Engineers should do pro-bono work.

Not all the time. Not for every cousin with a startup idea. Not for the guy at the party who wants you to build "Uber, but for sandwiches" in exchange for equity and warm beer.

But sometimes.

Lawyers have a whole concept for this already. If you're good at a thing, and that thing is expensive, and some people who need it cannot afford it, you occasionally give some of it away.

Software should have that too.

AI changes the math

A small useful software project used to be a big ask.

A website, a booking form, a little CMS, a small admin tool, a decent landing page with analytics and deploy previews and a database and a way for a normal human to edit the words without opening VS Code? That could easily become weeks of work.

Now? If you know what you're doing, a lot of that is weekend-sized.

Not because AI magically makes engineering free. It doesn't. You still need taste, judgment, deployment sense, security sense, and the ability to tell when the robot is confidently building a haunted little trash cathedral.

But agentic coding lowers the cost of generosity.

That's a big deal.

Small software matters

Recently, I built a website for a friend starting a company.

Nothing world-changing. No pitch deck. No "we're disrupting the future of whatever." Just a person trying to make a thing real.

The stack was boring in the best way:

• Next.js
• Vercel
• shadcn
• Turso
• GitHub
• CI with autodeploy when a PR merges

I tried a few CMS options and none of them felt right, so I wrote a tiny custom one.

Not because everyone should write their own CMS. Please do not turn that sentence into a religion. Most custom CMSes are cursed little filing cabinets with a login screen.

But this one had one user. One job. One editing flow. It could be shaped exactly around the person using it.

That's care.

Sometimes software is better when it is not generic. Sometimes the whole point is that it fits one person like a jacket they already love.

This is not startup fuel

There is a kind of engineer who sees every small need as a potential SaaS.

A local dance studio needs scheduling? SaaS.

A friend needs a website? SaaS.

Your neighborhood chess club needs member signups? Believe it or not, SaaS.

And sure, sometimes that's the right instinct. But god, what a boring way to look at the world.

Not every useful thing needs to scale. Not every kind act needs a pricing page. Not every form with a database behind it needs a logo, a waitlist, and three founders in Patagonia vests saying "community-led growth" at each other.

Sometimes somebody just needs a good thing, and you can make it.

Pick carefully

The boundary matters.

Do not become free IT support for everyone you've ever met. Do not inherit maintenance for a flaky nonprofit calendar until the sun burns out. Do not say yes to vague projects with vague owners and infinite appetite.

Pick small things.

Pick people who will actually use them.

Pick work where a modest amount of software can remove a real amount of pain.

Then make the thing clean enough that it will not become a raccoon in your walls six months later.

The payoff

The funny thing about this kind of work is that people remember it.

A good little website can make a new company feel real. A booking form can save someone ten annoying messages a week. A tiny CMS can let someone change their own words without texting you at midnight like "hey can you make the button say Book Now instead of Contact Us."

That's not glamorous engineering.

It's better than glamorous. It's useful.

And if agentic coding means strong engineers can make more useful little things for the people around them, we should do that.

Not because we owe the world free labor.

Because building software is a weird, expensive superpower, and sometimes the correct thing to do with a superpower is help somebody move a couch.