I lasted three months at a financial services firm in continental Europe before I walked. The resignation letter took about twenty minutes to write. The conversation that followed it — the one with my peer who got what was happening — lasted decades. That imbalance tells you everything you need to know about what actually matters in broken systems.

The role looked perfect on paper. Senior architect role at a major bank, which in the late 1990s meant solid salary, interesting problems, and a platform to do real work. The first week was optimistic. By the second week, the structural issues started showing. By the third week, I was certain I'd made a mistake.

The problems were deep and orthogonal. Management structure that made no sense — people reporting to people who didn't understand their work, decisions made by committees that included people whose opinions directly contradicted each other's. Compensation that didn't match the role. A direction that shifted weekly. It's not that these things are unusual in large organisations. It's that they all compressed into the same three-month window, with nothing balancing them on the other side.

But there was one engineer — let's call him Marc — who understood what was happening. Not sympathetically. Not performatively. Just clearly. We started having coffee together at a local café when we needed air. He'd worked through similar things before. He wasn't trying to fix the organisation or fix me. He was just being present to the fact that this wasn't working and that I should probably leave.

I did leave. The resignation process was straightforward. The job went away. The company continues without me, absolutely no one surprised by my departure. The kind of thing that should be a non-event in a functioning labour market.

But Marc remained. Still does, in fact. Twenty-five years later, we're still in contact. Not in a "we grab coffee when we're in the same city" way, though we've done that. More in a "if something mattered, we'd both show up" way. We've stayed tangentially connected to each other's careers, asked real questions when the other person had a decision to make, maintained the thing that actually worked when nothing else did.

Here's what I've learned: **in broken systems, people are the only stable thing.** Everything else fails. The job fails. The management structure fails. The compensation fails. The direction fails. But the relationship with someone who *gets it* — who sees the broken thing clearly and doesn't pretend otherwise — that holds.

This applies everywhere systems are complicated. Open source projects die. Maintainers burn out. Technologies become obsolete. Organisations fold. But the people who understood the work, who paid attention to the same problems, who didn't waste energy pretending things were better than they were — those connections last. Not as a consolation prize. As the actual valuable thing.

The worst advice you can get is "just choose the right employer." The right employer is usually a fiction. What actually matters is choosing the right people. The people who won't lie to you about how broken things are. The people who'd tell you to leave if leaving made sense. The people who'd stay in touch after you're gone because the relationship wasn't contingent on the paycheque.

I think about that bank sometimes, usually when I'm reading about financial services failures. I wonder if the management structure ever made sense. I wonder if Marc is still there or if he eventually walked too. I wonder what he'd tell someone else arriving now with that same optimism I had.

But mostly I'm grateful for the three months because they gave me twenty-five years with someone who understands how to live in broken systems without pretending they're working. That's worth more than the salary ever was.