Ep 20 - Cloud Costs and Values for Leaders with Eyvonne Sharp

Chris Miles: 0:01

Welcome to the Cables to Clouds podcast. Cloud adoption is on the rise and many network infrastructure professionals are being asked to adopt a hybrid approach as individuals who have already started this journey. We would like to empower those professionals with the tools and the knowledge to bridge the gap.

Alex Perkins: 0:31

Hello and welcome back to the Cables to Clouds podcast. Today's episode we're going to be talking about cloud cost and values for leaders and decision makers, and we have a returning guest this time, Yvonne Sharp. So before we dive into the episode, let's cover our bi-weekly news section. Chris, you want to go ahead and kick us off here?

Chris Miles: 0:57

Sure.

Chris Miles: 0:57

So, yeah, this first step in the category of AI news, which is, you know we've said this is going to be a recurring thing, because there's always something every, every, especially every day, let alone every two weeks Quick one here.

Chris Miles: 1:08

So there was an article from SDX Central that I thought was interesting, talking about how LLMs have this multilingual jailbreak problem. So basically, I guess it's a tongue twister. Basically, the summary of the article is that the obviously all the safety that's built into these LLMs to make sure that people aren't doing, you know, bad things with them, essentially is all based on the English language. So it was very easy to kind of get around these things using non-English mechanisms, right. So some of the researchers I guess that were involved have kind of proposed this new framework, security framework, to try to protect against these you know nefarious things and you know, potentially, we mentioned, like you know, if they could use the AI itself to translate it and then build the security with the translation. But I'm sure, tim, you might have some comments on that, as you've seen how well well these things do with translating. You know Japanese and things like that, and in the article they specifically call out you know these kind of like Eastern languages, right, and like Chinese and Mandarin and things like that.

Tim McConnaughy: 2:21

I've been using it off and on as I've been studying Japanese it is it will confidently tell you, of course, the absolute wrong thing. In fact, there's been many times when I'm like, are you sure, I guess. I'm like I don't think it actually works that way and oh yeah, you're absolutely right. So I can't wait until we can build our security frameworks using this type of ability.

Chris Miles: 2:43

So, yeah, I don't know how much confidence I have in that, but you know, it's a pretty simple problem to solve. I think just do it in other languages. Right, it's definitely doable. Yeah, what's next, alex?

Alex Perkins: 2:55

Yeah, so we got a couple. There were quite a few conferences this week, of course, a lot of AI stuff, like Chris mentioned. So there was GitHub Universe 2023. The big takeaway from this is really that they're set tag. They had some tagline that was basically GitHub was originally founded on Git and now it's being refounded on Copilot, which is kind of strange because, you know, copilot was basically built off of what people were putting into it, so it's I don't know. I see lots of memes and jokes about how you're paying for the privilege of reusing what you helped it build. So I thought it was kind of funny, yeah, and then we had this. This is the big one. This is a lot of crazy announcements here. Openai had a dev day. Big thing that stuck out to me is they said they have 100 million weekly users. That number to me is just nuts, crazy, like that's so many people and it's just continues to blow up and get bigger and bigger. What do you guys think about that number?

Chris Miles: 4:04

Yeah, I'd say I definitely expected it, to be honest. I mean it's like as much as we see it popping up in the news. You know that makes sense that it's permeating out to the non-tech verticals very quickly, right? Because I mean that's what it's used for, right? I mean we talked about this earlier. Like you have, you know your kids are using it to for the schoolwork and things like that, so it's yeah.

Tim McConnaughy: 4:27

I'm not too surprised. No, I was going to say that's the same thing Actually. 100 million actually is not that surprising to no weekly. Of course that's a weekly number. That's where the real that's where the iPop comes from, right, Is that? But yeah, I mean, it is very quickly. Ai is becoming very quickly, ubiquitous almost, and of course open AI, chat, GBT, you know kind of the first mover advantage there. So yeah, that doesn't surprise me, but I mean it's a big number.

Alex Perkins: 4:57

And then so the other thing. This is what really blew my mind. I think I was telling you guys this earlier, but I've been kind of, like you know, I've been paying attention to AI and it's kind of been in the background. I'm like kind of like, is it really that big of a deal? But they announced custom GPTs and this just seeing the demo of this really blew my mind and made me think maybe there's a lot more going on to this than I thought there was, and essentially what these are.

Alex Perkins: 5:24

The demo that they did on stage was like Zapier, right, which is like a home automation software. These, or just people, can build their own model that you can just like type into and it will do everything for you on the back end. So, like a simple example is you know you have smart blinds, right, and you're like, hey, I want to set up an automation that every morning, when I wake up, my blinds open for me and you just type it into GPT and it will, like you know, touch all your devices and like figure everything out for you. It's like a, it's like a turning GPT into like an AI assistant, but it's all companies can build it out for you and then you just interface directly, like you would with with chat GPT. So I think the possibilities here are insane and endless.

Tim McConnaughy: 6:14

So I got front end. Basically, what it is is like a custom front end that anyone can kind of interact with. They don't need to know how the code works. You know you could like build whatever you want and just talk to the app and you know, based on these GPT, I guess personas or whatever they can interact with the back end and actually do the coding essentially for you to make that work. That's pretty cool.

Chris Miles: 6:38

Yeah, to me this just seems like it's the next iteration of things actually being intent based. I feel like we've been hearing this phrase intent based for like years now, and like the name one product where you can literally just specify your intent in this space. It actually does what you want it to do. This is like that actually coming to fruition. So this is, I think it's pretty cool.

Alex Perkins: 7:01

As part of this demo, sam Altman, who's the CEO. He built a custom GPT with his voice in like five minutes, so it's like it's not even hard for these companies to build these things either. So I don't know. Lots of cool stuff to watch out for. Tim. What do you got for us next?

Tim McConnaughy: 7:22

So from our friend Duane Lightfoot over at AWS. Actually, he put out a blog on the AWS community site about network engineers getting acquainted with generative AI. So this is topical If you're a network engineer, if you're trying to think like how the hell do I start to get involved with this? We've got the link, of course, in the news, the show notes. So, yeah, take a look at that article. He lays it out really well and it gets yourself started and let's see, and the next one, the next one's a big one, big one, big one. So President Biden signed an executive order, basically on how to get the government to start governing AI. I guess there's a lot to unpack here. How big is the EO? It's like 20-something pages or something.

Alex Perkins: 8:11

It's huge. Yeah, it's very long In depth.

Tim McConnaughy: 8:13

Yeah, and I'll ask you guys opinion of it in a second. There's some of the highlights. It's very, very high level because you can tell that the people writing it don't really know enough about the technology, so it's very wide, and that's actually one of the criticisms that's already started to come out about this order from the open source AI community. But, yeah, it casts a really wide net, but it basically says the government's role in this is to make sure that American privacy and security and safety and everything is thought of and enforced when these companies are going out there and building AI. So you guys have looked through it too. Alex, chris, what do you guys think?

