Building the BETA invite and losing a round to Claude

On Monday, I said I wanted to write more and build less. That's still very true. I got completely in the zone and wrote several blog posts for TWO in one sitting, which put me ahead of schedule again.

That allowed me to build the BETA invite page, which was also planned for this week. My goal was to have at least ten posts published before I turned on any marketing for TWO.

Yes, I know that's an unusually specific goal to hit before doing anything else. But I want the site to feel lived-in before I send anyone to it.

I'll still run one more QA pass on the app and the website before that happens, because I'm the type who checks twice, but one thing was still open: a BETA invite page with a waitlist, so I could start collecting testers instead of just thinking about it.

The page itself

The page I built is deliberately plain. No wall of information, no feature dump, just an honest invitation to join the BETA and get lifetime access if you do. That's the pitch. I didn't want to oversell an editor that's still finding its shape.

The Resend fight

Getting Resend and Claude to cooperate was...sigh, a real joke. Resend wasn't the issue. Claude was, again, altering inputs it shouldn't have touched and missing a requirement I'd stated clearly: every signup needs a mandatory confirmation email before anyone counts as subscribed.

Skipping that step is easier to build, but it leaves you with a messy list, and a messy list isn't worth the shortcut.

What made it worse is that I'd already fought this exact battle once before, when I set up Resend for my personal newsletter on this blog.

I even handed over that repo so Claude could follow the same working model instead of reinventing it. It STILL got the API and segment ID misaligned. So I went back and forth, raised my voice at a chat window more than once, and eventually got it working.

It's a small thing, but it's the kind of small thing that eats an entire afternoon you didn't plan to lose.

In the end, victory came. The waitlist works, and the BETA invite landing page is done.

Who gets invited?

Now's the harder part: finding real users willing to test-drive TWO before I go back to fighting with the Mac app, which is still a nightmare of its own.

Read also: How do you vibe code a Mac app? Fuck I don’t know.

I come from hardcore B2B SaaS, so I know how to sell into that world. I don't have the same instincts for a mix of B2C, B2B, and small teams, which is closer to what TWO needs.

I already know my ICP in theory. I've decided to start with bloggers and content creators, and for the early BETA specifically, I want to keep it narrow and focus almost entirely on that group.

Not because other groups don't matter, but because I can't test a clear signal if I'm talking to five different kinds of users at once. Bloggers are also where I have the most credibility to reach out directly, and where I understand the daily writing habits well enough to know if the editor is solving something.

There's a more personal reason too. I write primarily on an iPad, and a chunk of my early ICP thinking comes from that: writers who want a docs editor built around that device as a primary tool, not an afterthought.

I belong to that category myself, so it's the easiest place to start listening closely.

The upside of staying narrow is that I can always widen later. I'd rather pivot away from an ICP that isn't working than dilute the signal by testing too many segments at once.

Pace over acquisition

Financially, I can afford to wait years for paying users if that's what it takes. That doesn't help the $168M challenge in the short term, but it buys me room to test, pivot, and eventually land on the ICP that sticks with the editor long-term instead of trying it once and disappearing.

My plan for the first wave is simple:

Email a handful of older contacts and invite them directly. Find a few more through Reddit. Cycle through some relevant Discord servers. Nothing complicated, just direct outreach to people who are likely to use the goddamn thing.

Eventually I want a hundred BETA users with free access forever. But I also know that managing a hundred people at once isn't something I want right now, mostly because of response time, and because my mental bandwidth isn't what it was ten or twenty years ago.

I need room to build, write, and record without every hour going into support threads, and I don't want to lose focus now that I have momentum, and I am not in the mood to hit my personal workload limits now that I have gathered some discipline.

That only changes once TWO or one of my other projects is actually profitable enough to justify bringing someone else in to help.

Writing all this out, I realize I did know how to approach it. I just needed to think it through out loud instead of assuming I already had the answer.

User acquisition and distribution have never been my real problem. The real problem is pace, and right now I'd rather move slower and get the first users right than move fast and get the wrong ones.

Either way, the week started strong, just like that dumbest thing I built last week, which apparently is also still going strong.