Telemetry News Now

Telemetry News Now.

Welcome to another episode of Telemetry News Now. We are at the tail end of May two thousand twenty five. A ton of headlines to cover today. But before we get into it, Justin, you are at AutoCon 33in Prague, I believe. Is that right?

I am. Yeah. Well, this is the first day of the actual conference. We just wrapped up the workshops yesterday. It's, well attended. A lot of lot of excitement, so I'm excited to see what the next couple days hold.

Very good. Network Automation Forum does a great job with these events, and, certainly, we are a friend of that organization and look forward to the next one in the United States next fall. So let's dive in. We do have a lot to cover, and, Justin, why don't you kick us off today?

Sure. So first up is an article from Network World, titled Oracle to Spend forty billion dollars on NVIDIA chips for OpenAI data center in Texas.

So, you know, we're covered a lot of different news articles on the podcast in in the past about other companies and their investment. This one is specifically, as the title implies, OpenAI, scaling out their environments to be able to continue to train, presumably, Chad GPT and some of the, you know, some of the the more advanced models that they're coming out with and OpenAI is coming out with. But, yeah, I thought it was interesting. It's actually Oracle that's buying the chips, and they're building the data center for OpenAI. So Oracle is buying four hundred thousand of NVIDIA's GB two hundred GPUs, putting those into a data center that they're actually building and gonna manage and maintain in Texas, and then they're leasing that compute power, basically, those GPUs back to OpenAI so that they can do, their training on on top of it. This investment is actually falling under the umbrella of the Stargate project if, folks remember.

Right. Yeah.

You know, shortly after Trump was elected, he announced the Stargate project, which is the United States government's five hundred billion dollar investment in AI infrastructure. So this investment that, OpenAI and Oracle are making in these GPUs from NVIDIA is actually falling under the Stargate project. So kind of an interesting, collaboration there, I guess, if you will.

Yeah. And it highlights this idea of the importance being placed on AI development in general, obviously, from the government with this initiative that you just mentioned. And then also just, organizations like Oracle spending huge amounts of money. This particular deal is costing them more than their entire cloud revenue from last year, twenty twenty four. I mean, this is a lot of money, a lot of investment, a big move. And it also speaks to the foundation of advancement right now is clearly scale.

There's research and development being done in creating more efficient workflows and models that operate more efficiently at a smaller scale. We've talked about that on this show, but this is clearly an indicator that scale still reigns supreme in our advancement moving forward with AI. Even in a practical sense, considering that many of the models that we talk about when we say AI, these foundational models are models as a service. So they are, you know, huge scale, and then they you know, you hook into them via API.

You're not building them at in your basement. And then, of course, the facility itself, that's really interesting. I mean, that's gonna be a couple years to stand up and then the multiple gigawatts of power. I don't remember what the, the article said.

So, again, speaking to scale in particular. Now, did you read anything perhaps in another article researching if there was any concern about, like, the power grid, you know, in the state of Texas in this case?

That wasn't actually brought up, in the article, but, I mean, I'd have read other articles where that's always an going concern. I mean, it hasn't been that many years ago that during the summertime, Texas had some rolling brownouts, right, because of the power demand on the grid. And when it gets hot in Texas, which, you know, there's three call it three or four months where it gets super hot in parts of Texas. Right?

So the air conditioning puts a lot of load on the grid and then you got all these data centers that are running all these GPUs that are putting load on the grid. So presumably, that's gonna be a concern. The other thing, Phil, that we didn't talk about that I pulled out of this article that I thought was really interesting is Microsoft is an investor in OpenAI. Right?

And there's been a really tight collaboration historically between OpenAI and Microsoft, but it's Oracle doing this data center. So presumably, this reduces OpenAI's, dependence on Microsoft. So they're kinda diversifying who they're, you know, who they're doing business with a little bit here. So that that's kind of an interesting development as well.

Yep. Absolutely.

Alright. Moving on to our next article, from CNBC.

