500 Hours of Foosball, 20 Hours on Your Career
There's a lot of fear in tech right now. Learn AI or get left behind. Learn automation or become irrelevant. Learn Python, cloud, APIs, or whatever new technology showed up in your feed this morning.
I've contributed to some of that urgency myself. After spending years resisting network automation, I eventually watched the industry change around me, and that experience changed how I thought about my own career. I started talking more about the importance of learning automation, AI, and other technologies that are reshaping network engineering because I didn't want other network engineers to make the same mistakes I did.
During a conversation with Mike Bushong at AutoCon 5 in Munich, though, Mike challenged some of that thinking. He wasn't arguing that network engineers shouldn't learn new things. Quite the opposite. He challenged the idea that fear is the right reason to do it.
Mike pointed out that we tend to overestimate the impact of new technology in the short term and underestimate it in the long term. Network engineering isn't disappearing tomorrow, and neither is the need for deep technical expertise. There will probably be network engineering jobs that change very little for years. At the same time, much of the new growth in our industry is likely to favor people who understand automation, AI, software practices, and other emerging technologies.
That creates a more useful way to think about career development. Instead of asking, "What do I have to learn so I don't become irrelevant?" Mike suggested thinking about the skills that create more options for you.
Create options instead of trying to predict the future
Nobody knows exactly what network engineering will look like ten years from now. We can make educated guesses, but trying to identify the exact skill that will "future-proof" your career is probably a losing game.
Mike described learning new skills as buying lottery tickets, except unlike an actual lottery, you have some control over the number of tickets you hold. Learning some automation creates another possible path. Understanding AI well enough to use it effectively creates another. Developing leadership skills, improving your communication, or learning more about the business your network supports all create additional opportunities.
None of those skills guarantees a particular outcome. They give you optionality.
That doesn't mean everyone needs infinite options. Someone who is seven years from retirement and values stability may make very different investments than someone with thirty years of career ahead of them. There's nothing inherently wrong with either choice. The problem comes when career advice becomes binary: automate or you're obsolete, learn AI or someone who does will replace you, learn this technology right now or your career is over.
Mike's point was that fear-based rhetoric can make change harder. If you convince an entire profession that everyone is perpetually one technology cycle away from irrelevance, people don't necessarily make better decisions. They can become defensive, overwhelmed, or resistant to the very changes you're trying to encourage.
We can acknowledge where technology is going without pretending everyone who doesn't immediately follow it is doomed.
But there's another side to this argument, and this is where the foosball story enters the conversation.
The 500-hour question
Mike once worked with a developer who desperately wanted to move into leadership. Every few weeks, they would spend hours talking about his career. He felt like he wasn't getting opportunities and that other people were preventing him from progressing. Mike would suggest things he could learn or areas where he could improve, but nothing really changed.
The guy was also extremely good at foosball.
He played in the morning, at lunch, and again in the afternoon. Eventually Mike started doing the math. If he was playing roughly two hours a day for about 50 weeks, he had invested somewhere around 500 hours in becoming better at foosball over the course of a year.
Then Mike asked him how many hours he spent improving his career over the course of that year.
That question gets to the part of career development we probably don't talk about enough. We can debate AI, automation, layoffs, certifications, market conditions, employers, managers, and every other external variable, but at some point we have to look at something much more uncomfortable: how much effort are we putting into the outcomes we say we want?
Mike uses a version of this lesson with his kids in sports. If your expectations are at one level and your effort is substantially below them, you eventually have two choices. You can increase the effort or lower the expectations. Maintaining a huge gap between the two is a recipe for frustration.
The same relationship between effort and expectations applies to your career.
Think about last year. How many hours did you deliberately spend becoming better at what you do? Did you read a technical book, build a lab, learn a new technology, attend a meetup, write something, teach someone else, experiment with an LLM, learn more about your company's business, improve your communication skills, or build something that didn't exist before?
Twenty hours over an entire year is less than 24 minutes per week. Yet it's easy to expect substantially more responsibility, more money, better opportunities, or greater job security without making even that investment.
This isn't an argument for spending every waking hour grinding on your career. It's an argument for being honest about the relationship between what we put in and what we expect to get out.