Alex Perkins: 8:56

There's a lot here and I have not read the whole thing and I don't pretend to be an AI expert or legal expert or anything like that, but, man, obviously the government had to do something. So I think this is definitely a good start, but my concern is, like Tim said, who in the government actually has the knowledge to be able to speak to these things and say what kind of guardrails should be in place? I've seen a bunch of stuff on Twitter. There was some giant letter that was written out by a bunch of the open source AI community, like you were talking about, sam and they're opposing this, and I haven't read that either, so I don't want to throw my thoughts out there about it. I just know that this is an explosive thing that's out there right now, and I'm sure there's lots of sides and debates to this.

Chris Miles: 9:49

Yeah, there's plenty of pieces in here that are really good. You'll read it and be like, yeah, obviously it should behave that way, but I just do not know how you don't. How do you enforce half of this stuff, especially without the heavy hand being a government? I'm sure thisnot only the open source. People are probably feeling a little attacked. I'm sure people with the tinfoil hat people are also very upset by this. They don't want government playing in this field at all. So it's yeah, I just like, Like I said, I agree with everything you guys said. But in addition to that, how the hell do they enforce this on anyone without being way more involved than they should be?

Tim McConnaughy: 10:34

I'm suddenly actually seeing Now you mentioned the tinfoil hat people, chris, I'm actually envisioning now the meme with the guy with the two buttons, and one button in this case would be like you know Skynet's going to destroy us all. The other would be like the government has to stop there. The government's taking my privacy away.

Chris Miles: 10:53

Yeah, lose, lose, all right. And next up we have a report from Gartner saying that cloud concentration has now made its way into the I guess a significant emerging risk for many organizations. So there's a list that they have in here of the significant risks, or the top five emerging. You know AI is up there, but now we've seen the cloud concentration piece come into this as well and really, just talking about it, I mean to me this just sounds like vendor lock-in but like at a cloud concentration perspective, right, because you're.

Chris Miles: 11:30

Ultimately, it's the same argument that you'd have, you know, if you're all in on AWS and you know there's points in here that you have such a high dependency on them.

Chris Miles: 11:40

You know you're involved in their blast radius, right, because you inherit that from them, right? And then you know there's this other comment about you know the regulatory compliance issues, you know per geographical area and things like that. So, yeah, I don't know. I mean this just sounds like vendor lock-in to me and while I definitely agree with that, but you know, I think vendor lock-in always comes with that you know the grain of salt, because it's the alternative to it is commonly not as any better than being locked in on one vendor and plus these vendors in this scenario have gotten to the scale where they're providing redundancy and things like that at a level that no enterprise would ever do right? So while, yes, usc is one going down, still fucks a lot of shit up, but you know, I feel like I was just thinking that, yeah, but at the same time, like I don't know, I don't know exactly how I feel about it, but that's how it came across to me.

Chris Miles: 12:45

I don't know about you guys.

Tim McConnaughy: 12:47

Well, we talked all the way at the beginning. We talked about some of the reasons you used the cloud and you know multi-cloud DR was something we specifically said was not something that was on the list of reasons that you would do it right. Yeah right, the idea of doubling your infrastructure and trying to build everything build everything in ABS and figure out how to build the same thing in Azure and hope all the services match up and everything is ridiculous. So yeah, I totally agree that that's Anyway. What do you think, alex?

Alex Perkins: 13:13

Yeah, I mean I agree with you guys and I think it does say I think there is a part in the article where it does say vendor lock-in right. So I agree with you, Chris, that that's exactly what this sounds like, and we did a whole episode right around the importance of multi-cloud. So I think we did talk about this quite a bit and howwhat ways around that and, like you said, it's so redundant, Like you can just design around a lot of failures. In the first place, You'll pay for it.

Eyvonne Sharp: 13:41

I mean, there's some things you can't avoid, you'll pay for it, but you can design.

Chris Miles: 13:44

You're going to pay for it, though, yeah. Yeah yeah, cool, yeah, all right. What do we have up next, alex?

Alex Perkins: 13:52

So this one's pretty cool for me. I love I follow a lot of the stuff that Cilium has been doing the parent company's Isovalent right. So what stands out about this? Isso Cilium, justthey're coming out with a new Cilium-certified associate certification. The thing that stands out most to me about this, though, is that it's comingit's not from Isovalent. It's actually from the Linux Foundation, which is like very interesting. It even writes in the announcement article that they're not the ones that came up with the questions or the blueprint or anything.

Alex Perkins: 14:27

It's like a bunch of different vendors that use Cilium, and, for anyone that doesn't know, cilium is basically the de facto like CNI container networking interface. So it's likeit's the networking piece of Kubernetes, and it's been adopted by pretty much every major Kubernetes distro that there is right. Like AWS, their Kubernetes service uses it, azure, gcp they all use it. So I just thought it was really cool that it's coming out in 2024. It's definitely something to look out for, and this article has links to ISOvailin's like map, with all their labs and everything that walks you through and teaches you all this stuff from the very basics. So maybe a new time to CCNA for the future. I don't know what you guys think about it.

Chris Miles: 15:16

Maybe, I think, the biggest benefit to the CCNA at least. Obviously it's been around for God with 30, 40 years at this point it almost seems like. But the CCNA at least has this fundamental piece that's built into it, like all the networking fundamentals that you can, even if you're not gonna work on Cisco, there's still a great deal of, or at least there used to be. I haven't taken the recent CCNA, but there was at least a great deal of fundamentals that you get to learn from that, just on how, even how a computer works in some capacity from a networking perspective. This one, I think, has a lot of implied knowledge that you're supposed to walk in with, at least be familiar with Kubernetes and have some hands-on experience with stuff like that.

Chris Miles: 16:00

Yeah, that's fair, I mean. But who knows? I mean with the same conversation people would be like so I get my CCNA or should I learn Python? Kind of do both. Right, Maybe we're at this point where it's like do I get into cloud or do I learn Kubernetes? Like maybe it's both. So I don't know, you might be honest something.

Tim McConnaughy: 16:20

I imagine it's pretty lightweight. Honestly, I bet if you I mean I, having said that of course I haven't taken it, but I would assume it's just not that I don't feel like it's not like the depth of a CCNA, for example, or something that's really basic about just like here's all of networking it's going to be. Here's how Sillium works, which is pretty small comparatively.

Alex Perkins: 16:43

And just I think it does say, prerequisite knowledge is like, I think it's the KCNA level, which is like the very entry level of. Kubernetes certification. So you're right, Chris, you do need to know a little bit.

Tim McConnaughy: 16:57

Yeah, I'm going to check it out. It actually sounds pretty cool.

Chris Miles: 17:00

Yeah, you can sign up to get notified once it's actually available. I don't think it's not released yet. I think it's coming, I think in the next few months or something like that 2024.

Eyvonne Sharp: 17:09

Yeah, okay.

Chris Miles: 17:09

You can there's. I had a look at the curriculum. The curriculum looks good at least for what's published so far Pretty evenly distributed across all the networking and security kind of frameworks that they want to talk about. So yeah, I'm excited for it. I signed up to get notified, so we'll see.

Alex Perkins: 17:27

Yes, maybe we'll all take it and report back.

Chris Miles: 17:29

There you go. Whoever gets the lowest score has to buy something for everyone else.

