For a few years, I was talking to myself on the internet.
Not poetically. Literally. I was on Twitter writing about tech, tutorials nobody asked for, reviews of software I'd been testing on my own time, and if you've spent any time on that platform you know the engagement is brutal. Mine was worse. It was silence. I was speaking into an empty room.
There was no strategy behind it. No job attached, no portfolio play, no recruiter I was performing for. I didn't even know Twitter was a networking tool beyond the haha's. I posted because I like writing and talking about things, and at the time I had no one to share my enthusiasm for programming with. I liked building things and breaking things and writing about what I learned, so I kept doing it into the quiet.
Then the quiet started answering. A reply here, a question there, someone asking how I'd done a thing. I answered every time. That's the whole trick, and I'll come back to it, but that's it. The empty room turned into a community, then into paid work, then into invitations to speak and travel and a network of people I'd never have met otherwise. In 2024 a package arrived from GitHub, sent because of my contributions. This year another one came from Flutter. These are real boxes with my name on them, mailed to my city because of things I started doing alone, for free, when no one was watching.
None of that was luck and none of it was raw talent. It was a choice, and it's the same choice available to you.
There are two kinds of people in technology
There are people who consume technology and people who build it.
Consuming isn't a sin. Everyone does it. You use the apps, take the courses, watch the tutorials at double speed. The problem is that consuming feels like progress when usually it isn't. You finish a course, collect the completion badge, and your brain files it next to actual accomplishment. It isn't the same thing. You moved content from one place to another and changed nothing in the world.
There's a line, and on the other side of it you make the thing instead of using it: you write the tutorial instead of reading it. You fix the broken app, or file the report that gets it fixed, instead of just being annoyed by it. You build the opportunity instead of waiting to be handed one.
Most people spend an entire career on the consuming side of that line. Not because they can't cross it. Because no one told them they were allowed to and no one showed them where the door was. So here's the door.
Nobody tells you the truth about open source
Open source has a branding problem. Say the words and people picture a genius alone in a dark room writing code that runs the world, and they decide, correctly, that this isn't them, and, incorrectly, that the door is therefore closed.
The truth is more boring and more useful: your first contribution doesn't have to be code, and most contributions aren't.
- Documentation. Bug reports. Translations. A tutorial. Someone answering an issue at midnight so a stranger can keep going.
Code gets the credit and is the smallest part of what keeps a project alive.
Let me be precise about a distinction people blur. Posting your work, writing in the open, sharing what you learn, this is not open source contribution in the traditional sense. A hot take on Twitter isn't a contribution to a project. A review on your own blog isn't either. But here's what I won't do: pretend it doesn't matter. It matters enormously. Building in the open is how anyone finds out what you can do. Somewhere out there is a person who can change your life, a collaborator, a client, a destiny helper, and they cannot help you if they've never seen your work. You post into the quiet so that one day the right person is watching. That's not noise. That's the on-ramp.
And it was mine. The posts and the testing gave me the confidence to walk into rooms I wasn't invited to, to look at software and have an opinion, to see a bug and change the comma that made the tool a little more usable. One day you file a real bug on a real project, and later a box shows up from GitHub, and no one ever announces the moment you crossed over.
So do both, and don't let one crowd out the other. Build in the open so people know you exist. Contribute to real projects so you go deep, not just wide. They're not rivals. They feed each other, and the people who go furthest are doing both at once.
If you want to make a real contribution this month, the path is unglamorous. Pick a project you actually use, not the most famous one, one whose rough edges you already know. Read its CONTRIBUTING file; the good projects tell you exactly how they want help. Find an issue labeled good first issue or documentation. Fix the smallest real thing, a broken link, a confusing paragraph you can rewrite because you just fought through it. Then write your pull request like a human talking to a tired human, because that's who reads it.
And stop waiting to be ready. Perfect is an illusion. The typo fix you're too embarrassed to submit is worth more than the flawless contribution you never make. Ship the small imperfect thing. Do it once and it stops being mysterious forever.