Salesforce is acquiring data management company, Informatics, for eight billion dollars. If you haven't heard of Informatics, they, Salesforce actually tried to acquire them a few years ago. The deal fell through. Informatics has struggled a little bit, and their share price has come down a little bit. So Salesforce came back in, was able to, acquire it for twenty five dollars in cash per share, which, equals the eight billion dollar purchase price. This is actually the largest deal that Salesforce has done since they bought Slack back in twenty twenty one for twenty seven point seven billion dollars.

So Yep.

I remember that.

Yeah. A couple of the other big, acquisitions you may or may not recall, Phil, were Tableau back in twenty nineteen. That one's fifteen point seven. And then, they bought MuleSoft back in twenty eighteen for six point five.

So, you know, Salesforce has a history of these large software company acquisitions, and, presumably, they'll be integrating it into the CRM, which is their kind of flagship product. But Mhmm. Right. Right.

Interesting that it's it has been a while since they've made a major acquisition like this, but they're putting down a lot of money for this company and it really focuses on high scale data for, being able to look at, presumably, I'm not all that familiar with informatics, but, presumably, like, large datasets for mathematical data and so forth. So I'm interested to see how this ties into their CRM.

Yeah. Certainly a lot of structured data, but but not that's not all that we use in business intelligence. And and so, yeah, we're talking about large scale data management, which is the lifeblood of what Salesforce is all about and the idea of the CRM and concept of business intelligence. So this is clearly a a complimentary technology, and and they're probably gonna see, I think, a return on investment pretty quickly, especially as they start implementing AI technologies that Yeah.

That was the angle I was wondering about is are they gonna start using this data to train on AI? Is this basically just them buying a big data set that they can train on AI to to help their sales teams who are their customers better do, like, targeted outreach? Or what do you think the the use cases in this long term?

Yeah. Well, I mean, as far as purchasing the company, will they then turn around and train models on that data set that they acquired as part of the acquisition. I don't know. And I don't know if that would be necessarily the main goal either, because a lot of the stuff that you wanna do is gonna be customer specific.

And so it's gonna be contextually relevant to individual customers, although there's a lot of overlap, so maybe I'm wrong. But I do think that it's the processes, workflows, and mechanisms in the overall data pipeline and data, you know, architecture and the way that they're gonna able to integrate that into how their underlying data warehouse and data lake and all that kind of stuff in data swamp, if you're not keeping things neat and tidy, underpin the technology of what Salesforce does. You know? And so yeah.

And on top of that, you build your business intelligence workflows, whether it be relatively, quote, unquote, basic statistical analysis stuff or, you know, very advanced AI and ML stuff. You know? So in this case, I think the the idea, yes, is definitely to strengthen the AI capabilities and AI powered products. Right?

They have, what is it called? Agent Force now?

Yep. And they did say it was gonna be part of Agent Force. So just the details on what exactly that means is gonna be is a little bit lacking in the articles from CNBC. Yeah.

Those things kind of trickle out, you know, as the integration I say integration like it's a small thing, but it's it's a big deal. So we'll we'll see, especially when we're talking about, you know, data pipelines and integrating that kind of stuff, not to mention the actual, you know, human beings that work there. So we'll see.

We'll be right along. The next article is from AT and T's newsroom, AT and T to acquire Lumen's mass market fiber business. So first, a couple articles we'll talk about here in the telecom world where one telecom company is acquiring the assets of another one. This is essentially Lumen's broadband service, the fiber service that they sell mostly to businesses. So, you can think of, like, enterprise businesses that the Lumen team would sell fiber optic connectivity to those. That's the assets that they're selling off here to AT and T. It includes a total of about one million fiber customers and reaches more than five million fiber locations across, eleven different states in the US.

Pretty big footprint fiber that AT and T is buying for Lumen for five point seven five billion dollars is the acquisition price on this one.

Did you catch how like, the scope of how many new customers this would reach?

The only thing I could, surmise from this is they did say that there's four million fiber locations that reaches one million customers. So presumably, there's like another three million buildings that the fiber is connected to that aren't currently customers for them. So, you know, what I take away from that is their sale AT and T sales team can now go out to these three million buildings that are presumably lit up that have their fiber connectivity into them and try and sell them service where they don't currently have service today. That, I think, is the the attractive thing to AT and T for this as it gives them, you know, some buildings and some reach that they didn't currently have with their existing assets.

