summer Offer: Exclusively on Apple Podcasts, try our Dateline Premium subscription completely free for one month! With Dateline Premium, you get every episode ad-free plus exclusive bonus content.
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!