Management is just the art of repeating the same ideas to different people without losing your temper.
Two runners broke the two-hour marathon barrier in London today. Genuinely impressive. But objectively, my 6-hour marathon is three times the effort. I’m still waiting for my press conference.
"You shouldn't use AI for coding, it can produce insecure, unscalable, unreliable code." You shouldn't cut vegetables with a knife — it can cut your finger. You shouldn't use a lightbulb — electricity can kill you.
Doctor Strange: I looked forward in time, I saw 14,000,605 possible futures. Me: Bro, that’s not magic. That’s anxiety disorder. I do it every morning before coffee.
Everyone's debating whether AI gets too powerful, too biased, or too regulated. The scenario I find most plausible for actually slowing it down isn't getting nearly enough airtime. Economics.
I’m from Ukraine but have lived in Germany for 12 years. I can’t call myself German yet, but I’m not fully Ukrainian either. Last time I was in Ukraine I ordered an Uber. The driver accepted, then called me immediately:
Can we normalize not spending the first 10 minutes of a 30-minute call doing intros? You have a calendar invite. It has names. LinkedIn exists. Get to the point.
Software engineers who think their job is to write code are like civil engineers who think their job is to draw blueprints. The blueprint is not the bridge. The bridge has to stand up. Handle load. Survive weather. Be maintained by people who weren't there when it was designed.
Writing code is the part of the job most engineers actually love. It gives flow, clarity, a sense of accomplishment. Everything else — meetings, tickets, coordination — we tolerate for it. Now it’s the part getting automated. What’s left is everything we liked the least.
One benefit of awful eyesight: bad CGI doesn’t exist. Everything looks fine to me.
There’s a state where you’re not doing the work — not because you’re tired, just… not doing it — but you also can’t relax because it’s still hanging over you. Feels like a race condition. Only way out is through. Hate it.
Harry Potter movies were, at best, mid — and a poor substitute for what are actually incredible books. I’m tired of pretending otherwise. Hoping the new adaptation finally does justice to the books.
Where’s the line between “I’m lazy and don’t want to go to the gym” and “I’m listening to my body and taking care of myself by skipping today”?
Fun fact about Pliant #3: Most startups can’t afford to “do tech right” at the beginning. Time to first release usually wins over everything else. In our case, external dependencies — vendors, licensing, legal work — bought engineering something rare: time. We used it.
Premortems are one of the most underrated reliability rituals. Instead of analyzing a failure after it happens, you gather the team beforehand and ask: “Imagine this project failed — what went wrong?” It’s much easier to spot risks before launch than to explain them after.
Most systems don’t collapse. They slowly normalize worse behavior.
Managers who obsess over velocity usually don’t understand what slows teams down.
Most ‘overengineering’ accusations come from people who’ve never been on-call for the thing they underengineered.
I’m at the stage of my career where sleep is a feature, not a luxury.
AI won’t replace taste. And taste is still the hardest skill to hire for.
Fun fact about Pliant #2: All of our tech teams follow the same naming pattern: two words starting with the same letter — an adjective and a noun. We started with Fancy Features and Mighty Machines as placeholders until we came up with something better. Seven teams later, we didn’t
GenAI is great at answers. Judgment is still a human bottleneck.
There’s no such thing as ‘simple’ in distributed systems. Only familiar.
Every abstraction saves time — just not always for the person who wrote it.
Fun fact about Pliant #1: In December 2020, when we issued our first card as a B2B payment card provider, the very first transaction was buying a giant plush pink unicorn 🦄 Priorities were clear from day one.
Experience is knowing which problems are not worth solving.
If your AI strategy is replacing humans instead of amplifying them, you’re just automating mediocrity