I mean, is this a significant enough? I mean, obviously, we're talking about five point seven five billion. So I understand that the actual scope of the purchase is large, but the actual benefit financially to AT and T, is it really that significant or do you think this is more of like a long term play here? Obviously, everything is going to be long term in the sense of it's an investment and you're talking about infrastructure, right?

That has to be a long term endeavor just because of the nature of what, providers do. But what what do you think there? Because it feels like it's not immediately that beneficial to AT and T.

Well, I think it's the the bigger picture. Right? So they're doing their own investment in bearing additional fiber. In fact, one of the, statements from the article was AT and T now expects to reach approximately sixty million total fiber locations by the end of twenty twenty three.

So that's only five years from now. Right? So over the course of the next five years, they, you know, expect in addition to these assets, they're now gonna get from Lumen. They're building a lot more of their own to where they'll be in sixty million total fiber locations, which will roughly double what they have today.

So today, they're at about thirty million. They're bringing in these four million locations from Lumen's assets, and then they'll continue to expand, well, I guess organically, putting fiber in the ground to get to additional locations to bring their total up to sixty. So I think it's just part of a bigger investment where they're just trying to continually expand their footprint.

Mhmm. Yeah. Yeah. And and there is regulatory approval that will be pending.

Oh, of course.

Yeah. Which, you know, considering this was just announced, that's going to be sometime next year, twenty twenty six, probably toward the end of the year. So Yeah. There is some time before all these things go through. And, yeah, certainly does seem like a buy and hold growth strategy in this case. Yeah.

Well, it'll be interesting to see because, like, since we've had a major telecom acquisition like this, we've had a change in leadership in the White House, and, we have a new FTC chairman. So it'll be interesting to see if they take a different stance on some of this stuff than previous administrations have. Whether that's good or bad, you know, leave an exercise to the listener, but I have a feeling, you know, this will be an interesting test for whether the government allows an acquisition like this to go through or if they see it as anti competitive. Right? Because now Lumen's no longer competing with AT and T in these four million locations to provide service to these customers. So definitely be an interesting one to keep an eye on.

Alright. Moving right along, yet another acquisition in the telecom world. Cable company Charter and Cox agreed to merge. Again, this one, just like the previous one, has to go through regulators.

So we'll have to wait and see whether it makes its way through the regulators.

But in principle, Charter and Cox have agreed to merge their two companies.

The deal is values Cox, their portion of the assets, at thirty four point five billion dollars in line with Charter's recent value. So, they're basically mergers of equals for, you know, lack of a better, at least roughly equal. The brand that they're gonna keep around is actually gonna be Charter Spectrum brand. So the Cox brand will actually go away.

They'll keep the Charter Spectrum brand and operate all the assets under the umbrella of Charter even though it's technically a merger of the two different companies. So I think this one will be slightly different in the way that the regulators will view it because cable operators don't really compete with one another. They have their own operating area that they operate out of. Right?

So so I don't know who you have in your area. I'm in charter footprint myself, but, there's no other cable operator, you know, in my area in St. Louis.

It's not like there's three of them that are competing with each other.

Right?

They pretty much have their own monopolies that they operate within. So this one is probably more likely to get regulatory approval. It doesn't really affect competition. Now the reason I say it doesn't really is because they do have things besides just the cable coming into your house, you know, at least I think of when I think of these cable operators. These days, they're doing Philip, they're doing all kinds of different services, right? And so some of those services, they do actually have, some competition, and there could be some anti competitive angles that the government will wanna look at.

Yep. And we are looking at a mid twenty twenty five, which is right around the corner. So I agree with you. This is probably gonna be a quick run through any kind of regulatory bodies to get that approved and then finalize the, the merger.

And I did see that the CEO of Charter, Chris Winfrey, is gonna lead, the combined company. Although, I mean, I have to assume that Cox's leadership is gonna have some sort of roles on various boards and things like that, you know, when all is said and done. Yep. That's how those things go.