Alex Perkins: 17:34

Oh God, I'm not playing that game again. Don't let Tim yeah.

Tim McConnaughy: 17:40

Never again.

Chris Miles: 17:42

I got tricked once.

Tim McConnaughy: 17:42

I'm not doing it again.

Chris Miles: 17:44

If you ever see Tim, a person ask him how high he scored on the CCNA collaboration because he did great.

Tim McConnaughy: 17:49

Actually it was the CCIE collaboration.

Chris Miles: 17:52

Oh, it was the real one. That's what we all took.

Alex Perkins: 17:55

This kind of goes right into the next point, which was KubeCon. Well, kubecon NA was this last week as well. It's very quiet news cycle for this year from KubeCon and I don't know if it's. You know, I talked earlier. There was the two big like AI conferences, there was GitHub Universe and there was the OpenAI Dev Day. So I don't know if it was just overshadowed, but I personally I think a lot of this is that Kubernetes is becoming kind of just boring a little. It's still complicated to use and implement and work on, but I think that it's just become such an entrenched part of, like cloud infrastructure. I think it's more about what people are building on top of it and not so much how to build it or use it anymore. What do you guys think?

Chris Miles: 18:47

Yeah, pretty much like you said. I think Kubernetes is getting to the point where it's almost what do you say run rate. At this point it's like it's no longer the cool kid on the block it's supposed to be. It's like it's at the point where you should have been doing it for quite a while now, not to say that it's at that adoption rate, but yeah, I think it's just. Yeah, there's not a lot of coverage yet. I mean, I've seen some kind of news here and there, like you know what happened at KubeCon, but like I haven't seen any major announcements so far, I think that maybe it should be a victim to the cool kid thing, but I'm sure there was some cool stuff. I don't know about what you saw, tim.

Tim McConnaughy: 19:30

No, I totally agree, and I think that Kubernetes is now like mature enough that you're not going to see huge, amazing, like earth-chattering innovations that need to come out of it anymore. It's performing its function pretty well, so that doesn't really surprise me, although I mean we have KubeCon, so I mean obviously people aren't sitting around twiddling their thumbs while they're there, I assume. So obviously there are talks and stuff, but I'm curious to see coming out of there like what's the, if it's not about innovations and new features and all that like what's new? What's interesting about you know, kubernetes moving forward?

Alex Perkins: 20:07

All right. So, tim, what do we got left?

Tim McConnaughy: 20:10

Oh, yeah, of course. So let's see, we have the AWS Fault Injection Simulator is a offer by AWS. It's free and it's a very small article. There's actually very little to say about it. But if you can think of, like Netflix's chaos, monkey type of thing, it's a little bit like that where you can kind of schedule this thing to come in and start breaking stuff and essentially and see how your environment reacts. Obviously you probably don't want to necessarily roll that straight to production, but probably really good for testing failover scenarios, testing fault scenarios. Just, you know, in your lab, spin up your lab, that's just like production run all your test simulations, make you know, write it down the results and then make changes to your production, hopefully based on what comes out of that. Yeah, it's pretty, I mean it's.

Chris Miles: 21:02

I'm just going to comment on this. It seems like AWS is releasing a lot of really short articles around this time. I don't know if it's just to like not put out things, save it up for rebate teasers.

Eyvonne Sharp: 21:13

I think so.

Chris Miles: 21:13

Yeah, so maybe, again, maybe this will lead into something bigger at reinvent. But I think that, yeah, it's kind of like the chaos monkey thing, but like I don't know, look, I will not complain to be or complain, I will not admit to being a compute expert or anything like that. But one of the examples they give here in this kind of a fault injection simulator experiment thing is that you can, you know, increase CPU on your instances up to, you know, 90 to 100%. How are they getting access to that? I don't know how that works. Like is that? Is that like concerning to people? Are they? Are they getting CPU level access on things that you're running?

Alex Perkins: 21:55

You signed the Euroma, come on.

Chris Miles: 21:57

Oh, that's true. Yeah, that's right. Maybe it's in there.

Tim McConnaughy: 22:02

Well, there is a, you know, there is a package that you can, that you can yourself install on your box to tax it. So I assume that's something like that where on the back end. They can like install the package and then start running it, but yeah, yeah, so it probably requires an agent.

Chris Miles: 22:16

Yeah, okay, makes sense.

Tim McConnaughy: 22:18

I'm sure, I'm sure it requires an agent. I'm sure, yeah, that's pretty cool though.

Chris Miles: 22:23

And you'll sign your life away when you install the agent, I'm sure.

Tim McConnaughy: 22:25

Oh well, they're going to collect all the telemetry and then use that to improve products and services.

Chris Miles: 22:30

They're building that, building that, building that model baby.

Tim McConnaughy: 22:33

All right, I got one more and then we got a fun one after that. I believe so Linkerdee, the Susie partner, is partnering sorry with Susie on Susie Linux. God, I've got me doing it now, before the show. Chris was like how do you pronounce it?

Tim McConnaughy: 22:50

How do you pronounce it? You're like it's Susie. It's like I keep hearing people calling Susie. He's got me doing it. Sorry, susie. Linux is partnering with Linkerdee to do edge computing like Kubernetes on the edge. So Susie has a lightweight Kubernetes package called K3S, which is again very lightweight, meant to run on OT networks. So the idea is that you can build tiny little worker nodes, essentially cube type environment, and then use this Linkerdee service mesh to connect to a control node elsewhere in the OT network that can do all the orchestration and stuff. So this is kind of interesting. So we've got edge computing and Kubernetes and service mesh and to me it's all. It makes sense, right? You look at an OT network and I'm like, oh yeah, that makes perfect sense. But I'm trying to think of other situations where something like that might make sense. What do you guys think about that?

Alex Perkins: 24:00

Yeah, I mean, there's a lot of use cases for this. Think of farming. You can have sensors for your crops. You need a way to put these. The thing is both K3S whatever you want to call it and Linkerdee are just super fast and lightweight, so they run on arm processors, so you can literally stand it up on a Raspberry Pi, and I think that's what the appeal is here is that they're both super fast, super lightweight and honestly, I'm kind of surprised it took this long for edge stuff to start kicking off. But I think this is really going to drive some more innovation and competition in the edge space.

Chris Miles: 24:43

Yeah, for sure. I mean, we remember the rumblings about IoT and everything just a few years back. So I'm with you, alex, I'm surprised it took this long, but it just makes more sense that the lightweight compute element that is now being used for modern applications is making its way to the edge. So, yeah, nothing too surprising here, nothing huge in this announcement, just that it's coming. I guess We'll have to see what comes, but yeah, it should be pretty cool. Might be a game changer for that space.

Tim McConnaughy: 25:13

Yep, and speaking of game changers, we're going to round out this week with an amazing announcement Chris, go ahead and bring us home.

Chris Miles: 25:22

Yeah. So, Alex, you sent this article to us and I almost like put it in a bucket that we called cringe news Not that it's I don't know what to think about it yet. It's crazy. So there's this company called Humane that is basically to develop this wearable technology for, called an AI pin. It's like a little thing that you clip on your lapel, has built in camera, like AI processing chip, and it's backed by this whole cell network that is built to build all this AI stuff for you. But it's just like. It's really strange. It's like it projects the display on your hand so you can see the weather thing, I will say it's kind of cringe.

