Social Good Is a Daily Choice, Not a Grand Project

When people hear “social good,” they picture something huge: AI that forecasts floods for two billion people, models built around the UN’s Sustainable Development Goals, a company-sized initiative with a press release. Those matter. But if I’m honest, that framing lets most of us off the hook - because I’m not going to ship a flood-forecasting system this week.

What I’ve come to believe is simpler: social good is a daily choice. It’s the willingness to pay a small personal cost - time, ego, credit, comfort - because the outcome is better for more people. And as an engineer, I get to make that choice at my desk far more often than I get to make it on a stage.

Yielding beats “winning”

The most common place this shows up is a design review or a pull request. There’s a version of engineering where you dig in on your approach to win the argument, or block a change to teach someone a lesson. It feels good for about a minute. But a blocked PR is like a car that refuses to let anyone merge: you protect your lane and jam the whole road.

The pro-social move is to yield when the cost of yielding is small and the benefit to flow is large. Disagree, make your case once, and then commit. Merge the junior’s reasonable approach instead of your slightly-cleverer one. Optimize for the person who has to maintain this in a year, not for the person reading your diff today. You give up a little ego; the team keeps moving.

Doing the work nobody claps for

A lot of what keeps a team healthy is invisible. Writing the tests. Fixing the flaky CI job. Paying down tech debt. Taking the boring integration ticket so someone else can do the exciting one. None of it earns applause, and some of it quietly makes you look slower than the person shipping shiny features on top of the foundation you’re repairing.

I’ve made my peace with that trade. I’ll take the reputational hit of “what has he even been doing this week” if the alternative is a team that’s blocked. Saying the uncomfortable true thing in a retro, or owning a mistake to unblock everyone, costs you a little standing in the moment. It’s worth it, because the organization is better for it - and the organization is a lot of people.

Sharing what you know, even when it’s “not cool”

The third one is my favorite, because it compounds. Open-sourcing a small tool. Writing docs nobody asked for. Answering the same beginner question patiently for the tenth time. Mentoring a junior. Giving the internal tech talk. Writing a blog post like this one.

None of it is glamorous, and some of it gets the “why bother, that’s basic” treatment. But knowledge shared is one of the highest-leverage forms of social good an engineer has: it lifts many people at once, and it keeps lifting them after you’ve moved on. The mild embarrassment of putting something imperfect out in public is a tiny price for that.

Why I choose it

You might have noticed the through-line with things I’ve written before - disagree and commit, building rooms where people feel safe to be honest, keeping humans in the loop. It’s the same instinct. When I walk into a meeting, a review, or a repo, the honest shared goal is that the product and the team get better. Social good is just choosing that over the smaller, more personal win - again and again, in small doses.

And there’s a version of this for what we build, too: an AI feature that genuinely helps a user (accessibility, real time saved) is worth more than one that inflates a vanity metric, and a model that does the job on a fraction of the compute is a quiet gift to everyone downstream, including the planet.

None of these choices are heroic. That’s the point. Social good isn’t one grand project - it’s the sum of a hundred small decisions to spend a little of yourself so that more people are better off.

Further reading

• Google, “AI for Social Good”: blog.google/innovation-and-ai/products/ai-social-good
• United Nations, “The 17 Sustainable Development Goals”: sdgs.un.org/goals