Yep.

Okay. Moving on from Fierce Network dated May twentieth, so about a week ago, Arista is reportedly considering a one billion dollar acquisition of VeloCloud. That's Broadcom's SD WAN business, which they acquired just a couple years ago in twenty twenty three and which was first acquired by VMware before Broadcom bought VMware. There's a lot of going on there.

And prior to that, you might remember VeloCloud as one of those more prominent SD WAN startups when, you know, like, they were everybody was talking about SD WAN maybe almost ten years ago. Now analysts are saying that this is gonna benefit both Arista and Broadcom. That's kinda how these things work. It's supposed to be a win win.

So that's sort of the point of these kinds of deals, and it's gonna strengthen Arista's position in yeah. It's a pretty competitive SD WAN market. It's been consolidating over the past few years for sure, but I think it's still pretty competitive. And that to me, it makes sense because Arista would be directly expanding into the enterprise market, which I you know, we've kind of been seeing happening lately and then therefore going head to head with other networking vendors.

You know, think Cisco, for example. Now in my opinion, VeloCloud has always had kind of a strong market presence. It's a solid SD WAN platform. I did I remember doing a four day VeloCloud training, ahead of a big project that I had back when I was working with VARs.

I was impressed.

The thing is that, Arista's current SD WAN product is technically functional. They have an SD WAN product, but it kinda lacks features, and it really you know, that makes it not be very competitive in what many consider a mature market segment at this point. So I think integrating VeloCloud would give Arista a big advantage. They'd be combining their strong routing capabilities, obviously, their data center switching capabilities with VeloCloud's mature, I guess, you call it established, you know, very feature rich SD WAN technology.

That's the Arista benefit. Broadcom, on the other hand, it seems to be moving away from SD WAN. And according to analysts, they view it as kind of peripheral to its core focus on, like, cloud infrastructure, virtualization, that kind of stuff. So they're shedding lower priority, perhaps lower revenue businesses.

Maybe that is, a mistake. I don't know. But that's kind of the idea here. So these kinds of acquisitions have all sorts of disruptions.

Right? So we're gonna see the disruptions to personnel, business strategy, and all that. This case, probably most with VeloCloud employees. So if this does go through, which we're still gonna see, we'll see the ramifications of that, I'm sure, on on social media and and the Slack channels that we're in.

Right?

Yeah. I mean, I I think this one, I always thought it was interesting that VMware acquired VeloCloud to begin with. I agree with what the analysts are saying. It seemed like a little bit of a departure from a company that you consider to be more of a cloud virtualization type of a of a company.

Now they had a lot of network assets. Right? They had, Nicero that they acquired, which became basically their network virtualization or NSX product that they had within their portfolio. So they had some other networking, but it was very much focused on inside of the data center.

So the acquisition of VeloCloud, I always thought was a little bit strange, but it seemed to be working okay with them. But I can see now that Broadcom has acquired the VMware assets, and, presumably, in order to make that deal make sense, they have to get some money from some of those assets by spending them out or shutting some things down that just don't make sense going forward. Like, that's how you get the, quote, unquote, synergies when you acquire another company. Right?

You have to figure out how to how to make it, all make more money together than they did separately.

Right? So I think Broadcom getting what did they say was the actual acquisition cost here? About a billion dollars is what they're gonna get out of this Right. Selling it to Arista.

So, you know, that helps them recoup some of the cost that they paid for VMware to to acquire it. So yeah. And I think, you know, Cisco has Viptela and Arista, like you said, doesn't really have a strong competitive SD WAN solution. So I think from Arista's perspective, it makes perfect sense to really strengthen their product portfolio to better compete against Cisco and the enterprise by bringing in VeloCloud to compete against Viptela.

Well, that assumes that they wanna compete in the enterprise. Right? And I think that's the key here is that Arista, who, you know, for years and years was like a a foundational technology in data centers, especially larger data centers, and Mhmm. Especially when, you know, their platform lends itself to more programmability and openness, which, you know, has been changing among all the vendors in recent years, but that's kind of where they were for a while.

