Two months, three countries, seven projects

Two months in. Three countries. Seven projects in flight, plus a couple of collaborator slots that don’t quite count as projects but do count as work.

This is the inventory. The point of writing it down is partly to mark where the launch arc ends, and partly to be honest about which projects are shipped, which are mid-build, and which are still early enough that calling them projects is generous. I’ll name what can be named and anonymize what can’t. The work matters more than the labels, but the labels still have to be right.

The seven, in country order

ExplainTheHack is the first one and the only one I can name without checking with someone else first. It’s a small product for offensive-security practitioners who need to explain their work in their own words, for interviews and writeups. It’s the thing I built as a test in March, and it’s still the test, two months later.

Habemus is the second. A tournaments platform live in Spain, currently running real events with real participants, currently iterating on the parts that did not quite hold up at scale. Public-facing enough that the name appears here without an asterisk.

Flamenco is the third. A store for a client in the Netherlands, mid-build. Catalog, checkout, the boring parts done well, the marketing site that goes with it currently being wired up.

Fourth is an art gallery I’m building for my mom. Small site, large stakes, in the way that family projects are. It ships when it ships, and the deadline is mine, which means the deadline is the strictest one on the list.

Fifth is an IV-therapy and lab site for a client in Santo Domingo. Booking flow, treatment menus, lab integration. Live, iterating, the kind of project where the second release matters more than the first.

Sixth is debt-collection software for a law firm in Santo Domingo. Workflow engine, document handling, the parts that no one writes about on Twitter because they are not glamorous. The constraint here is regulatory, which is the constraint that quietly governs most software written for law firms.

Seventh is helping a client in Spain ship her first book. Production, layout, distribution. Not the kind of work I would have predicted I’d be doing in February. Currently the closest thing to “done” on the list.

On top of these, there are two collaborator slots. A book-reading app where the build is mostly someone else’s and the parts I touch are infrastructural. SMB automation tooling where the work is somewhere between consultative and hands-on. Both are real. Neither shows up as a flagship project.

What each project taught me

ExplainTheHack taught me that the right wedge product is the one that solves a problem you’ve personally felt at 11pm the night before a job interview.

Habemus taught me that a platform with real participants is a completely different animal from a platform with hypothetical ones. The bugs that matter are the ones the third user finds, not the ones the first user finds.

Flamenco taught me that “build the store” is shorthand for thirty smaller decisions about taxes, shipping, and refund flows, none of which look like building.

The gallery for my mom taught me that family-project deadlines are the strictest deadlines in the world, even when they are deadlines I set for myself.

The IV-therapy site in Santo Domingo taught me that a client who knows his own distribution mechanism end-to-end is the rarest client a small shop can have, and the most useful to learn from.

The law-firm debt-collection software taught me that regulated industries do not give you a learning curve so much as a wall, and the move is to climb the wall before writing the code.

The book project taught me that production has its own version of the engineering problem: the work is mostly the parts no one mentions, done in the right order.

The collaborator slots taught me that “I will help with the infra” is a job description that expands until you redraw it. Twice, so far.

What’s still mid-build

Half of the inventory is mid-build, honestly.

Habemus is live but iterating on scale issues that only show up at scale. Flamenco is not yet selling. The IV-therapy site is in its second release. The debt-collection software has shipped the core flow and not yet the reporting half. The gallery is two weeks from launch and has been two weeks from launch for two weeks. The book is one of the few that’s nearly done. ExplainTheHack is shipping new features weekly and will be shipping new features in six months.

The story is not that seven things got delivered in two months. The story is that seven things are in flight in two months, with one person, no offshore team, no PM overhead, and Claude in the loop on every workflow. Some of them will land in the next quarter. Some of them will reveal that they were the wrong scope and get cut. Most will be different shapes by year-end than they are right now.

The next six months

The next six months will show whether the bet holds.

I don’t have a finished sentence about this. Two months is a thin signal. It is enough to keep moving, and not enough to know whether the model holds up over a quarter or a year. The work has produced more work, which is the only signal that matters at this stage, but it has not yet produced enough work to know what the steady state looks like.

I’d rather be honest about that than pretend the picture is settled. The launch arc was four posts about how the shop got started, what it bets on, what it runs on, and what it has shipped so far. The next six months are about what happens after the launch arc ends.

Starting next Thursday, the posts change shape. Less narrative arc, more shipped-work. Each one will be anchored in a specific thing that happened that week, with the same vulnerability and the same specifics, just without the four-post structure carrying it. The arc was the introduction. The work is the rest of it.