Chris Miles: 26:12

You can watch this, the introduction video that they have on their website. You can tell that this it's a married couple, their ex Apple, and apparently they worked potentially with Johnny Ive and there's no surprise to that whatsoever. It is. It looks like an Apple promo, like through and through and oh my gosh. And then Alex also. You found that New York Times article where they're talking about some Buddhist monk was basically telling them that the company was going to be amazing. I don't know if it's a bit, I genuinely don't know. Maybe they have gotten me and I'm just like it's crazy, dude, it's the Silicon Valley bit.

Alex Perkins: 26:55

It looks real.

Chris Miles: 26:56

If you've ever watched Silicon.

Eyvonne Sharp: 26:57

Valley, the Hooli thing.

Chris Miles: 26:58

It's exactly that. It's crazy, dude. I don't know how do you guys feel about it.

Tim McConnaughy: 27:06

I speechless, I can't even, I don't.

Alex Perkins: 27:11

It's like what's the? I think it was the movie grandma's boy. Have you guys seen that?

Tim McConnaughy: 27:15

Oh, I love that movie so much too, Isn't?

Alex Perkins: 27:17

it when Steve Martin is like the dude with long hair and he's like sitting cross legged on the table and you like lighting up incense and stuff and he's like I was swimming this morning with the dolphins.

Tim McConnaughy: 27:28

That was Kevin.

Chris Miles: 27:29

Nealon. That was Kevin Nealon. There's Kevin Nealon, okay.

Tim McConnaughy: 27:31

It was Kevin Nealon, but yeah, he's the head of the software company or whatever. Right, he's like super hippie.

Alex Perkins: 27:39

Yeah, that's exactly what this, this New York Times article that Chris is talking about, was reminding me of. It's like they they found a Buddhist monk named brother spirit through their acupuncturist who introduced them to like Mark Benny off, and it's like they were at his house and they were like here's two products this one, this is the one that's gonna be the the big one, craziest story ever.

Chris Miles: 28:06

I'm amazed wearables are coming back man.

Tim McConnaughy: 28:08

I you know what with AI. Like they're gonna want to put AI assistants and think about it, like about the, the progression there, right, like it's all cyclicals. Like okay, now we got AI, ai can be an AI assistant what? But what's a better assistant than a wearable? Like that's what they're.

Chris Miles: 28:22

Yeah, I mean this is obviously like there's there's two ways this can go. It's almost like funny and and cringe, but at the same time it's totally like a black mirror moment. To me, it's like what, what bad could come from yes. I think could be Far worse, but I don't know. We'll see what happens.

Alex Perkins: 28:40

I saw like one of the best Comments or arguments I saw for this, though, which was surprising is so it costs $700, but then it's also I think it's like $24 a month subscription, yeah, which is like that sucks. It's a subscription model, but I saw a point that was really good. That it's it's it's got its own number, it's so it's like cheaper than paying for a cell phone, but it gives you the cell phone number. So I don't know that was a cool like. If you're gonna do it anyway, that was the argument is that you're gonna pay for a phone. Why not just pay for this and you get all the extra features? So that should wrap up all the news. I know we went a little long, so let's, we're gonna go ahead and get into the episode again.

Alex Perkins: 29:24

We had a guest on, yvonne Sharp returning guest, and this time we talked about cloud cost and values for leaders and decision-makers. You guys have anything to say about this episode? I thought it was. It was really good. Yvonne is always, you know, I enjoyed a talk to. She's got lots of insight. She is Constantly in the room with with the people that make these decisions, so she had she had a lot to say, tim Chris.

Tim McConnaughy: 29:50

Yeah, I mean I, I couldn't say it better, right? It's always a delight having Yvonne on to talk with us. She, could you just she just talk about anything. She can open the phone book and start talking, probably just as just as excited. But no, I mean she. It's always good to hear her insights and I'm this is. This is a great episode. This is good stuff.

Chris Miles: 30:12

Yeah, absolutely, and just a plug timing. Oddly enough she, I think, just put out a blog on like the packet pushers network about what business outcomes mean to it and there's a lot of overlap with that. So I feel like you know, listen this episode here, what she has to say, and also read that article. It's, it's really good. But yeah, it was a great conversation, as always with Yvonne.

Alex Perkins: 30:35

Awesome, all right. Well, let's get to the episode.

Tim McConnaughy: 30:42

Hello and welcome back to the cables to clouds podcasts. I'm your host this week, tim, and we have a very special guests with us this week, a returning guests, yvonne. Go ahead and reintroduce yourself, please.

Eyvonne Sharp: 30:57

Yeah, sure, I'm Yvonne sharp. I am a technical lead in customer engineer with Google cloud and you may know me on Twitter, or the platform formerly known as Twitter, as sharp network.

Tim McConnaughy: 31:10

It will always be Twitter. I don't care how many times he renames it.

Eyvonne Sharp: 31:14

Yeah, we're not saying anything. Yeah, that's yeah.

Tim McConnaughy: 31:17

Okay, so this is. This is cool because we released an episode of God I guess would have been a couple months ago now episode 12, which is where are the cost savings, where we were talking about things. Basically we tackled things like where's all the savings from the economy of scale, and Yvonne listened to that episode and she had some, some points that she wanted to make, and we really like having Yvonne on the show, so this was not a hard sell. So we would love To to talk about this with Yvonne. She had, she brought some great points and I'll actually let let her start making them.

Eyvonne Sharp: 31:55

Sure, yeah, I was. I was listening to that episode and I think one of the things that's helpful, as you know, as infrastructure folks who are working in the network day-to-day, we we always think through the lens of technology and and in the last few years I've had the opportunity to be in the room where decisions about cloud are being made Are we gonna go to cloud, what workloads move to cloud, and how do we think about what cloud does for us and Whether or not to make that decision? And I think it's important that when customers decide to go to cloud, it's not purely a dollar-for-dollar cost Conversation. So I'd love to dig into that and the conversation that happens in the room when folks are making decisions and Start there.

Tim McConnaughy: 32:39

That's a great place to start. Actually, I think we I mean and to start us off, I think we all agree that it moving to cloud is definitely not a strictly cost conversation. I mean, if it was so, many organizations probably just wouldn't move to, you know, wouldn't go to cloud in the first place. So but, but yeah, I think that's really good starting point.

Eyvonne Sharp: 33:00

Yeah, I think one of the things that happens of course on infrastructure teams for the most part, you have a budget and that and we've we've Kind of learned to think about that budget in a very particular way that you have a three year or five year, seven year cycle, whatever that is, and you, you buy your hardware, you and you buy as much infrastructure as you think you're gonna need for a long period of time, and then After that point, like if you need another VM, it's kind of free, like you're not thinking about the cost per VM. But there are other challenges that come in as you're, as you're deploying in that kind of a model, and that is what if all of a sudden, we need a ton of capacity that that we don't have, or what if there's a new capability, or what if we need a new app that requires, you know, gpus to use the hard to get hardware the day, and and those become big problems. And so I think part of what happens when, as as a networking person, infrastructure person, we look at, okay, well, I can run a VM Cheaper on-prem, then I can run One in the public cloud, I think what happens when we do that is we're missing a big part of the reason why, and Wanted to, you know, explore that a bit. Talk through the conversations that happened in the room.