AI, and the question everyone gets wrong
I'm not going to pretend AI isn't the thing right now. It's the biggest new door technology has opened in a generation, and most of the world is standing on the consuming side of it: typing prompts, getting a clever answer back, feeling that same warm hit of fake progress. Fine. But some people are going to walk through and build with it, on it, and decide what it becomes.
There's a conversation that happens around this and it's aimed wrong. It goes: how do we get more women into tech? Into AI? It sounds supportive. Listen to what it assumes. It treats tech as a house that already belongs to someone else and frames the job as inviting women in. "Invited" is the tell. You get invited into places that aren't yours.
That's the wrong question. Women shouldn't be invited into tech like guests in someone else's house.
The right question is: what happens when more women build the technology that shapes the future?
Those aren't the same question dressed differently. The first assumes permission. The second assumes power. And the stakes aren't theoretical. The systems being built right now decide who gets seen, who gets heard, whose problems are worth solving and whose get ignored because no one in the room ever had them. Who's in the room isn't a diversity footnote. It's most of the outcome. So this was never about being let in. When you build, you don't ask for a seat at the table. You build the table and decide who else sits there.
How people actually build with AI
The phrase "build with AI" sounds like it needs a PhD. It mostly doesn't. It's a ladder, from reachable to hard:
- On top of a model. Wire an existing AI into an app through its API, which is just software talking to software. A study helper that sends a question to a model and shows the answer is building with AI. You didn't make the intelligence; you built the useful thing around it. Most people start here, and sooner than they think.
- Customizing a model. Feed it examples until it's good at one job, a local language, one company's documents. That's fine-tuning: shaping a model, not inventing one.
- The system around the model. Connect it to real data, give it tools, make it reliable and fast and safe. The model is the engine; you build the car. This is most of what "AI engineering" actually is.
- The model itself. Enormous datasets, training runs, serious compute. The hardest rung, and the freeing part is that almost no one needs to climb it.
There's also the quiet version, and it's the same on-ramp as open source: test AI tools and report where they break, write the docs, build datasets in languages the big models handle badly. That's building too. You don't invent a model. You build something small on top of one and climb.
Building is bigger than code
I used to think building meant only writing code. Everything else was the soft stuff around the edges. That belief kept me small for years, because by that definition almost nothing I did counted, so I discounted almost everything I did. It was wrong.
You can build people. Teach someone, answer their question with actual patience, and you've built a person who goes on to build a hundred things you'll never get credit for. That compounds harder than code does.
You can build community. My empty room became one without a single line of code. That's how Women in Tech Malawi started. Once I saw how much visibility a post could create, I wanted the world to know women in tech in Malawi exist, so I started using the hashtag #womenintechmalawi. It became a community, and now it's a registered open community supporting women in the space. A hashtag. That's where it began.
You can build opportunities. Every gig and talk and collaboration I've had traces back to something I put out first, not knowing what it would return.
And you can build the systems that let other people build, which is the highest form of it. Documentation is that. Tutorials are that. Open source is that. It's building the thing that makes the next person's build easier.
If you've read this far thinking you're not technical enough to build anything, drop it. If you can teach, write, organize, explain, or show up consistently, you're already building. You've just been refusing to call it that.
The 30-Day Builder Challenge
Inspiration is worthless by the time your coffee's cold, so here's a deadline instead.
In the next 30 days, do one of these. Not all five. One. Pick it, put a date on it, do it.
- Build something. A tiny tool, a script that saves you ten minutes a week, a one-page site. Small and finished beats big and imaginary.
- Contribute to something. One open source project you use. A typo fix, a real bug report, an answered issue.
- Teach someone. Take one thing you know and hand it to someone who doesn't.
- Document something. Write down how you figured out the thing you figured out, so the next person doesn't have to bleed for it.
- Start something. A blog, a channel, a project. Talk into the empty room and see who shows up. Someone will.
I started by talking to myself in an empty room, for the love of it, with no plan and no audience. It turned into a community, work, travel, and boxes from GitHub and Flutter with my name on them. Not because anyone chose me. Because I chose to build, in the open, before anyone was watching.
Same choice is sitting in front of you. The door's right there and you know what's behind it now.
Don't finish this inspired. Finish it building something.
