More episodes
Telemetry Now  |  Season 2 - Episode 23  |  December 5, 2024

Telemetry News Now: Suspicious Submarine Cable Cuts, AI Model Causes Security Concern, New OpenTelemetry Certification

Play now

 
In this Telemetry News Now episode, Phil, Justin, and Leon discuss the launch of the CNCF’s OpenTelemetry certification, security concerns after recent damages in the Baltic Sea, Cisco’s partnership with NTT DATA to enhance global 5G connectivity, and Arista Networks’ growth driven by AI data centers. And as always, we end with a quick recap of important upcoming events.

Transcript

Telemetry News Now

Welcome to another episode of Telemetry News Now. We are recording today a few days before Thanksgiving here in the US, though you're likely listening, the week after, perhaps two weeks after Thanksgiving, which is the way it goes with coordinating schedules to record and scheduling and and and all that kind of stuff. So happy belated Thanksgiving to our listeners today, especially those in the United States. So without further ado, let's get into today's headlines.

Okay. So I am delighted to report that our first item has nothing to do with hacking or AI. The headline reads, OpenTelemetry Certification Launches.

This is a new certification program from the CNCF or Cloud Native Compute Foundation, which is part of the Linux Foundation. The goal is to validate people's skills in utilizing open telemetry and configuring it for monitoring distributed systems. The certification is called OpenTelemetry Certified Associate, and if I know anything about certifications, there will be an OpenTelemetry certified engineer and probably an OpenTelemetry certified professional certification coming after this. That's just me guessing based on what I know from my Novell and Microsoft and Cisco certifications in the past.

It is abbreviated OTCA, And it's going to cover the fundamentals of observability, the OpenTelemetry API and SDK, OpenTelemetry collectors, which is where you lose your mind, and finally maintaining and debugging observability pipelines.

The exam is gonna cost about two fifty bucks. It's, that's bucks in US bucks.

And, the scheduling slots open up in January of twenty twenty five. I love everything about this. I'm a exam hound. I always have been.

I just which is just a character flaw. I know. I love studying for exams and taking exams like that. They're just fun.

Even if I don't pass, and I don't always pass, but the process is really rewarding. You learn a lot. Even the weird bass ac words language that they use to try to trip you up because it's an easy question, but they the people who make tests just can't live with the idea of an easy question.

Also, the fact that it's sponsored by CNCF and not a vendor lends a lot lends a lot of credibility in my mind.

And I I feel like I I am honor bound to mention that I am actually on a working group in the CNCF, the deaf and hard of hearing working group, which helps, create more accessible events and, materials within the organization.

I have to echo a lot of what Leon said. I really like certifications that are vendor neutral. I think that, you know, vendor sponsored certifications are fine. They can help you learn how to implement a certain technology on that particular vendor's platform, you know, like in their my early days of my career, I spent a lot of time doing, like, Cisco and Juniper certifications, you know, learn how to implement those technologies on their platforms. But I think there's something special about a vendor neutral one where you're really just learning the foundations of a technology. And then once you get that foundational knowledge, you can apply it to any particular vendor's, implementation.

I also kind of agree with Leon. I still enjoy certifications. Early in my career, it was more because I needed that on my resume so I could try and get the next job to move up the corporate ladder. These days, it's more just a learning exercise. There's nothing, nothing like proving to yourself or proving to yourself you don't know something than taking a certification exam. Seems like you're mastering the material if you're going through and reading it and studying it, but once you actually have to go and, like, test that knowledge in the form of a certification exam, you really find out where the gaps in your knowledge are. So I still enjoy them for for that perspective unless it's for career growth at this stage in my career.

Right. And I'll also say at this at the same time, it also proves, like, when the test doesn't match reality, that it it can be very frustrating for a lot of people. Like, why do they do that? But at the same time, like, I know that it doesn't match reality.

So it's good to have that confirmation.

Yeah. Yeah. I mean, OpenTelemetry was a question mark for a while, wasn't it? So this is interesting. And and I and I also think that it speaks to this interest in observability that I'm seeing out there in in the industry. And and, of course, what our what our company does, network observability, but really looking at this entire system of applications and subsequent application delivery as a whole and really understanding how things interact with each other. And and, you know, OpenTelemetry plays a big role in that for, for folks in trying to understand the health of these systems.