Eyvonne Sharp: 34:23

When those decisions are made, I, I guess in the back of my mind, I'm thinking of a former co-worker I were I worked with, who was, you know, always the curmudgeon, and everybody who made any decisions was always Clueless. They, you know, they, just you know, they were the stereotypical Dilbert management boss, and I think so. He was always like grousing. I don't know why we did it this way and it doesn't make any sense to do it this way, and who decided to do this? And I, I can, I can just hear his voice in the back of my mind as we have some these conversations and there there are good reasons that are bigger and more holistic, when, when we're talking about, you know, public cloud.

Alex Perkins: 35:07

Well, who are the types of people I guess that you see come into these conversations, right, because it's different. I know you've been in a lot of different spaces over the years, so, like I know, you worked for VMware for a while, right? So if you had on-prem Discussions about this kind of stuff, are the people the same, like the key players that make these decisions? Are they the same as on-prem versus the people that are coming in to decide these things in the cloud as well?

Eyvonne Sharp: 35:33

Well, I think it really depends on the organization, but but the personas are different. Sometimes you'll have a customer that really wants to evacuate a data center and very often those decision-makers are pretty much the same because they're the folks who run and manage it on the data centers, and so what? Their concerns and questions are different than, say, a line of business leader. But often you'll have a line of business leader. That's that's in healthcare, for example, we have a problem in our ICU units with sepsis, right? Is there something you can do to help us detect sepsis cases? Right? And that's a very different conversation. Then how do I, how do I move my data center right?

Eyvonne Sharp: 36:21

That's, that's a very you know, business focused conversation, and so a Lot, of, a lot of times, I think it's easy, in the day-to-day nuts and bolts of making the technology work, to lose sight of what the technology, what real business problems, the technologies there to solve, and so, and that could be, you know, and, and sometimes those problems are more technical, or sometimes they're less technical. I mean, we have an entire universe of a video library If you're talking about a media entertainment company, right, and we want to be able to find every the Barbie movie releases, and we want to be able to find every video or every picture that we've got that has pink lipstick in it. Right, we may not want to brand it Barbie because it's going to cost us money, but we would love to update our socials with pictures that are pink. Can you help me find that Right? And so that's a very different conversation.

Eyvonne Sharp: 37:16

Then, how many VMs do I need? Or you know, are we gonna run, are we gonna run on GKE? Are we gonna run OpenStack, on-prem, or are we gonna run, you know what? What technical platform? And so I think it's it's always important to tie back to you, okay, so what technical problem is this trying to solve? And then what is the business getting out of it?

Eyvonne Sharp: 37:41

as opposed to what does it take? You know what's, what's the price per VM, for example. I keep going back to that, but it's an easy number.

Chris Miles: 37:48

I'm curious so, like in your position as of today, are you more so having the conversation around Expanding capabilities, like adding new capabilities, like like the ones you just mentioned, or optimizing infrastructure?

Eyvonne Sharp: 38:03

I think the customers that I see that are most successful when they're moving to cloud have a very clear idea of what it's going to do for their business. I think customers struck and and and that's, and it's fine if what it's doing for their business is it's, it's, it's, you know, eliminating the cost of maintaining a physical data center and dealing with the power and the, the site development and site management and cooling. And you know there's a lot beyond just the technology. There's a physical plant details, and if they know that's what they want to accomplish there, they're much more successful than you know. The customer that's like, hey, I've heard cloud is a thing and it's the new place to be and we should be there too. Right that that is almost a recipe for disaster. It's, and customers have done that. You know what, what's the new hotness, and we, we want to be in on that. And if they don't understand what what problem they're solving with it, that that's when they get into trouble throw everything into microservices and put them in the cloud, right.

Tim McConnaughy: 39:12

Everything is magical.

Tim McConnaughy: 39:14

Yeah well there was the gold rush of you know, everybody is like all the CIO's, the world got together and decided that they were all going to evacuate their data centers without any Consideration of what they were trying to evacuate, or, and all of that. Right, we've gone over that a hundred times. You touched on something I wanted to go back to just a little bit, which is this idea of you know, if you wanted to do agility on prem, you basically have to over build, like you have to, like you mentioned. Okay, well, the business all of a sudden, for whatever reason, needs a ton more capacity than sits in their data center today, and that's actually very difficult to scale up to or create more VMs out of thin air. Right, and you know we're talking about not just a cost discussion, but a business agility discussion.

Tim McConnaughy: 40:01

Now, this is interesting because, of course, that is a very technical discussion about alright, well, how many more VM capacity do we need? But, but we, as the technologists, tend to forget that. You know we just didn't decide out of nowhere that we need all this extra capacity. The business on the back end is doing something, that's that's asking us to create all this new capacity, and it's very easy for us to lose sight of that, that driving force, if you all the business reason to do it. But no, I just wanted to point out that just what you said you know it's difficult to do on prem and the agility of being able to suddenly, all of a sudden, you got the capacity you need to do that thing you're trying to do is is definitely one of the reasons why people make that decision to move their workloads or their technology to the cloud.

Eyvonne Sharp: 40:51

And from a technical standpoint, like you have to be at a certain size to be able to provide any kind of level of geographic redundancy. Right, if you know it's most smaller orgs it is. It is very capital intensive to build out to data centers and especially to that are not that are that are more than a hundred miles apart, at which really you just about have to have you know if you want to avoid any kind of major like geological weather event, and I think those are. Those are other like more technical important capabilities, but but honestly that I don't. I don't see customers choosing cloud just because of the regional capability. Usually that's an incredible bonus, it's something that the cloud has to have, but it's not it's typically not not the deciding factor.

Alex Perkins: 41:56

That I mean that's a, that's a really good one, because I mean I see it all the time with customers that I have nobody ever even talks about that as a benefit. Just having that available and then comparing the costs of what it is to spin up and maintain to entire geographically. You know disperse data centers and you know running between them as well, because I think all all the CSPs have some kind of like backbone. You know some kind of feature where you can ride on the backbone and get really low latency between them as well.

Eyvonne Sharp: 42:29

If you look at object storage, for example, everybody's got object storage. Ours is Google cloud storage. But you know it is a fraction of a cent per gigabyte to use a regional bucket or to use a multi region bucket or a dual region bucket than it is to use a single region bucket. You don't have to, there's no engineering effort there for you, right? And those those kinds of capabilities like it just takes a whole layer of complexity off the customers plate and that provides a ton of value. Because as much as we, you know, is networking going away? Is it is this going away or whatever that there are not enough people skilled enough to do the jobs that we need done? And so I feel like often those conversations are almost a red herring, because everybody is starved for talent and, and the more we can do to bake that into the system that you know, the the the higher value work that folks can do. And those are all conversations that are had, you know, I think a lot of times were individual contributors.

