Exciting Announcement! In celebration of launching our AI Certification, we’re thrilled to offer a 50% discount exclusively. Seize this unique chance—don’t let it slip by!

Transcript

 

Welcome to the Aviz podcast, I’m Ilona Gabinsky and today we are joined by Ethan Banks, tech thinker, storyteller, and co-founder of Packet Pushers. So Ethan’s voice has guided a generation of network engineers, but today we’re going deeper. What really drives him, what’s changing in networking, and what lessons from outside tech have shaped how he leads.

Ethan, welcome to the show and we are very happy to have you. Thank you, Ilona, it’s nice to be here. You can ask some probing questions.

Okay, we’re getting into it today, I see. All right, let’s go. Let’s go right into that then.

So Ethan, you’ve built a career, a platform, and really a voice in this space, but what really drives you underneath it all? Yeah, a career and a voice in the space. A lot of us that are in the public’s eye, we’re in the public sphere, we’re content creators. If you want to get into a psychological discussion of what drives people like us, sometimes it’s the need for approval, I think.

That just looking for that need for approval. And when you make something, you create a podcast, you share your knowledge, you write a blog post. I mean, I started out as a blogger, as a writer, really, in 2007.

You get feedback, you get approval, you get people that were helped by what you wrote, or were interested in what you wrote, and wanted to talk, and you can find your tribe that way. And so when you’re looking for, if what’s built into your psyche is that need for approval, it’s fed when you write something, or in modern days, you post a short, or you create a YouTube video, or you do a podcast, and you get that feedback from it. It’s very, makes you feel good, makes you feel like, you know, people accept you, makes you feel like your tribe is there, and you’re all together.

What I have found is, because of this weird job I have, podcasting, is primarily what I am known for, and how people are familiar with my voice, I’ve met a ton of people. And because of that, I’m in this neat position of being able to connect people. Oh, I just met you, and you have this need, you got to meet this other person who you guys should talk, definitely, you know, and I connect people that way.

And I’ve done that countless times at this point, and love doing that. I mean, there’s real value in that, that kind of thing. Another thing that’s come up for me alone, and especially in the last year, let’s say, this is 2025 that we’re recording this, there’s a new generation of networkers that are coming along.

They’re in their 20s, they’re maybe in their 30s, but they’re early in their networking days. Who’s mentoring the younger generation that’s coming up and networking? And so that’s become a driving force for me, where it’s like, okay, we really got to talk to that younger generation and help them understand. It’s terrible to be in your 20s as a up and coming networker and inheriting the stack that people like me helped to build, or at least we’re got to see it being built.

When 20 years ago, networking was a simpler stack. And today, it’s a rather complex stack. And to have to figure that out is just, I would not want to be in that, in those shoes of trying to learn all of that from the ground up.

I know the foundational stuff and how we got to where we got. And so I think I have a perspective where I can share with people, this is how we got to where we got. And to understand what’s happening way up in the overlay that everybody wants you to focus on now, you got to start with what’s going on in the underlay and the fundamentals, because that’s still all there and it still all matters.

And being able to share a perspective on that has been as motivating as well. This is great. And this is great to see that this constant search for approval really drives you.

And that’s what I think really made you successful as well, because you’re kind of also never satisfied with the results. So you constantly need to move forward and forward and constantly getting this approval. So yes, maybe it’s a psychological barrier to some people, but it’s amazing to see how you actually turned it into your strength and using it.

Yeah. Yeah. It’s funny you put it that way in that people that are driven that way, whether it’s a need for approval or a need to feel successful or that along those lines, it is a driver, right? It is really a thing that helps you to accomplish things that maybe you wouldn’t otherwise achieve.

And so there is this upside. Sure. Psychologically, it’s some kind of a failing or a shortcoming I feel in some way.

But then if you look at how it motivates behavior, I mean, man, people like that tend to be great employees. They tend to go after certifications. They tend to take on responsibilities because they need someone like me.

Would someone please like me? Exactly. Exactly. That’s amazing how you can turn your biggest weakness into your biggest strength.

So that’s amazing. And by the way, thank you very much for sharing this with us. Sure.

And you also touched on that aspect, being able to educate the younger audience on what’s really important in networking. And we are all working in the networking field right now. From your perspective, what’s truly new you think in networking and what do you actually think matters most right now? Man, this is a big question because what has happened in networking is that we’re baking a lot of functionality into the transport layer, into moving packets around the network that before that was the thing the network did.

