Learning Out Loud

Learning Out Loud

This week on The Art of Network Engineering podcast, Andy Lapteff 🛠️💬 sat down with friend and frequent instigator of “weird lab stuff,” Lexie Cooper . We covered space stuff, learning in public, why messy home labs are a feature (not a bug), and the pressure engineers feel to look perfect when the value is often in the struggle.


Learning in Public: Why Vulnerability Wins

We've been streaming Andy's journey learning Python. It’s messy. Sometimes it’s reading from a textbook. Sometimes it’ not-so-quietly yelling at a for-loop. And yes, Jeff Clark jumped on one episode and told us to “just code” with AI’s help. (Love you, Jeff.)

Here’s the real talk:

  • The gap is real. Most network jobs now include automation in the requirements.
  • The iceberg effect is real. Polished YouTube tutorials hide the grind. You don’t see the 40 minutes of “why won’t you run” edits.
  • The audience needs the messy middle. Seeing someone struggle, in a competent, curious, honest way, helps more people start.

Lexie’s take: "we’ve created an aesthetic around perfect labs, perfect racks, perfect code. But the most useful thing to share is the process, including the failures and fumbles."

“There’s magic in vulnerability. If I can learn in public and be lost, maybe it pulls someone else along.” - Andy

Weird Lab Stuff™ (And Why It Matters)

Lexie thrives in what she calls weird lab stuff. That’s not resume bullet points. It’s curiosity with a camera rolling.

Recent experiments:

  • Cutting cables on purpose. Take two auto-negotiating NICs, snip the blue/brown pairs, and connect only the orange/green. Outcome? Auto-neg downgrades to 100 Mbps (because you need all eight conductors for gigabit). It’s obvious in hindsight—but feeling the negotiation happen teaches layer-1/2 intuition you can’t skim from a doc.
  • Oscilloscope on the wire. Turning off features that should stop certain link pulses…and watching pulses anyway. The kicker? Behaviors vary by PHY—the physical transceiver silicon that bridges the ASIC to the medium and houses Ethernet MAC-layer logic. (No, not your MAC table—that typically lives in the switching ASIC. Different “MAC,” same layer, different role.)

If that last paragraph felt new: same. Most cert tracks barely touch PHYs, reconciliation sublayers, or PMD specifics. You don’t need EE depth to be great at networking, but peeking under the hood sharpens your instincts when the “impossible” happens on a wire.

“People think perfect cable management equals ‘real.’ In a learning lab, perfect often means ‘unused.’ The messy stuff is where the learning is.” Lexie

TikTok vs. Twitch vs. YouTube (and How to Actually Stream)

Quick streamer notes from the trenches:

  • Twitch: best for multi-scene, multi-camera, polished OBS setups.
  • TikTok Live: unmatched for “flip phone open and go.” Great for reach, perfect for spontaneous lab vibes.
  • YouTube: we stream our episodes via Riverside to YT; it’s already wired into our workflow. We're still figuring out multi-platform streaming without summoning a gremlin.

Pro tip we learned the hard way: load your “Starting Soon” bumper inside OBS as a scene, not as a screen-share of a looping MP4. Your future self will thank you.


Career Talk: Networks, Automation, and Being “Allowed” to Be Wrong

We went somewhere a lot of us avoid: the pressure to be infallible, especially when you move from operator to vendor.

  • The persona tax. As your platform grows, it can feel like you’re “not allowed” to ask dumb questions. But the industry doesn’t need more invulnerable experts; it needs more honest ones.
  • Automation anxiety is universal. Many network pros don’t want to become programmers, and many don’t have to. But some fluency in Python, Git, and toolchains is increasingly part of the job. AI helps, but basic programmatic thinking still pays dividends: data types, control flow, “how do I think like code.”
  • There’s still a human in the loop. Automation isn’t a panacea. When the unexpected happens, it’s comforting, and often critical, to have a person to reason through it. Especially when the stakes look like…space.


Space Dreams: How Far Would You Go?

We also let ourselves dream. Would you go?

  • Mars? Hard pass for both of us (for now). Months in a tin can = existential nope.
  • The Moon? Lexie: yes—if it’s autonomous. (Same.)
  • Pilotless planes & pilotless rockets. The automation bar moves. Comfort follows capability. But for the edge cases, the “what now?” moments, we still want humans nearby.

“I trust automation. I also trust having a person when something weird happens.” —Lexie

Hiring, Mentorship, and What’s Next

Lexie’s team is hiring to backfill her as she shifts to a related project (yes, it’s very cool). On-site work is part of the gig, which narrows the field, but the interview panels have been strong.

Lessons she’s absorbing on the other side of the table:

  • The req is a wishlist, not a gate.
  • Fit and curiosity often matter as much as checkbox tech.
  • You learn a ton watching senior engineers probe, guide, and evaluate.


Why This Conversation Matters

Because the industry is changing, and we’re all renegotiating our identities:

  • From CLI lifers to automation-aware engineers.
  • From polished outputs to visible process.
  • From lone wolves to community learners.

If you’ve been waiting to start your lab, your stream, your learning path, this is your permission slip. Start ugly. Hit “Go Live.” Snip a cable (safely). Break something you can fix. And let people see you learn.


Watch, Hang, Build With Us

Thanks for listening, reading, tinkering, and learning out loud with us. See you in the lab, and hopefully someday on a beach with a perfect line of sight to LC-36.

To view or add a comment, sign in

More articles by The Art of Network Engineering

Explore content categories