So, yeah, definitely, agree with with Justin as well.

Regardless of whether you use it or not, it is it is a learning framework, which I very much appreciate as well.

Okay. So the next article, is a bit of networking engineering, catnip clickbait. Here we go. Ready?

Undersea cables, submarine cables, ocean floor network cables, sharks biting cable. Okay. Not that one. I just figured I would help our SEO a little bit, make it more reachable.

It's gonna resonate a lot, I think. Anyway, the actual headline says, and I quote, FCC launches first comprehensive subsea cable rules review since two thousand and one, unquote.

So the article goes on to explain that the demand for additional and more advanced runs of underwater cable continues to grow. So the FCC feels it's time to streamline and improve the rules for deploying cable while also guaranteeing the security and resilience of what's become a critical aspect of global network infrastructure.

Of particular interest to the committee is interest in beefing up security around CLSs, the cable landing stations, where the undersea cable comes ashore and connects with terrestrial equipment. So I'll start with the obvious joke, where is Doug Midori when you really need him?

That aside, I'm sure he's listening. I'm sure his ears just perked up. I think the article, really does a great job of articulating why this is important, this review and the beefing up of the of the rules about this is important in general, and also why now specifically in terms of and we're actually going to get to this later, but the fact that undersea cables are more at risk both from encroachment but also from perhaps, purposeful assault.

They also have some really cool statistics that, I didn't know. There's nine hundred thousand miles of undersea cable currently in operation, and ten trillion dollars per day of financial transactions go across those cables. So that, gives you an idea of why this may be getting a lot of attention.

Yeah. And and, certainly, the FCC has had some oversight to an extent over submarine telecommunication cables for years. But these cables, they they do cross international boundaries across entire oceans from continent to continent, country to country, sometimes those countries being adversarial in in their relationship. And so there is there is significant complexity in the FCC trying to extend its oversight. We do have the International Cable Protection Committee, the ICPC.

So it is it'll be interesting to see what kind of role they they play in in, coming years. And, and I think this all stems from the growing understanding that the the system of cables, that powers the global Internet and, therefore, commerce and the spread of information and the connectivity that we experience in our modern day is, is both a very real attack vector. And, also, you know, there's an element of fragility there. I mean, just drag a ship a fishing trawler dragging its anchor can accidentally break cables and, therefore, Internet service to an entire geographic region. That's not necessarily a a a nefarious attack, but certainly still speaks to the fragility of the system and perhaps the requirement of additional oversight beyond what we have today, especially as as we pretty much rely globally on on this system for pretty much everything that we do. Right? Yeah.

I mean, as Leon alluded to, our coworker and colleague Doug Madore does a lot of reporting on undersea cables, and he might have even been involved in this work with the FCC. I know he donates a lot of his time back to policy making and protection of a lot of this stuff.

For a lot of places on the globe, this is critical to connecting them to the Internet, which is critical for how they interact with society. Right?

They may still have satellite communications as a backup, but those undersea cables is really their preferred method of communication because better performance, better bandwidth, lower latency, provides them a much better experience.

And, you know, as Leon also alluded to, there's a lot of rumors around shark bites. That's not actually one of the more common causes of damage to the cables. Much more common is, human damage, whether it's intentional, or unintentional, like anchors from a boat. You know, it's interesting to see the FCC putting some rules in place in trying to help protect this infrastructure that is critical for a lot of these places. But, Phil, like you said, the FCC only controls the US portion of this, and this is really more of a global problem. There are other nation states involved, so we're gonna have to see more, I think, action from some of the more global, organizations beyond just the FCC to really see a lot of improvement in this. But it's still still good good, good action on the FCC's part for sure.

Yeah. And to be clear, the specific actions that the FCC is taking, is it's in it's in the context of a growing security concern. So there is a review happening with regard to the, rules and regulations and processes to obtain a license to operate a cable landing station, operate and deploy and build, and then, of course, the subsequent submarine cables, and, and and also a review of and maybe even a prohibition of using certain equipment that is manufactured.