We move packets around the network, that was the part and parcel of it. That was the point of it. And it still is.

It’s just that as time has marched on, we’re now looking for this additional functionality to be baked into the network where you’re delivering a complex security design down at the packet layer and baking that kind of functionality into a switching fabric, let’s say, with some sort of an overlay. And there’s been various strategies that have happened there to create this complex stack. So one of the things that’s happened, and this is interesting, if I back up a second, one of the podcasts I started last year with Holly Metlitsky.

Holly is a pre-sales engineer at Juniper and she’s younger in the networking field. And I’m the old guy. And so we both play those roles in the podcast called N is for Networking, where it’s about fundamentals.

One of the questions we get from the audience for that podcast is, hey, can you talk about EVPN VXLAN? It’s like, we haven’t talked about VLANs yet, let alone VXLAN and EVPN. We haven’t talked about BGP as a routing protocol yet. And you want us to just dive into that? It’s like to set you up so that you could understand what’s going on in that overlay.

There’s so many other foundational pieces that you need to grasp to get your head around it. But yet that’s where that younger generation is, what they’re looking for to help understand. And that really does highlight paradigm shift in networking.

It’s kind of always been there in service provider networks on some level or another, some flavor of overlay, underlay, multi-tenancy and all of that. But now it’s becoming commonplace in almost every network, really, any network of size. So that’s a thing that’s come up.

Automation certainly is another thing. And that’s another challenge, I think, for those younger engineers, where I got to be an automation engineer. I got to know how network automation works and know all these tools and understand Python and all the tools that go around infrastructure as code and so on.

Well, yeah, but no, not now. What now you need to understand how packets move around a network. That’s what you need to understand, because you can’t automate a process you’re unfamiliar with.

You really need to know at a deep level, like how does a VLAN work? And then advance to how does routing work? Well, how does a routing protocol work? And understand as you deploy those fundamental technologies on a network, how that is done, what that workflow looks like, how you think about that, what a change control looks like. And then you can get to a point of being able to automate it, jumping in with both feet saying, ah, I’m going to start with Python. Are you? Because you don’t know what you’re doing.

You don’t know what it is that you’re automating, really. So it’s good to have those tools, and eventually there won’t be a CLI. But if you skip out on the fundamentals of how the network works, automation is going to be tough, because you won’t understand the risks that you’re introducing with your automation process.

You might be a heck of a coder, but if you don’t understand the fault domain that you’re introducing that code to and those processes to, you can, as I like to say, fail at scale, where you can break a lot of things in a hurry. You’ve just enabled your ignorance to run wild. And in a sincere way, I worry about that with younger engineers, because that’s another big change.

That’s another fundamental technology shift that really matters, going back to your question. But this whole hype, obviously, that’s happening right now in the infrastructure with AI, not just automation itself, but AI. What do you think about that from the networking perspective? Because I also talk to a lot of customers, and when we were just starting with AI, maybe a year ago, everybody was pretty skeptical about this, and AI, can it really do what I’m doing? Can it really help me? Will it take my job? So a lot of people are really concerned about that.

And right now, obviously, it is shifted towards that all people now understand AI is here. If I learn AI today, the better leader I will be tomorrow, kind of a thing. So everybody now understands that the AI is here, and we obviously have AI product already also in our portfolio.

But this is really an AI assistant to engineers, not something that will replace them. It’s more like making them more effective, or taking away the mundane work that they are doing right now. But I just wanted to ask you, what’s your opinion in AI, and where do you think it’s all going? And obviously, there are a lot of talks, Bill Gates is talking about that AI will take all of our jobs, and everybody else.

Bill says a lot of things, yeah. What’s your point of view on that? You want to do this one, huh? Okay. All right.

Let’s have this conversation. Maybe briefly, yeah. AI is an umbrella term that introduces a lot of different technologies, and we end up talking about them as if they’re one thing, but they’re not.

AI, from a networking perspective, has significant challenges in that everybody’s network is unique. So let’s take one component of LLM, or one technology that we’ve become familiar with as a society because of its wide use, and that’s LLMs, large language models. We interact with chat GPT, and we ask it a question, and it gives us back an answer, and it sounds intelligent.