Eyvonne Sharp: 43:41

We don't think of the I know it was the case of me, especially for the first 10 or 15 years of my career is like I'm worried about what I have to do and I have to get done, and how I can make a project successful. But there's a, there's another problem of like, how do we make sure we have all the people we need to do all the things that need to be done? And then there's also what exactly is it all those people should be doing? And then and then the layer of what's going on in the marketplace and what you know, how do we differentiate ourselves as an organization? Like there are these layers of problems to solve, and and we can get myopic on the technical problems and not see the bigger picture problems that the decision makers are really concerned about, and I think the thing that I've grown into lately those problems are no less real problems than how to architect our leaf spine fabric in our data centers, right, yeah, that's, that is a great point.

Chris Miles: 44:41

So back to back to what you said, like referencing your old coworker. You know the typical curmudgeon like how, how do we try to bridge that gap and and and help make sure everyone understands the common goal? Right, because I don't know. I see this plenty of times as well. You know you have that person that's arguing. You know why would we pay for this? You know why are we having talks about capacity? We've never had, you know, capacity concerns with our existing, you know, on prem environment. Is it? Is it? Do we think it's rooted in, like FUD, fear, uncertainty, of doubt, with with talks about certain aspects going away in the cloud? Or is it just that there's difficulty for you know operators and things like that to see the same? You know, 10 miles down the road, that the that the C level executives have.

Eyvonne Sharp: 45:35

Well, I think that problems can occur at every layer of of every layer in the organization, right, that you can have a problem from the C level down to the you know, vps of the senior directors, of the manager, like there are different problems that happen at each of those levels, and so, and there can be problems of multiple layers, right? I mean, you, some orgs run where a decision gets made at the top and that gets pushed down, and this is what we're doing and we're gonna do it this way and you, like you really don't get a vote that that is not a way to run a well functioning technology organization. Their talent, but your innovative people, that they may get some skills there, but eventually, if they don't have a voice, they're gonna go somewhere else, right, and so that's that's. One problem is that leadership is not transparent at all about what they're doing or why they're doing it. If you look at the, the other, from the individual contributor level, I think a lot of one of the things I've said this before, I'm gonna say it again that you know, if your company has an all hands where your top level executive leaders are speaking, you should go, you should listen and, and a lot of the and again one of my curmudgeon friends like that. They're not gonna say anything that matters to me. You know I'm, you know I'm gonna pull cable in the data center yesterday and I'm gonna pull it today and I'm gonna pull it tomorrow.

Eyvonne Sharp: 47:11

Well, but what those leadership conversations and what those all hands do if if they've done reasonably well at all is, first of all, you pick up the vocabulary what, what's the vocabulary are the leaders are using, and so you can talk about how your work fits in in their language. It also helps you understand what the business priorities are. It helps you see what your leaders care about. And, yes, at the end of the day, is there a huge difference in running a voice over IP phone system, you know, for a tech company or for a health care company or for manufacturing company maybe not because everybody's a phone system, but there may be some special capabilities that you need or that you are implementing, that can, that can help, or a new emerging technology that has an interesting use case for your company. And you need to understand what those priorities are because at some point you may be in an elevator or you may be, you know, in a lunch line or you may be part of a conversation somewhere when you can speak intelligently, and so it's it's always valuable to know and understand. So I think that's something that individual contributors can do is understand their leaders, listen to them, even when maybe You're not sure how it's relevant, relevant to your day to day.

Eyvonne Sharp: 48:43

I think the other thing is understand how your organization budgets and what they care about and how those numbers percolate up. You know, do we do? We is it. Is it CapEx heavy? Is it OpEx heavy? What? Which of those categories is most painful?

Eyvonne Sharp: 49:09

I've worked in companies where they would spend CapEx all day, every day. But if you were going to do, if the dollars were OpEx, there was a whole different level of scrutiny, and understanding those things makes a big difference in what you can buy when, how you present what you want to do. So that's that's also important, and I think all of those things become part of those high level conversations and part of the equation beyond what particular technology that we need. I mean, I've seen and this is certainly not unique to cloud, but but sometimes those financial drivers between how, how something can be financed and structured will radically impact the solutions the company will choose because they have financial constraints, like you know, just like you have a budget at home and you have financial constraints that detect, dictate, you know what model car you're going to get. So that's what I think is important for the company, and understanding those things matter as well.

Tim McConnaughy: 50:15

I always like to refer to the Venn diagram fast, easy and cheap pick two right Very much and that's the. And I think when you're talking about like business drivers to go to the cloud, you know fast, easy, cheap, it really is like pick two of those as well, right.

Eyvonne Sharp: 50:35

I keep wanting to believe that that paradigm is wrong, but I'm hard pressed to find a scenario where it is.

Tim McConnaughy: 50:42

Yeah, right, but I mean, those are the kind of drivers that the business has to figure out right, like do they value agility? Okay, well, what's the dollar value the agility is worth? Do they need a new technology? Because the competition is utilizing this new technology to get an edge on them? All right, well, what's that worth? And how much is in the war chest to try to catch up or overtake? You know, like those, I think those are the kind of conversations that are most interesting and that basically nobody outside of the C-suite is probably ever going to see.

Eyvonne Sharp: 51:12

I've been part of conversations with IT leaders where they're like you know they want to pull out like their hardware vendor costs, like this is what we paid Dell or EMC, which they're now the same thing in Cisco, and like here's what we paid those guys last year. Now here's the price you have to beat, and it's like that's not really how this works right Right.

Eyvonne Sharp: 51:40

Because you don't have a line item for power in your budget. You know you're not responsible for paying for those things. And I think and it gets challenging when you're having this very myopic conversation like, well, you know, this is how much our switch is cost, and this is how much our server's cost, and this is how much our licensing costs. Okay, Like that's the cost.

Eyvonne Sharp: 52:04

It's like, well, not, exactly Not exactly and and, but at the same time, it is unrealistic to be like well then you know we don't need all these people anymore. Well, that that equation doesn't work either, because things still have to be built and managed and operated and maintained and the hope is that you can have your folks doing more high value work and I believe that's often the case and you can take some things off off your plate. But you also you have to articulate that value more in what you're, what, what the people can accomplish, as opposed to in very rare situations I've seen customers like hey, we've got, you know, a hundred offshore people that are bandating this really awful, broken, hard process and we would like to eliminate that. Never now and then a situation comes up like that, but it's usually like that. It's very low value work. Any already that that have got people like pushing papers around or or checking boxes or, you know, alt tabbing through 30 different interfaces.

Tim McConnaughy: 53:22

Stuff that's hard to automate.

Chris Miles: 53:23

Yeah, yeah.

Eyvonne Sharp: 53:26

Um, so you know, I think it's just never. But I think the other thing too. I've I've worked with customers, um, especially in a former life, who would be like you know, they would switch an entire platform vendor for 3% cost savings and, like they believed their value, was choking out additional 3% by being a jerk to their vendor. Um and and that was a win. And what I will tell you is that once you throw in the disruption of that technology change and the migration costs and the retraining, training that you've not saved 3%, you know.

Eyvonne Sharp: 54:11

So there has to be a bigger reason to encourage a technology change than a few percentage points in price. It just doesn't make sense to do it for, for, for a couple of percentage points.

