Make the Call

Three days into Boston’s biggest blizzard since 2015, I’m pushing a snowblower three blocks down to clear the path for middle school kids in my neighborhood. My back aches. My hands are numb. And I can’t stop thinking about my father. He’s the type who’d come out to shovel himself instead of waiting for me to get there.

I’m also thinking about AI. Everyone’s waiting for clarity on what Large Language Models (LLMs) will become, how they’ll reshape work, which jobs survive. We’re all frozen, trying to calculate outcomes we can’t see.

My father never had that luxury.

Oil Lamps in the Early 90s

I grew up in a post-war “strategic hamlet” in South Vietnam. We used oil lamps. When we finally got a TV, it was black and white with exactly one channel and thirty minutes of Soviet cartoons. When I tell my kids these stories, they think I’m describing the early Colonial era.

I was growing up in the early 1990s.

Electrical lines were crossing through my village, but the government had no interest in lighting us up. They were running power to the train station beyond us. We were just in the way.

What can you do against a government agency? You don’t have the data, leverage, or the connections.

My father got the whole village together and negotiated: if we paid for our own equipment, would they allow us to connect? When they said yes, he went to the city, worked his connections, and throughout the whole week, we saw reels of electrical wire and light poles slowly gathering in our front yard.

After a few more weeks of negotiating, being as flexible as possible with workers on and off the books, we got electricity. He made the call before he had certainty, because his village deserved better than oil lamps.

Đả Đảo

Vietnam was still very much a post-war Communist country. The head of our commune was embezzling money and creating real duress for poor families who had nothing to spare.

My father got several key figures together. Meeting after meeting, they planned every step, scripted who would say what at the next commune assembly. I listened to them planning, not fully understanding but sensing its weight.

When the day came, I stood outside and watched through the window.

One by one, each person stood and said their piece. Laid out evidence. Made their case. When things looked like they weren’t progressing fast enough, they activated something I’d heard them discuss during those night meetings in our house. They all stood up and started shouting: “Đả đảo…” (which means “down with…”) until the whole board relented and initiated a vote of no confidence.

The corrupt official was removed. A new head was installed.

My father didn’t wait until the outcome was guaranteed. He built a coalition, planned meticulously, then moved when the moment demanded it.

Go Home and Die

Sometimes people come up to my father when we’re visiting the village after many decades living in the US. They express gratitude, tell me things I never knew about my father.

“He saved my wife.” “He saved my son.”

One man, voice still shaky after all these years, told me this story: the commune hospital had declared his wife couldn’t be cured. They told him to take her home and get her affairs together. “They practically told us to go home and die.”

He came to my father. Together, they loaded her onto a moped and drove her to the city. My father used his connections to secure a bed in the big hospital.

She made it.

The husband refused to accept the verdict. My father agreed to help when he could have said it wasn’t his problem. Neither had complete information or guarantees, but they had a moment that demanded action.

No One Has the Data

We’re in a similar moment with AI. The data is incomplete. The outcomes are uncertain. Every prediction about what LLMs will become, how they’ll reshape industries, which capabilities emerge next, all of it is contested terrain.

Some people are waiting for clarity before they act. Waiting for the perfect framework, definitive research, or the consensus view.

That clarity isn’t coming. Not in time to matter.

Part of my work is helping clients build great engineering departments. Right now, teams are asking: do we hire the same headcount or fewer engineers with AI tooling? Do we restructure workflows around AI-augmented output? Nobody has clean data. The companies moving fastest are making calls based on conviction, not certainty.

The question isn’t whether you have enough information. You don’t. None of us do. The question is whether you’re willing to move anyway, to build coalitions, to plan meticulously, and then to act when the moment demands it.

My father didn’t have electricity, money, or power. He had conviction and community. He made calls that everyone else thought were impossible, not because he knew they’d work, but because waiting wasn’t an option.

More Than Just Yourself

One thing that keeps nagging at me as I clear a path for all the kids in the neighborhood, mine included: you have to care about more than yourself.

The electricity wasn’t just for my father’s house. The corrupt official wasn’t just affecting my family. The woman on the moped wasn’t his wife or his relative.

He cared anyway.

That’s what separates useful action from noise. If you’re making bold moves on AI purely for your own positioning, you’re just adding to the chaos. But if you’re thinking about your team, your community, the people downstream of your decisions, then the hard calls become clearer. Not easier, but clearer.

We got ourselves a breather after 3 days, but more snow is coming this weekend.

And I’ll keep thinking about what it means to make the call when the data is incomplete, when certainty isn’t coming, and when the people around you are counting on someone to move first.

What hard call on AI are you avoiding because you’re waiting for data that isn’t coming?