And and I really do think that it's clear with other platforms, not just this acquisition of VeloCloud, that they are expanding into the enterprise market as well and thinking about how to and that's how you make money. I mean, you can't just stay focused completely on one little sliver of the market and then just sit there. I I guess you can if you want to be on cruise control, and there are some companies that do that, but you always need to be thinking of growth, new markets, and expansion. So I think I think this is a move in that direction into the enterprise world.

Keep in mind, though, that it's not like they're slowing down in the data center space. I mean, Arista is one of the prominent switching players in the AI, workload data centers now too. So they are working with organizations like the UEC and with very, very big, prominent customers, web scale customers even in designing and then, you know, fitting these AI data centers. So they're still they're certainly trying to capture revenue from more markets without abandoning, you know, their their core, their foundation.

I think they've only just begun in the AI data centers, Phil, because, like, they you know, a lot of that infrastructure is still InfiniBand. Right? A lot of people are buying the buying into the closed ecosystem from NVIDIA and buying Mellanox InfiniBand equipment to put into those AI data centers. But I think over time, the more people start doing this, the more people are gonna be, you know, happy with Ethernet, wanna see Ethernet, the more the UEC stuff rolls out and Ethernet catches up with some of the things that Infiniband can do in those environments, you're gonna start to see more open, architectures for AI fabrics, and I think that's where Arista will really start to see the benefit because that drives huge traffic on the switches, huge volumes of ports and ASICs and, you know, the next evolution of Philip. So, yeah, I think their data center revenue is going to continue to explode over the next few years in addition to this this acquisition of the VeloCloud assets.

But it doesn't hurt to hedge your bets. So here we are.

No. Of course not.

Alright. Moving on from the Extreme Networks newsroom, also on May twentieth. Extreme Networks has launched new capabilities with Extreme Platform One. One is all capitalized, so that makes it extra fancy. And it makes it the first enterprise networking platform to integrate conversational, multimodal, and agentic AI. That's the claim, and I haven't cross referenced all of those claims yet against every networking vendor out there. But, you know, we can say safely that this is a big step forward in applying the latest AI technologies to networking.

Now platform one, as it's called, is in limited availability and according to extreme according to extreme, right, aims to simplify and automate network operations by eliminating silos between networking and security, significantly reducing manual tasks, and enhancing visibility and control. That's the blurb from their their site. So it's a little bit ambiguous, but clearly, it's the idea of we're using AI, probably some underlying traditional ML models to surface some insight in the data and the telemetry in order to do some sort of root cause analysis, probably, hence the agentic workflows and that sort of thing.

And then when they say enhancing visibility and control, my guess is that we're gonna start seeing some Philip to push config, perhaps some auto remediation. We know we're gonna have to dig into it to see when I when it comes out. But, I mean, ultimately, what they say is that it features AI powered diagnostics, real time dashboards, intuitive workflows, deep network visualization.

They are talking about security. They are talking about the network, and they even mentioned, simplifying their licensing model. We'll see about that. But they have had the platform adopted by over a hundred and thirty customers thus far, and we'll see general availability in q three of this year, so later this year.

Yeah. I mean, I think they've had some amount of observability and automation in the platform for a while, but being able to being able to bring AI on top of that, especially when, you know, they kind of vertically stack this and integrate it, right, makes a lot of sense. And I really like the idea of, like, the the access control. Right?

So being able to run AI, look for malicious traffic in presumably your data center is what we're talking about here because that's where your stream is the strongest, right, is in the data center. So if I've got, you know, some sort of, malicious traffic in my data center, whether it's north south coming in and out of my data center or east west within my data center, being able to detect that, send down updated firewall filters and be able to block that traffic in the data center, that actually makes some sense to me. I can kinda see where the security piece of this would come in. So it'll

be interesting to see where this goes. I have a feeling customers who are in a, multi vendor environment are gonna struggle for this. I'm assuming that this product works really well if you have extreme switches, but you have Arista's or you have Cisco or you have some other data center switches in your environment.