Basically, it's from companies, on an FCC list of companies that are deemed to pose a threat to the United States and to US national security. So that'd be companies like Huawei and ZTE and things like that.

Alright.

Next up, we have a press release from a small hardware vendor out in San Jose called Cisco Systems I've heard of them.

This week. They and, yeah, I never heard of them, and I don't know what they do. We'll have to dig into that in a future episode. But, this week, they announced their partnership with NTT DATA to simplify access to five g connectivity.

The goal here is to give the global workforce more connectivity options. The article states that by using Cisco's eSIM technology, workers can access NTT DATA's Transatel network, I believe is how it's pronounced, from a hundred and eighty different countries. You know, it's not really clear from the article whether they're using the five g time slicing technology that's part of the five g standard, but that seems to be what they're doing here, which is I always thought ever since I first started reading about five g architectures that that could be really interesting to be able to provide, users, corporate users, kind of like a VPN style connectivity, private connectivity over five gs wireless network. So presuming that that's what they're doing here, that's really exciting to see, you know, an actual launch or something like that.

Alright. So my first question is always is, is this gonna let me get, faster or cheaper Internet at home?

Because that's, of course, where my brain always goes. But, beyond that, beyond the again, the easy joke is that anything that enables remote work and pushes back on the invented narrative of people can't work unless I can actually see their actual button in actual chair, which, as someone who's worked remotely for over a dozen years just burns my buns.

Anything that, you know, pushes back on that narrative is good in my mind, you know, let people work, do the work they need to do from wherever they need to be when they need to be there and be done with it.

So I am excited to see how this opens up opportunities for people on the ground.

Yeah. I mean, I think I mostly agree with that. It's a serious trust issue in your employee employer relationship if the only way you can manage your workforce is by seeing them physically in their chairs in the office. Right?

At least for most job functions. I mean, there's probably some that are, you know, still require a physical presence, but I think for most, you know, at least in tech, most jobs in tech, remote works fine as far as being productive as long as you set proper, goals for that role and provide people with coaching and feedback to help them accomplish those goals and so forth. I will say, though, I personally like hybrid a little better. I do miss, some human interaction office kind of lended itself well, too.

It's one of the reasons that I really love going to conferences, is I learned so much from, like, you know, Phil said we were out at AutoCon last week and, like, just collaborating with each other and talking, bouncing ideas off of, you know, your coworkers or people at other organizations that are attending those conferences. Like, it really compresses the learning down into you learn a ton of stuff in three, four, five days at a conference that it would take weeks or months if you could even replicate it on a web meeting.

Well and that's why I was very deliberate about the way that I said it because, again, I've been doing this for a long time is, you know, people need to be able to be where they need to be to get the best work done that they need to do. And sometimes that's in a conference room together, but a lot of times that's, you know, at home not being interrupted or whatever it is. So, anyway, that's just that's one of my soapboxes.

Yeah. I mean, Transatel makes, this cloud based, like, network for it's not a not a cellular service provider, but for the cellular components and then the subsequent network that it traverses over, the data traverses over for, like, IoT devices. And I think one of their customers is even Airbus that I when I did a little bit of research. And so all that stuff calls home.

I I don't know what the it wasn't clear to me what the goal is here with Cisco, whether that's for, like, out of band remote management, zero touch deployments, or if it's for just, like, sending telemetry back, which I is a thing. So I or or maybe all of the above. Either way, I I wasn't quite sure what the, what the actual goal here is incorporating this technology into, into network devices. There is this one line about enabling secure remote branch services on demand.

And I don't know I don't know if that means zero touch or, you know, just an easy way to to spin up new sites, which I guess, I mean, that's that's zero touch. So in any case, not exactly sure.

Yeah. I mean, I presume just from the fact they said it was based on Cisco's eSIM technology that it it has something to do more with being able to have customers provision, you know, electronic SIMs on their phone to connect to different wireless networks much more easily than having to go in and get it spun up in a Verizon store, an AT and T store, whatever allows them to roam from country to country and still be on the same logical network, even if they're on a different cell tower, different provider. But, again, I'm connecting a lot of dots. The article wasn't real clear on a lot of that.

Yeah. Yeah. Agree.