Tim McConnaughy: 54:25

Yeah, um, what you see this a lot, what I see this a lot is when the incentives of a business to the, to their you know whatever say their procurement or finance department and it's not solely that right, it's really incentives across the board, across the whole business when they are ones that are easy to quantify on paper but bring very little true value to the business, like that's what you that's. When you see this kind of like why would you, why would you do this type of stuff? Right, and the procurement people probably are either incentivized or metriced on being able to bring down vendor costs 3% and, of course, really the business is actually paying a shitload more money on the back end, like you pointed out. But on paper, hey, I saved the business 3%, that was my metric, right?

Eyvonne Sharp: 55:14

I have lots of thoughts Um none of them fully formed about, you know, misaligned incentives. And how do we, you know, yes, create incentives that are measurable, that give people a sense of individual, like performance and accomplishment, but also work toward a collective goal? It is almost, it's almost, like that is another one of those pick two um situations, because, um, you can create a system that really highly rewards individuals but also disincentivizes them from working as a team, and that is a net loss for the business, and this happens, as for individuals, it also happens for departments.

Eyvonne Sharp: 55:59

And so if, if you know the, the and, and it also depends on how an organization silos, because if you've got, you're like these are our on-prem data center people and these are our cloud people, and they're in different groups and they have different budgets, well, all of a sudden, your on-prem people are fighting your cloud people for workloads because they want business relevance as opposed to saying what's the right target for this workload. What problem are we trying to solve? You know, does this? Where does it make sense to run it?

Eyvonne Sharp: 56:33

You end up with with, yes, with this battle between teams because, um, well, everybody wants, you know it's, it's existential at that point, like we, we want to exist, and as opposed to creating an environment or, and and, frankly, like there is no perfect answer, because you have to divide people into teams and you have to organize them and manage them in small groups, because you just can't have a team of 50 people and then figuring out where to draw those lines and how to do it, I've come to appreciate those as very real and significant problems. Um, that are that are that require just as much skill and effort to solve as you know how to how to connect three clouds and with on-prem and make sure that the packets all get from point A to point B.

Alex Perkins: 57:24

Well, the hidden costs in in all this right, um, I've I've actually seen this quite a bit where, um, like, like you were talking about, a finance department will make a decision for something like a support contract and they don't talk to the people that now have to deal with this new support team and it it just causes all these hidden costs between these departments that nobody's thinking about. So it seems like it saved them money, but in the in the back end right, it's just causing all kinds of other issues and causing raising costs elsewhere, um, or like missed SLAs or you know like even not even financial costs. There's just other other costs that aren't taken into account when those decisions are made in the first place.

Eyvonne Sharp: 58:05

Yeah, and this is this is not even a cloud example, but but I had a in my in my past, I had a situation where we had a partner and and the partner had really, really gone out of their way to help us. They'd given us spare gear, they'd helped us build out a POC, they had provided, you know, um, customer engineers, sales engineers, who were incredibly helpful, um, and provided all kinds of really relevant guidance. And then, and so we, we made a product selection and then, when it went to procurement, they, they chose, like the bargain basement company, to procure for every time.

Eyvonne Sharp: 58:43

And our partner was like, really justifiably, livid, um, and, and it it became a thing right, and those problems don't go away. Um, like, the last thing you want is your FinOps team or your, your procurement department going well, why did you use this low balancer instead of that one right? Or to create, you know, circumstances in which you're choosing suboptimal architectures because of pricing? Now, that's, that's a. That's a different scenario. Um, but um, and I think that's why we see a lot more talk about FinOps. Um, and you know, because there is an art to managing cloud costs, but it's it's got to be holistic. It can't just, be well, this one's cheaper Now justify why you have to use the more expensive one. Um, it just shifts the conversation around a little bit.

Tim McConnaughy: 59:46

It reminds me a little bit of the uh and and not to not to plunge us into the political realm. That uh. It reminds me of the American uh health insurance system. You know, when you talk about medicine, the doctor prescribes you something and then uh, now you got to go justify that to the health insurance company while you need that medicine.

Eyvonne Sharp: 1:00:06

Yeah, yeah, and, and we can. We can set up systems where that wasn't intentional, but that is the ultimate outcome. Right, when the, the, the people who need to make the right decisions, um, for, you know, the health of a uh infrastructure, a health of the environment, um, uh, those decisions are called into question by tons of other people. Now here's the thing checks and balances are important, and I think our world is so complex now that, like, very rarely can a very small group of people make it, make a decision, especially if that small group of people has limited scope of understanding. But, um, also, we, we, we got to communicate and, and those, and, and, and, and. They are hard problems Like I don't. I don't think any of these have an answer. You got to know your business and you got to know what, what makes sense.

Tim McConnaughy: 1:00:56

Yeah, there's a lot of a lot of tribalism, a lot of walls to knock down, uh, to make the whole thing work really just as a company and, and especially when we're talking about teams, um, you know, making decisions about cloud. Just to bring it back to uh, to the topic, you know there's a lot of of tribalism and who makes decisions about what uh, who's relevant to what work? Fighting over workloads is a perfect example, right Like and, and somebody needs to be making those decisions in a way that, like you know, is best for the business but also takes all of the uh information that you know, both teams, that whoever's got input uh, you know has has to give and and some uh, you have to pay some attention to that.

Eyvonne Sharp: 1:01:45

Well, and one of the things that I've seen too is a lot, of, a lot of times we don't always think through, like, what value is being unlocked for the business? So so let's say you, you adopt a new technology public cloud or otherwise but what it does is it enables a team not just to uh, operate, operate better, but it enables an entire new line of business, or it gives the business insights that creates revenue right.

Eyvonne Sharp: 1:02:16

At that point. It's not just about the cost of whatever the thing is, it's how much additional revenue can, can we make? And I find that if, if you can make a really valid, clear, believable argument for those things, then for, for, for unlocking new business potential, then all of a sudden, like the infrastructure costs, it's just the cost of doing business, and I think sometimes uh and and, and it's easy to lose sight of that when your entire world is is operating, whatever that thing is, because a lot of the levers that you have control over are those day to day operational expenses.

Eyvonne Sharp: 1:03:00

But I also think it's important if, if another good thing to do is not just listen to in all hands, but you, if your company's public trade, like an earnings call, like if, if you can tie something that you were working on to, uh, something that comes up in an earning earnings call in a positive way, like that's, there's huge value there, and so so many of the customers that I see that are really successful. Successful in adopting cloud, it's because it's tied to that kind of an initiative. If it's something that's going to come up in a board meeting, it's a big deal. If you can tie it to major positive revenue changes, it's. It's. It's no longer just about, um, the comparison of costs Now, over time, as that, whatever that new thing is, as that becomes part of business as usual in the industry, of course they're going to optimize, they're going to squeeze, they're going to try and and get, because all of a sudden like okay, so we've got all that revenue, now we want incrementally more, right, but you've always got to be looking for that.

Eyvonne Sharp: 1:04:09

What's the next thing that's going to change? Um, the, the revenue curve. And then the technology is just not that hard to justify.