I'm guessing it probably works less well for mixed vendor environments, but who knows?

Yeah. I mean, the thing is, in my mind, as a former operator and engineer, I do think that having some sort of orchestration platform that runs a block of your network is still great. Okay. So I use this for all my data center stuff, but then I end up using this for all my WAN routers or my firewalls.

Mhmm. And and, I mean, I've heard people decry like, oh, you know, look at that. You have to touch multiple things. It's like, well, so what?

You're not touching eight hundred and fifty thousand devices. You're touching your five different orchestration platforms, you know, your firewalls, and then here for your whatever. I think that's still great.

We're going to talk about this in the next headline, I know, because we're going to talk about Cisco certifications and vendor neutral and vendor agnostic and all that.

The idea that you can have a vendor networking vendor develop some sort of orchestration platform, whether it's using AI or not, and then have it be completely vendor agnostic or your vendor agnostic white box environment. I think it's kind of a pipe dream. And I also think it's there's really no huge market for that. Maybe it's just my naive experience, Justin, but I have never met a customer that was anxious to run eight different switch brands in their data center.

Sure. They'll run, like, a brand in their data center and a completely different brand for their firewalls or SD WAN solution, you know, like, again, like the network blocks. Right? Mhmm.

But never, like, eight different brands in the data center itself, except for some one offs.

Right? You know, except for two one offs. In that rack over there, we have these couple switches for this, you know, consultant or whatever, that kind of thing.

Or if you build it in blocks. Right? Maybe I have row a few rows that are one vendor that that was my legacy architecture, if you wanna call it that. And then when I went and I went from core and access switching l two VLAN trunked network, I switched vendors, and now I have a VX LAN, EVPN architecture, and I've switched vendors.

Right? I could see keeping those separate and having separate architectures for, you know, legacy until some of that stuff is is phased out and so forth. But I agree with you. Like, building an ongoing design and architecture that had that just swaps out from one vendor to another is I've never seen it done unless customers are doing white box, right, and doing something like Del Sonic.

Right? That that's the whole purpose behind having some of that kind of stuff is that literally the hardware itself doesn't really matter to its abstract in a way.

But Exactly.

Exactly. So interesting, advancement from extreme networks.

Lastly, from the Cisco Learning Network, quote this is quote. On February three two thousand twenty six, learn with Cisco will introduce exciting changes to its industry leading certification portfolio shaping the next generation of networking careers in the age of AI, network automation, and cybersecurity.

Yes. Our name is also new, end quote. So that was the first sentence of the press release about the revamp of Cisco certifications clearly focusing on automation and AI. And they have been focusing on automation for a while. Remember DevNet?

So basically, Cisco is rebranding its training and certification program as, quote, learn with Cisco. And the idea here is that it's gonna help or they're gonna help networking professionals stay relevant amid what we're seeing with AI right now, automation, of course, and public cloud. So they're streamlining the certification framework focusing on, you know, the familiar certs, like CCNA, CCNP, CCIE brands, but launching, like, updated tracks in automation collaboration and cybersecurity. This isn't new. This has been going on for a generation. Right? We've seen new tracks being launched.

Mhmm.

This is interesting, though, because the DevNet program will be rebranded as CCNA, CCNP, and CCIE automation.

I actually think that kind of that part of it actually kinda makes sense, Phil. Yeah. Because, I mean, like, if you look at what some of the other vendors do, they have the same thing. Like, you know, Juniper, for example, I know them a little better because I spent ten years there.

But, like, they'll have JNCIE, JNCIS, JNCIE for routing. They'll have one for enterprise. They'll have one for security. I think they now have one for automation, if I remember correctly.

Right? So it kinda makes sense to drop the DevNet certification namings and go to the CCNA, CCNP, CCIE that everybody already knows and loves and just have it be for doing automation. Right? Doing Yeah.

Absolutely. DevNet type stuff.