Alright. And now it's time to talk AI for a minute. This in this case, we're talking about one of the hardware vendors is benefiting from the AI boom. The Nasdaq released an article this week talking about why Arista is still undervalued even as their stock has performed so well year to date.

I think we all know InfiniBand is being pushed hard in these distributed GPU clusters, especially by, NVIDIA, who's the eight hundred pound gorilla, when it comes to the GPUs being provided. I think Arista's recent performance is proof that, the rumors of Ethernet's death are greatly exaggerated. Arista is benefiting pretty heavily from a lot of, and maybe more so than any of the other switching vendors from network spend in these distributed GPU cluster environments. So So I think it's proof that Ethernet is still relevant in these environments.

You know, one thing I think is really interesting is and it was kind of teased out a little bit in the Nasdaq article is this is mostly still hundred, two hundred, and four hundred gig speeds that they're benefiting from that they're selling today. Right? So they're just now starting to see the tip of the iceberg, if you will, on the eight hundred gig deployments. Presumably, we'll see a lot more of that next year in twenty twenty five, but there's still a tremendous amount of upside for Arista in this market, I think.

Yeah. I mean, Arista has been certainly trying to expand or broaden their portfolio in recent years into the campus and and and into the WAN. But they they've always been heavyweights in data center networking and and in serious large scale data data center networking as well. And, and then in the past few years, their heavy involvement with the UEC and in, providing the the hardware to support AI workload data centers, I mean, that has been a a big deal for Arista in the past few years, and I'm sure a big reason for why they're doing really well, among other things.

I think that they're just a great, open platform. But you know why why they are undervalued? I guess I'm just not as adept at, like, understanding business things. So may maybe, Justin, you can elucidate me a little bit.

Why do you believe that they are an undervalued company?

I think it's just they're they're looking at the future value of what their stock price is gonna be. Right? And they're saying that for what they're currently trading at and what their future income potential looks like, they're still an undervalued asset if you were to invest in their stock is basically what they're getting at.

Alright. So, basically, their growth trajectory doesn't match doesn't align exactly with their current stock price. And so folks are saying that they are an undervalued company based on potential future earnings. Is, is that about right?

Yep. Exactly.

And so do you think that cloud repatriation also plays a role here with folks that are now considering bringing considering and actually following through with bringing resources back on premises?

Oh, a hundred percent. I mean, I think, I don't remember where I was reading this article, but, somebody was talking about how what a lot of enterprises are doing, what a lot of companies are doing is using things like AWS Bedrock, right, where you can run, AI in AWS's cloud to experiment, play around. Like, some of the the thoughts that they have, some of the use cases they have in mind for, AI may or may not pan out, may or may not work, or they may or may not be financially viable. Right?

Once they find out like, hey, that experiment actually paid off, that actually has legs, there's actually a tangible ROI there, then they can repatriate the workloads back into their own AI data center. So then they start to, you know, either go to a data center provider, have them build them a cluster or do their own training, or they build their own data center with their own GPUs and do their own training. I mean, there are some enterprises that are now just now starting to do that because they're realizing that some of those experiments actually have legs and there are some actual tangible use cases for it.

So it's kind of a spin on the workload repatriation. It's a very specialized use case for it, but, you know, it it is out there.

Well, I will say it's not it's not as unique or specialized as you think because originally, I think I'm going back to twenty sixteen, twenty seventeen, maybe as late as twenty eighteen, nineteen, but ninety percent of Azure's workloads were test and dev, and ten percent were production workloads. People were using cloud at the time specifically for basically as a cheap sandbox before they brought that test back home.

I I think, you know, to reuse your joke from the beginning of this segment, you know, the death of of the data center has been greatly exaggerated. I think that bringing things back on prem continues to be a thing that we're gonna see for a variety of reasons, both in terms of a specific use case, but also in terms of particular organizations that, you know, the the numbers aren't right for them even if they might be right for other companies.

Right. Right. Yep. Yep. Yep.