I have to ask myself the same question. I can easily spend an hour at night staring at my phone, bouncing between apps and convincing myself that I'm learning something. An hour a day is 365 hours a year. The issue often isn't that there is literally no time available. It's what we're choosing to do with the finite time we have.
You don't need to learn everything
Once you decide to invest more in yourself, another problem appears almost immediately. Look at the technology landscape and the list of things a network engineer is expected to know is ridiculous: Python, Ansible, Terraform, Git, CI/CD, containers, cloud, AI, APIs, telemetry, MCP, and whatever comes next.
Mike's advice here is much more practical. You don't need to improve along every axis, and trying to become an expert in everything isn't the goal. Instead, build enough exposure to different areas that you can recognize opportunities and then go deeper where it makes sense.
He compared this to how dendrites form in material science. A tiny protrusion on a surface attracts more material. As material accumulates, that protrusion becomes larger, which gives even more material somewhere to attach. Knowledge can behave similarly.
If you know absolutely nothing about a subject, information about it can fly past you without sticking. Learn a little and suddenly you have context. The next time you encounter the subject, you recognize it and connect the new information to something you already know. As those connections accumulate, learning becomes easier and knowledge starts to compound.
That means you don't need to spend 500 hours becoming an expert in every technology that might matter. Sometimes you need to spend five hours getting enough exposure that the next five hours become much more valuable.
This is probably one of my favorite lessons from the conversation: build enough exposure that knowledge can compound, then go deep where it matters.
Mike is able to speak intelligently about a wide range of subjects because he developed some exposure across many different areas. When one of those areas becomes important to his work, or simply becomes interesting enough, he can double or triple down on it. Over time, that creates a combination of breadth and depth.
AI could make that depth even more valuable. As access to surface-level information becomes commoditized, there may be less incentive for everyone to memorize it. Mike believes that could create something resembling a barbell, with a large group of people who have broad but relatively shallow knowledge and a smaller group with extremely deep expertise.
If that happens, deep networking expertise doesn't become worthless. It may become scarcer.
The CCIE who understands what is happening underneath all of the abstraction still matters when the abstraction breaks. The automation engineer benefits from understanding the network being automated. The person using AI to generate code benefits from understanding whether the output makes sense. Tools can abstract complexity, but they don't eliminate the systems underneath them.
Start before you're ready to be an expert
There's another problem in our industry that makes all of this harder: we sometimes expect people to become experts before they're allowed to be beginners.
Want to learn network automation? Great. Install a pile of tools, understand Git, know Python, figure out Linux, configure your environment, understand APIs, learn YAML, and then maybe you can finally automate something useful.
The barrier to entry becomes so high that people never start.
A better approach is to lower the initial investment. Spend a few minutes with something before you spend hundreds of hours on it. Run the example. Ask an LLM to explain something you don't understand. Clone the repo. Break something. Fix it. Learn enough that the next time you encounter the subject, it isn't completely foreign.
You don't need to invest 500 hours tomorrow. You need to invest enough today that tomorrow's knowledge has somewhere to attach.
That's where our conversation landed. We started by talking about network automation, but this isn't an automation story, and it isn't an AI story either. It's a story about agency.
Technology will change. Markets will change. Some skills will become more valuable and others less so. We don't control most of that, but we do control whether we're paying attention, whether we're curious, whether we expose ourselves to new ideas, and whether we're willing to occasionally go deep.
Most importantly, we have some control over how much effort we're willing to put behind the expectations we have for our careers.
Instead of constantly asking what you have to learn so you don't become irrelevant, ask what you want from your career. Then ask the harder question: How many hours did you invest in getting there last year?
If the answer makes you uncomfortable, that's useful information. You don't need 500 hours. You don't need to chase everything, and you don't need to live under constant fear that AI, automation, or the next technology wave is coming for you.
But you do need to keep growing. Build some exposure, follow what interests you, go deep where it matters, and give that knowledge enough time to compound.
Just make sure your expectations are reasonably aligned with your effort.
Listen to the full conversation with Mike at AutoCon 5: 500 Hours of Foosball, 20 Hours on Your Career.
More from Mike Bushong on the show: his guest page.