It isn’t. It is basically doing mathematically-based word completion based on a lot of text training data to spit something out at us. There’s a use case there where we could take an LLM, and maybe it, from an IT perspective, maybe it helps us troubleshoot a particular situation because we’ve got a custom LLM in-house, a private model that we’ve trained on our own internal knowledge, and that will help us actually walk through troubleshooting a problem as a bot.

Yeah, I’ve seen those demos. They’re getting better and better and better and better. That’s real.

What about an LLM that can, hey, I need a BGP paragraph that is going to deal with this autonomous system and give me these sorts of policies and deal with these routes and so on? Yeah, maybe. Again, it depends a bit. Where LLMs are falling down in this space is how well were the models trained? How accurate was the data? How applicable were the paragraphs that the model was trained on to your specific environment? Networks are snowflakes in their current iteration.

Everybody’s network’s a bit different. They all do the same thing. They’re all kind of alike and pretty similar in a lot of ways, but different enough that it becomes less predictable to know how to train an LLM and that the resultant LLM spits out at you, assuming it’s not hallucinating, assuming it’s actually giving you something that’s even valid, will work for your network in the way that you want it to.

We still have not built as a networking industry a modular, predictable network that follows a specific engineering-style framework like you might have to build a house to code, like electrical code. There are standards that you’ve got to follow. We don’t have that in networking.

We don’t have those kinds of standards. We have best practices. We have pools of knowledge.

We have people’s experience and people’s preference. We have vendors that publish validation documents. But because everything’s different, this is a long way to answer the question.

I apologize, but I’m getting to a point here. AI is always going to be a bit challenged to just take over, because take over what exactly? What are we taking over? Well, you’re taking over this unique snowflake network that’s got these particular business requirements and has this particular vendor that it’s working with and these particular cost constraints. You have to train the LLM to be able to react to not a network, but to your network.

This is something that as we talk to networking companies that they specialize in AI, this is what they’re finding. We’re going in as a startup AI company dealing with networking data and having to learn, train our models on that customer’s dataset to maximize our product’s value for them. It takes a while.

It takes time. Can AI take all of our jobs? I mean, it’s not inconceivable. I’ve been talking about LLMs.

There’s all these other AI use cases, too. But right now, I don’t see the path there right now. I don’t see that we’re going to get there.

Is it going to help augment an engineer? You bet. Is it going to pull trends? Is AI not LLMs, but more machine learning-based training of data that is trained to look for trends more along the lines of statistical analysis or root cause analysis where it’s taking multiple data sources and it can tell you, this event happened on the network which affected the Kubernetes cluster in this way, which affected these apps running on the Kubernetes cluster in this way. That’s why user experience declined from 1042 to 1153 on this particular day.

AI is great for that kind of stuff. Does that replace an engineer? It does not. Does it help the engineer get to root cause, MTTR, more quickly? Yeah, you bet.

You bet. That’s great. Can we make a help desk agent, that triage person who’s taking a call in from someone who’s struggling with some kind of a bad user experience problem or inability to connect to a service? Could AI, like a chatbot, replace an engineer? In a sense, yes.

Maybe the help desk person can resolve the problem with the help of the chatbot, but that didn’t replace an engineer like replaced headcount. It meant the engineer didn’t get tied up dealing with supporting that ticket. It meant that they were able to stay focused on more important higher-level things.

Again, it augmented them as opposed to replace them. It replaced them on that call, and that’s a thumbs up. You want that, but actually reduce headcount in IT with AI and networking? Again, I’m not saying it’s impossible, but I don’t see the path there yet, not from any of the AI products I’ve seen.

They’re all helpful, but this is an engineer in a box now. I don’t see it. Exactly.

Ethan, thank you very much for sharing that. That’s exactly what my opinion is and what we are seeing as well. We are building AI to really help and assist engineers with their mundane work so that they can actually focus on something that they like and want and need to do.

Mundane work, yes. As a guy who used to have to travel, she’d be like, oh gosh, I have to log into these 10 boxes to figure this out. I don’t want to do this right now.

Yes, I want AI to help me not have to do that. Yes, exactly. Awesome.

That’s great. It seems like you’ve been really this networking expert and you had this all hands-on experience, but if networking disappeared tomorrow, what would you do instead? Have you ever thought about that? What other paths have always intrigued you? There’s several things. Some of them are more hobby things and some of them are more career-oriented things, but let’s talk about some hobby stuff.

I love music. I sing. I’ve tried my hand at guitar a little bit, but I’ve never really had the time to put into learning an instrument.