So so moving on, it all over the news, Reuters, CNBC, CBS, Network World, TechCrunch, two undersea Internet cables in the Baltic Sea were recently damaged, raising concerns about a possible sabotage linked to Russian hybrid warfare. Now, of course, the Kremlin does deny involvement, and, what we're seeing now are investigations focusing on a Chinese cargo ship, the Yipeng three, which was in the area at the time. Now Finnish and Swedish telecommunication companies did confirm damage to their cables with Sweden launching a a a criminal investigation.

Now the incidents follow earlier reports of Russian interest in the area and Nordic infrastructure, and it does echo of past attacks. So if you remember the Nord Stream pipeline explosions, which is still a question mark. And, of course, this, heightens tensions in the area as Nordic nations now have to bolster their defenses and potentially even prepare for a a crisis of sorts.

Yeah. I mean, I think a lot of there's there's a speculation for sure in this article. Right? I think we've mentioned on previous episodes that there's only so many ships on the planet that do these type of repairs.

And it seemed like from the article, the ship hasn't gotten there to do this repair yet. Right. So until that takes place, we won't have all the facts. But I think given the timing and the fact that there were two cuts and given where it took place in the Baltic Sea in such proximity to Russia, who we know is obviously in a war in Ukraine, like, there's a lot of speculation going on here, which, you know, has some some merit to it.

But it'll be interesting to see once all the facts come out whether this was in fact accidental damage or an intentional sabotage. But either way, like we said earlier, like, this is critical infrastructure for the people in those regions that they're now having to do without till the ship gets there, can get this cable repaired, and then get back online.

I mean, I don't think the facts are gonna come out.

So Well, that's true.

You know, I don't think we're ever gonna really know.

I I think it can't be understated that both Finland and Sweden recently joined, NATO and did so over Russia's strong objection. Mhmm. Yeah. So, you know, I'm just continuing to beat the drum of geopolitical.

Finland joined in twenty twenty three, Sweden in twenty twenty four. I I agree that this particular cut didn't have much issue. If I look back at Doug Madore's reporting on it, you know, they didn't lose any real connectivity because there were duplicates or other routes. However, you also know that this particular nation state is very good at effectively doing scare tactics.

You know, I only caught one line today, but, you know, that other line looks really fragile. So it'd be terrible if something, you know, almost like the mafioso, like I'm bending my nose here. It would be terrible if something happened the way, you know, like that kind of thing. Yeah.

So who knows? And, obviously, this article ties nicely with the one that I brought up earlier with the FCC beefing up security and protecting the infrastructure.

And I think that all of this, both the news articles talking about it and also the work to protect the infrastructure, make it more robust. All these critical elements are going to continue as both technology and politics progress.

Does does politics progress?

I don't know.

Political situations continue to progress and grow, especially over the next Evolve.

Maybe maybe that's the progress is the wrong term.

Maybe evolve is the right term.

Yes. To develop. Yes.

Right. Okay. So moving on from the silicon angle, November twentieth, Chinese startup DeepSeq's newest model surpasses OpenAI's O1 in reasoning tasks.

So let's go through this. DeepSeq is a Chinese AI startup launched which launched a reasoning model called DeepSeq r one. That's dash r one. And that's, it's mostly designed to excel in solving math and science problems, but with much better accuracy of what I guess you could call, like, traditional large language models, you know, quote, unquote.

Now the news here, there's there's two real, pillars of news in my opinion. The first is that this rivals OpenAI's o one model, which was, like, big news when this came out recently, and therefore focuses on reasoning by using chain of thought techniques to minimize errors. And, what that means is that deep seek is gonna take a little longer to respond to a prompt just like, you know, a GPT o one. And that's because it's sort of, methodically going through the chain of thought process, taking its time, breaking down the prompt into, various parts.

And the idea is that we're okay with that. Right? Because the delay is a good trade off for higher accuracy.

Now in the benchmarks, DeepSeq outperformed the GPT o one in a couple of a couple of tests evaluating its performance. That's good. And also math and complex word problems.

Apparently, it did, pretty well with some, like, trick questions that folks like to use to test these LLMs. But that to me, I've always been kind of suspicious of these reports and benchmarks based on on these types of tests. Right? Because I've seen, that sometimes a new model is released.

Right? It does very well in these controlled benchmark tests and the way that the prompts are structured and all of that. And then we have these reports of how accurate it is and all these kind of things. And then it does much worse in regular production.

