近况-en

A couple of days ago I submitted the internal review draft of my thesis—about five months of capstone work are basically wrapped. From barely telling Docker and Kubernetes apart to prepping for CKA and CKAD, my cloud‑native picture keeps deepening. My own verdict on the project: not satisfactory at all. The topic was too big, so whatever I did felt insufficient and never “good enough.” I wanted to do a lot at first, but as the project went on I saw how limited my bandwidth was and kept compromising. It reminded me of my freshman course project—starting without enough understanding of the work. Same mistake as a senior as in year one—like treading water—and that sometimes really got me down. In the end it’s still too little project experience and not learning enough from past mistakes.

One sleepless night I finally made peace with it: imperfect is normal—utterly normal. Imperfection is where improvement lives. I used to rush to the next task whenever something wasn’t good enough, so I never stacked depth anywhere—that was wrong. If it isn’t polished enough, take the lesson and do it again (cue Again from Your Lie in April). The next pass should be better and teach more—if you’re committing to a field long term, you have to keep drilling; you can’t decide you “know enough.” “Half of a hundred li is ninety.”

Take observability in this capstone: I did a lot of work but had few concrete metrics to show. Early on Kubernetes was mostly the dashboard; packet capture was mostly Wireshark. Since this run didn’t observe enough, next time I should use more serious tooling—maybe tcpdump for packets (haven’t dug in yet), Prometheus-class stack for cloud‑native monitoring.

The more I learn about cloud native, the more I feel the early years should go deeper. Simple example: with KubeEdge, when an edge pod misbehaves, the cloud often has no idea what happened. I first thought it was my proxy config; reading an OSPP 2024 topic made me realize—they simply hadn’t implemented it yet… Building on others’ work always means constraints. Top-level apps also lean on low-level principles; I’m too green to unpack that (runs away). So grad school can’t be restless—foundations matter.

Anyway, the capstone chapter is closed. Then I went to a close friend’s birthday—KTV, a bit stiff, people looking out for me (thank you; I hope one day I’m grown enough not to need that). One thing is clear: everyone’s genuinely kind—I like that vibe. First time at KTV was with my sister; I sang one song, the Xuan-Yuan Sword drama theme One Kiss, Wasteland (I was obsessed with that show, lol).

I show up to unfamiliar social settings because I believe social skills are learnable—as you experience more, parts of you change. People have many sides; different contexts surface different ones—not always “performing,” just that other sides don’t show as easily in the moment. So binary labels like MBTI feel like they cage growth. “I’m I*** so I’ll act like…”—no; I might surprise you.

The next day I met someone I’d known on the Fediverse for half a year—finally talking in person about online life. Two hours, really fun. We roasted Anno; they said getting close to Gundam brings misfortune, which only made me more curious. They also talked academia—I learned a lot. Face‑to‑face is still my favorite.

After submitting the paper I started a Misskey instance—domain neuromansser.tech (neuromancer → neuromansser). Roughly like this now:

vnil.de was great—the admin is kind and helpful; I felt comfortable there with folks. Thanks, vnil.de friends—if you’re still shopping for an instance, consider vnil.de~ I’m an adult now; time to look after my own server (hey!).

Once the instance was mostly up, I migrated my blog—finished last night. Not much clutter; overall clean (not as minimal as the admin’s, and JS included), and I’m okay with it.

That’s the snapshot ^ ^—see you~