Alex Perkins: 1:04:21

Yeah, like the first mover advantage right Um especially if you're opening up something new that you know your competitors aren't doing. Yet you don't even know what the infrastructure is going to be like half the time, right, you're just. You just want to spin up things quickly. Um, I, I see that a lot too, Uh, as as reasonings for people moving.

Eyvonne Sharp: 1:04:40

Sure and um, you know, that's, that's a double-edged sword, right, I mean. But but at the end of the day, uh, often the you know we have to make it to market first is a big deal Um and it it can be um, you know, and if, if you're cutting corners on reliability or things like that early on, a lot of times the market will be more forgiving to. I mean, I think, think about the early days of social media.

Chris Miles: 1:05:14

Right.

Eyvonne Sharp: 1:05:15

I mean, how many outages did we have? Um? And yeah, they were um, they were a problem. But I also think people are like we're doing something new here, right, and it. It was understandable that you know we were still trying to figure out how to scale those systems. Now you know one of those marquee services. You know there was a big like S3 outage and it was several years ago now, but I mean it was, it was a thing, man, it was a big deal, it was a problem, um. And so I think, like the market can be more forgiving and if, if, something allows you to move more quickly and get the market faster, that's also a huge deal. It's just, it's just there's, it's, it's, it's. It's hard to quantify, but it's real, and those are also things we have to consider.

Alex Perkins: 1:06:08

But, yeah, and your industry matters a lot when it comes to that too, right, like there's some industries that you just you have to move faster than everyone else, and it's a difference between a lot of money, yeah.

Eyvonne Sharp: 1:06:20

And over time, as as capability stabilized and as the technology stabilized, I think I think we're almost going to see a um, I'm picturing sort of a kind of a wheel of technology where you know, you, you have a, you have a thing that's, that's new and hot and we need to be fast, and but in over time, that that thing is going to standardize and it and maybe it runs faster Runs somewhere different after a certain period of time. Right, because it's predictable. We know exactly what the, the, the workload profile is going to be. We know exactly what compute it's going to need. But then we have these other things, um, the. The example for that right now is is all these AI workloads? Right, everybody's constrained on capacity. From a hardware perspective, folks are still trying to figure out, like, what's a good thing for a language model to do, what's the thing that we shouldn't let a large language model do, because it's going to hallucinate and put somebody at risk.

Eyvonne Sharp: 1:07:17

Like and and those things, if you wanted to build out a platform to do that on premises and you don't have um data scientists, like it's just not something you can do, right. And and and so where does it make sense to really innovate, move fast? And and those are also part of that cloud conversation they're, they're value. That may be our. You know the, the folks that are um, managing and configuring our infrastructure right every day. Just don't see um, but they're, they're equally important, yeah.

Tim McConnaughy: 1:07:52

I think the models themselves will end up being a product. You know, not just like here's the compute to run your models, but I think the models themselves will definitely be a product, especially for people who can't create the data. Science Like and we talked about this with Peter Jones, we were talking about the whole AI revolution and LLMs and the networks we'd need to build for them and whatnot is that? Uh, you know the data that goes, I mean, let's, why did you know? Why did Cisco just buy Splunk? I mean, yeah, great, splunk's a household name, but I think they wanted the data. I think they wanted the models and the AI and the that. That was way more to them than than the name. So I mean, yeah, I completely agree.

Tim McConnaughy: 1:08:32

All right, uh, looks like we're coming up on time, so let's go ahead and wrap up. Uh, we always love having you on, yvonne, every you're, you're always welcome, truly, um, is there any uh closing uh thoughts? Anything we forgot to cover, anything you wanted to, any points you wanted to make that, uh, we didn't get the chance to hit yet.

Eyvonne Sharp: 1:08:53

Yeah, no, I think I think it's just um and and and you may stop having me on because I feel like I'm repeating myself often but I think it's so super important to to pull back from our day to day view of you, know our day jobs and understand the bigger picture of what the business is trying to accomplish, where we fit in that and then how to take what we do.

Eyvonne Sharp: 1:09:17

And I have a I have a peer who calls it a steel thread right To pull that steel thread from, from where we are, what we're building all the way up through the business, to understand how that helps our companies accomplish their mission. Um, for most of our orgs that's, there's a revenue target, there are shareholders, they're they're, they're those considerations. But even for nonprofits or service-based organizations, like, how does what you do align to the mission of the organization? And the more you understand that, the more you can make good decisions where you are. And also, maybe, um, hopefully, um understand the decisions that have been made and why certain platforms make sense or don't. And speak, speak that language, um, because it'll, it'll add value to you, to your organization, um, and it'll help you be more effective, which I think most of us, if we're listening to something like this and showing up like we want to be effective and we want to know how to do our jobs well and we want to know how to contribute where we are.

Tim McConnaughy: 1:10:18

So yeah, that's my stick, awesome yeah.

Chris Miles: 1:10:23

I think that I think that's important because, just that, that whole exercise of trying to see where you fit into the you know if you're an individual contributor, where you fit into that strategy, um, it's important to to make that mapping to see you know if that aligns with with what you even want to do, right. Um, sometimes that exercise can can end in you determining that you know what. Maybe maybe this organization isn't for me, right May, but you at least can can see how individual contributors can can map to that and and where you may fit in someone else, right, or at at somewhere else, right yeah.

Eyvonne Sharp: 1:10:58

And I think it can provide context to you know, is the what you say, chris? Is this the right place for me to be? Or maybe maybe it's okay for now, but maybe I need to start pivoting, you know, in this direction. Or maybe I need to gain these skills here, or maybe I need to to get to know somebody in this pure, you know, organization, to understand a little bit more about their, their, their world. I think if, if we pay attention to those kinds of things, um, we, we can add a lot of like intangible value. That's not going to be, you know that, that we can't get a cert, a cert for that there's not an exam that we can take, but, um, it's, it's still meaningful, right?

Tim McConnaughy: 1:11:41

Alex. Any closing thoughts?

Alex Perkins: 1:11:42

No, I mean, uh, I was going to say a lot of what Chris said, so, um, I will, I will echo that. Yeah, I mean I I think it's a great. It's practical advice, right, it's um, something that that's not talked about often is is looking at that as a kind of way to integrate yourself into a company and advance yourself as well as your career. Um, so I just I just love, I just love that advice. I think it's it's really great.

Tim McConnaughy: 1:12:08

Okay, well, this has been uh, cables to clouds. If you liked what you heard and or saw, depending on where you are consuming, uh, please, uh, let's see, smash that like button on YouTube and subscribe to us. Uh, subscribe to us on, uh, your favorite pod catcher. Uh, follow us. Uh, you know, I'll give you if you're interested. I'll give you Alex's uh home address where you can find him, and uh, we'll see you guys next time. Hi everyone, it's Tim and this has been the cables to clouds podcast. Thanks for tuning in today. If you enjoyed our show, please subscribe to us in your favorite podcast catcher as well as subscribe and turn on notifications for our YouTube channel to be notified of all our new episodes. Follow us on socials at cables to clouds. You can also visit our website for all the show notes at cables to cloudscom. Thanks again for listening and see you next time.