So I'm not quite sure how to really interpret those, benchmarks and those tests just yet.

But what's also very interesting to me, probably more interesting to me staying on the theme of geopolitical, ideas and things, is the the model is controlled entirely by the Chinese government. And that means it's literally built into the platform that it won't answer questions about what the Chinese government deems as politically sensitive or contrary to its, communist.

Actually, the article says socialist values. That's a big deal. I mean, that's a big deal to me at least because we have very powerful models coming out now. Rivalry rivaling the very biggest and most advanced foundational models like GPT and Claude, especially GPT.

Right? But they're wholly controlled by potential adversarial governments or, in fact, right now, adversarial governments. And what does that mean going forward? I mean, you guys know I've been talking about AI a lot, but I've been talking about specifically AI in the context of security.

I just recorded a podcast with TJ Sayers from the Center for Internet Security recently, and, it just enlightened me on how this technology is being used, as an attack vector, both from nation states, and the those sort of, actors. And, and also from, I guess you can call your, like, everyday cyber criminals. Right? But it does feel like a new effective way to manipulate information at a broad scale for millions and millions and millions of people, which I guess is probably already happening and is always happening, you know, as part of any psyop and everything.

But it's definitely something that's concerning for me and a new direction that I think we should be aware of.

Yeah. I mean, I hadn't really thought about the, I'll call it, censorship angle that you just talked about, right, where the Chinese government, if they control this model, can control what information it will and will not return, and they can censor out things that don't, follow their narrative. That's kind of concerning. Mhmm.

But the more, obvious thing that kind of came to mind as I was reading this article is the words AI in Chinese in the same sentence, meaning that one of the biggest concerns around AI models is something you talked about on your podcast with TJ Sayers is around security risks of these LLMs. Right? Where where does the data go? Like, who has access to that data?

You know, those type of things. And that was what really concerned me. We know the Chinese government having looser regulations than the than some other countries when it comes to how they handle data and privacy around said data, you know, and how they they have been known to use it as surveillance for surveilling their own citizens, but other countries as well. That was really the more concerning part of this is, you know, if they get this out there and it becomes as ubiquitous as, say, ChatGPT as a chatbot, people go out there and put private information in there, either their own personal or their employer's private information into this chatbot and it winds up in the hands of the Chinese government.

What, you know, what can they do as far as, you know, potential attacks, you know, as a nation state against some of these companies? So, yeah, a lot of concern for me on on the security ramifications of this. But I guess just continuing to see companies investing in making models better and training, especially around reasoning, is really interesting. So I think, you know, like most things AI, there's there's both good and bad that come along with it.

I mean, from a purely technical perspective, great. It's there's advancements happening, and that's certainly interesting. And there is a net positive there from that angle. But and and I'm not opposed to guardrails.

I mean, we've talked about this many times on on various podcasts and blog posts that we have guardrails to protect private information, sensitive information, and all of that. And that and that's important. But we're talking about now, you know, the manipulation of information and the withholding of information. Well, I guess that's part of manipulation.

So that that is a concern. Now, of course, you can choose to use that model or another model or whatever, but I do foresee a consolidation of the largest foundational models over time where they are offered as a commodity model, a commodity service that I think folks are gonna use, whether they realize it or not, because they might be using said companies, you know, chat system, which on the back end is hooking up hooking into DeepSeek's, you know, API. Who knows?

So, we've already, you know, talked about how I'm a curmudgeon when it comes to all things AI, so we'll leave that out. And also we've already hit the, geopolitical side, and I don't really wanna sound xenophobic. What I do worry about also and what this article highlights is the complexity issue that these the technology is becoming incredibly complex more than any single or group of people can reason about to the point where almost nobody can know what's going on there. And and just by way of comparison, a friend of mine, Avishai Shalom, gave a talk back in twenty sixteen at DevOps days, and he showed, like, this very simple five line Python script.

Right? Oh, it's so simple. It's so easy to debug. Except he pointed out it's running you know, there's forty thousand lines of Python modules that were called from that original script.