I’d love to do that. That wouldn’t pay the bills, but man, it’d be fulfilling. I would love to do more with music.

Love that. I am a trail runner and a hiker in New Hampshire. I like to run.

I like to do ultra marathons. I’d love to devote more time to training or to hiking long-distance trails. The Appalachian Trail is on my bucket list at some point.

As in nonstop, I want to put a bag on my back. I actually live close enough to the Appalachian Trail from where I reside that I could literally put my pack on my back, walk out my front door, and get on the Appalachian Trail. I could be on trail from my front doorstep in probably, I don’t know, two, three hours of walking, something like that from where I live.

And then not come back until I’m done. That’s a thing. Yeah, yeah.

And walk home too. Literally walk back to my front door. That would be, it’s a fantasy I’ve indulged in because it’s there.

I just need five, six months to do it. So that’s a thing. But for career stuff, I’m at a point in life where I’m much less interested in tech and digital and screens and content creation and all of that, and more interested in my local community.

So I live in a small town. We got challenges in this town. There are people who struggle with opiate addiction, for example.

There are homeless people in this tiny community. Yeah, I run rail trails here and there are little tent cities of folks that are in need. And we have different groups within our town that reach out.

I’d love to work with those towns, work with our community more, work with some of our town’s outreach and help people that need help. My life is, I couldn’t ask for more out of life. I don’t need more.

I don’t, how can I help other people has been uppermost in my mind. I’m a man of faith and I do some work like that with our community church, just reaching out to different groups that need help. You know, we go to the, go to the shelter and share a meal with people, you know, help feed them, you know, stuff like that.

That’s way more interesting to me now than, because your mind kind of shifts that way. I think when, especially when you see the need and in a small town, the needs are a little more, you feel it more, you see it more because it’s like, it’s right there. There’s only 4,000 people that live in this town, something where I live, something like that.

And that’s a, that’s a big town. So when you see someone who is really hurt and really in need, you’re like, okay, they’re right there in front of me. It’s not something on a TV screen happening in San Francisco or something.

It’s like, it’s, it’s real. It’s right here and people need help. How do you help? And it’s not enough to just give money to a charitable organization.

That’s not a bad thing to do, of course, but if you can actually get involved, um, that’s good. And I, I, I would like to do more of that kind of stuff where I not doing the packet pushers thing that keeps me very busy right now, to be honest, I mean, you know, yeah, I don’t know. It’s after so many years in tech, you do burn out on it a bit where the novelty of it, I mean, my brain still loves this stuff, right? I have an engineer mind that loves tech and loves solving problems and loves building things that, uh, that solve a business need.

Let’s say I, I don’t suppose I’ll ever get away from that. Um, you know, I like to teach, I could see a career of, uh, you know, I’ve done some classroom instruction over the years. I’d love to get back into that.

It’s a hard game to get into. Um, but yeah, I’d love to get back to, uh, some classroom instruction and so on. I mean, I kind of do that informally now in my podcasting work, but, um, but at a point you get tired of the tech in the sense that is a cycle to it.

The, it, it has a life, it breathes. And you see, after you see enough cycles, come and go and come and go and come and go and try to play the guessing game. Is this the thing I need to pay attention to? Is this going to work? 15 years ago, we were talking about ethernet fabrics almost constantly.

It was trill and it was going to be shortest path bridging. How much of that technology has gained wide acceptance? Very little of it. It’s still out there.

You can still buy SPB solutions. For example, there’s still some big trill installations in the world as I understand it and so on. Did it become the way we do networking by standard? Nope.

Um, I don’t mean standard like, you know, RFCs, but you know, it’s not the typical way you build a network. If you bet everything on learning, whatever that tech was, you were important for two or three years and then it faded out and now no one cares anymore. By and large, for most of that tech, it’s become very specialized.

You go through enough of those cycles. It’s like, is this the thing I got to care about? Is this thing working through the IETF draft system and it’s going to become a standard? The thing I got to care about? I don’t know. You know, cloud are, are we all moving to cloud or are we all coming back out of the cloud? I don’t know.

Oh, we’re not. It’s it’s it. We’re going to just do everything.

We’re going to stay in cloud and we’re going to come back to our own data centers and we’re going to turn it into a cloud too now. Oh, okay. You know, it just, it’s like the old tech never dies.