Yep. And, though minor all expired, I had I never got the CCIE. I failed. But I did have a CCNP in four different areas and then other certifications outside of Cisco in specific areas, just like you said. And so those tracks make sense, and I think this rebranding makes sense.

But I I do wonder about this.

The world of AI is so so big and vast, and it goes so deep that I'm interested to see what these certification, like, blueprints are actually gonna look like when they come out. You know what I mean? Because you're not just talking about, like, vendor specific stuff or how to configure OSPF, which, you know, is relatively finite. There's still a lot to cover there, of course, and go all the way up to the CCIE level. But what will they choose to focus on, and then what are they gonna omit? Right? I mean, how often will they be updating these tracks?

That's the part of it that I find interesting. Right? Like, to your point, the IETF RFCs for OSPF, like, at least the basic foundation of how the protocol works that you'd wanna learn to pass CCNA, haven't changed in a long time. Right?

So it's easy enough to build training material and build a certification that doesn't have to iterate very frequently. But AI is like the polar opposite of that. Right? We're like, how many times every two weeks, we're talking about a new model on the show.

Right? I mean, like, AI, the innovation is exploding. But I guess maybe the foundations of how you train a model and and some of that isn't changing very fast, and maybe you or how you apply it to automation isn't I don't know.

Well, that's what I think. I mean, the thing is that Cisco would update tracks every so often anyway. Right? So you'd you'd go from, like, CCNP voice to CCN collab five years later, four years later.

And I think it's gonna be generally the same here because supervised learning, unsupervised learning, you know, they are what they are, and you're gonna learn them. You're gonna learn, you know, some Python. You're You're gonna learn what a four loop is, you know, as far as the automation stuff. So you're gonna learn some of those foundational things and then get more advanced as you get into the CCMP and CCIE.

And those aren't coming out every two weeks. Those aren't changing every two weeks. And like a new model coming out, the underlying technology isn't all that different. The the new model that comes out is either a greater scale, some tweaks to efficiency, and so it's or some updated training data and you have a new model.

But the underlying thing, like with a large language model, it's still a transformer, still back propagation and gradient descent, things like that. If it's more traditional, AI ML, you're still doing, like, cosine similarity and DB scan and whatever other models you're applying. Those things haven't changed in twenty years either. So I think that Cisco does have a solid foundation for, you know, just pure education.

And then, of course, you're gonna have those updates, like, for example, you know, MCP comes out, everybody jumps on that bandwagon.

Do you go back and redo all your certifications to cover that? You know? No. So there's gonna be some of those very, very up to date technologies that are gonna be lacking.

It's it's inevitable, with what's going on, like like you said, because things are moving quickly. But I think that's really neat. I am really interested specifically in the blueprints when they come out. I'm going to study them and see, like, okay, where do I need to go and fill in the gaps in my own knowledge?

Yeah.

We'll be interested in to keep an eye on for sure, at least.

And now moving up to upcoming events. We don't have too many to discuss today, but we do have the Chicago networking user group, part of the USNUA on June five. So keep an eye out for that one if you're in the Chicagoland area. We have the Missouri networking user group, also part of the n USNUA on June five, the same day. Justin, I believe that you are the leader of that particular event.

I am one of the local organizers and, our mutual friends and friend of the show, Scott Robon, will be coming to do the keynote for us on his concept of total network operations.

I just saw him here at AutoCon. He was telling me he's, refreshing the material, so we'll get to hear some breaking news. It'd be the first time this particular talk has actually seen the light of day, so I'm excited to see where he where he takes that.

Excellent. Good. Good. Yeah. Looking forward to hearing about it. We have NANOG on June nine through eleven in Denver, Colorado.

And then, of course, one of the bigger events in the of the year in the networking space, Cisco Live, also June nine through eleven in Las Vegas.

Justin, are you gonna be at either of those or maybe bouncing between the two?

I thought about bouncing between the two of them. That seemed like a lot of time on a plane, so I'm gonna do NANOG, in Denver. I'm part of the mentorship committee there with NANOG, so I think I'll stay put and participate in that one this time around.

Great. Alright. Well, those are the headlines for today. Thanks for listening. Bye bye.