And then, of course, you've got the five million lines of CPython and C libraries that make up the Python library language itself. And that's all running on thirty million lines of code that make up the Linux operating system. So it's not so easy to debug. And when we think something as simple a five line, it's only a five line query.

It's only a whatever, yes, but you first have to invent the universe.

So to to paraphrase Carl Sagan, so I worry when not only are these models becoming more and more complex, but they're also now being firewalled behind geopolitical barriers. I think ownership of this technology and openness of this technology is gonna become increasingly important Mhmm. To everyone in terms of management and just trust.

Yeah. Yeah. In fact, that's probably one of the reasons that a lot of the smaller models are becoming so popular, because some of the smaller models, when implemented properly in a greater workflow, are just as accurate, just as useful as, the the the largest, you know, general purpose models. Fact, they're probably more useful than general purpose models because you don't need a general purpose model.

You need a model for that specific, you know, domain of knowledge or whatever it happens to be. So a small model is just fine. But I don't I don't think it's xenophobic to to be concerned about, security issues, especially when there are apparent, you know, nation state adversarial relationships. I mean, that's part and parcel of differing interests among nations.

But I do understand what you're saying that what is what is the true nature of what's going on? What's the purpose of the development of DeepSeek? And, you know, is it is it to perpetrate these sort of information gathering and manipulation attacks on general populations, or is it just a model and, you know, people are gonna use it the way they use it?

I wanna point out that Pokemon Go was openly sponsored by the CIA.

Oh, really?

And it was, you know, we're not just collecting your geo like, we're not just mapping. We're determining if we can figure out your gait of walk from this app, and they openly collected all of that. And so I think anything that we use is a vector for information gathering and insight.

I am so ready to get a tiny house and move to the mountain and be off the grid. I say that, but, you know, I like having a big house.

But, I mean, seriously, sometimes I'm ready to just go fishing and then stay Off the grid.

Yeah. Yeah. Yeah. Really. Yeah.

Alright. So moving on to upcoming events. And, unfortunately, because of when we had to record this episode and when we had to publish it all because of the timing with the Thanksgiving holiday, some of these things are happening right now while you're listening to this podcast. But first on the list, we have DevOps Days Paris, December four and five in Paris, France. One of the biggest events of the year, AWS re Invent, is December two through six in Las Vegas. And, gentlemen, I believe both of you are gonna be there. That's correct.

Yes. We will.

We have Gartner's IT infrastructure operations and cloud strategies conference.

That's a mouthful. December ten through twelve in Las Vegas.

And one of the largest tech events, though not specifically networking, and much more broad, tech events of the entire year is a CES happening January seven through ten in Las Vegas.

We have CodeMash in Sandusky, Ohio, January fourteen through seventeen, I believe. And Leon will be giving a talk on technical empathy.

And last but not least, we have AI field a six, January twenty ninth and thirty, both in Silicon Valley and livestream. So make sure to tune in to the livestream.

But I do wanna backtrack. Leon, can you tell me a little bit about what you mean by your talk called technical empathy?

Right. And that's that's part of the problem is, we as IT practitioners and professionals design things, whether we think we do or not. It's not just if you're a front end dev or whatever. We all design systems, and we have to be aware and empathetic to the user's experience. So we have to put ourselves in the driver's seat. And so I illustrate that starting off with a story of a really bad lunch date with my wife in Switzerland.

And moving forward from there, I have some really horrific design examples. And it's in sunny, beautiful Sandusky, Ohio. So, you know, why would you wanna miss that? It's it's the event of the year in my mind.

Well, you know what? I'm sure it's gonna go amazing, and, I I'm looking forward to a good report.

So for now, those are the headlines for this week. Bye bye.

About Telemetry Now

Do you dread forgetting to use the “add” command on a trunk port? Do you grit your teeth when the coffee maker isn't working, and everyone says, “It’s the network’s fault?” Do you like to blame DNS for everything because you know deep down, in the bottom of your heart, it probably is DNS? Well, you're in the right place! Telemetry Now is the podcast for you! Tune in and let the packets wash over you as host Phil Gervasi and his expert guests talk networking, network engineering and related careers, emerging technologies, and more.
We use cookies to deliver our services.
By using our website, you agree to the use of cookies as described in our Privacy Policy.