There’s an endless supply of new tech and trying to keep your head in the game and trying to wrap your brain around what all’s going on and trying to formulate a career where you’re like, is this the thing that’s going to put food on the table for my family? I don’t know. You know, should I learn, should I go get that certification? Maybe. It’s a lot of work.

It’s kind of hard. It’s going to be expensive and take a lot of my time. And then if no one cares, it doesn’t help me pay the bills.

Why did I do it? There’s a certain grind to being in the industry as a, as an engineer that can be frustrating. I’ll just go into management, you know? Okay. And then some people do that.

Well, okay. Then you’re not trading on your technical skillset as much as you’re trading on now your ability to manage people, which is a whole different skillset that you got to learn. So do you want to go in that world? Because managing people’s hard.

It’s really hard if you get it to be good at it. Yeah. It’s, it’s not like, oh, I’m a good engineer.

So now I should lead all these people. No, those are different skills. Those are different things.

I mean, maybe you’re good at it, but don’t assume just because you’re the, the awesome person who can make the thing, do the thing that now you should be leading a group of people. Maybe, but maybe not. And don’t assume your technical prowess translates into effectively leading people.

It’s a career step people take because they get a little tired of the tech grind and learning the new things. But, but Alona for me, I mean, I’ve been doing this since 95. I started, you know, really in the, the, you know, the tech thing and I’ve somewhat reinvented myself as a content creator, which has kept it alive.

And I love doing that part of things, but at some point I don’t want to read a new press release from somebody and try to make sense of it. I don’t, I just, I just don’t. At some point I run out of gas.

Well, Ethan, thank you very much for sharing all that. And I really wish you that you will have all this time to do what you really want and enjoy. So maybe you will be able to use AI for some things that’s taking a lot of your time and just, you know, do things more efficiently and have much more time for things that you actually enjoy and want to do.

So I really hope you’ll be able to do that. And I really hope, you know, we’ll all be able to do that in the near future. This is great that you’re sharing that.

And, you know, I really see that maybe, okay, let’s give this some part of AI, do that. And maybe it will decrease my working hours 50% so that you can actually go hike and bike and I guess start feed all the press releases I get into AI as I was just saying, Hey, summarize these for me, explain like I’m five, give me three lines and tell me what they’re trying to say. So I don’t have to read it.

Yes. Yes. So that you can go run on the trail and, you know, Oh, I do get out some, I do get out some, don’t get me wrong.

I mean, I’m out that I’m out there, but you know, you’re never free when you got a job, it is what it is. Yeah. No, but I hear you.

Absolutely. Ethan, we are really, you know, at the top of the hour. So any last, last piece of advice you would like to share with the audience, anything you would like to say before we wrap up? If you are new to this, don’t be so don’t be overwhelmed by all the tech that’s out there.

There are some major lanes you can pick and scale up on and have a pretty good opportunity for jobs. If I made it sound bleak, I mean, I don’t, I didn’t mean to, you’re listening to a guy who’s been at this for three decades and has got a bit of cynicism baked in. Tech can be a great and very rewarding career in a lot of ways, especially if you’ve got the head for it, you got the right mind for it.

But get, get advice from some people that have been around and, and then pick a path and stick with it. What technology you want to be good? You want to be the cloud person? You want to be the network person? You want to be the cybersecurity person? What do you want to do? You pick your lane and then, and then, then jump in. And, and when you do that, be a little patient would be my, my chief thing to keep in mind if you’re young.

It’s been great having you. And, you know, all the insights that you shared are really precious and especially all the advice and, um, to the younger audience. I think it’s very important.

I also have, uh, uh, two younger sons and one is in college and another one is still in middle school. So thank you very much. I really see a lot of value in what you’re sharing right now.

And I will share this podcast with my kids so that they can listen to it and get some great, uh, information out of that. So thank you very much. It’s been great having you.

And hopefully we’ll meet again pretty soon. Thanks a lot. I appreciate the opportunity to come on.

It was a, it was a privilege. Yeah. Thanks a lot.

Contact Us

Episode 7: AI in Networking: Ethan Banks on Burnout, Automation & What’s Next

  Welcome to the Aviz podcast, I’m Ilona Gabinsky and today we are joined by Ethan Banks, tech thinker, storyteller, and co-founder of Packet Pushers. So Ethan’s voice has guided a generation of network engineers, but today we’re going deeper. What really drives him, what’s changing in networking, and what lessons from outside tech have […]