I was sitting in the parking lot of my daughter’s dance studio when my wife texted: “You’re on afterschool signup duty. Link’s in the school email. Limited spots. Go.”
I tapped the SignUpGenius link. The page loaded slowly, irrelevant ads taking up half the screen. I refreshed repeatedly, fighting through a wall of clutter to catch the moment the limited spots went live.
When they finally appeared, the real fight began. I was forced to create and verify an account, a ninety-second detour that dumped me onto a generic dashboard instead of the signup form. Once I navigated back, I tapped what I thought was right and hit submit.
The slow horror followed: every slot was already full. All of them. While I was battling verification emails and confusing layouts, other parents had gotten through.
When my daughter came out, I had to tell her she didn’t get into the program. She cried on the way home. She didn’t fully understand why she couldn’t go. I did, and that made it worse.
The Problem That Stuck
I work in solutions engineering — I’m paid to spot broken technology. Most software frustrations are mild. You adapt. You forget. This one stuck, because I kept hitting SignUpGenius links for every parent duty: school fundraisers, soccer snacks, teacher conferences. Every time, the same hostile experience. The group texts always lit up: “Which slot is which?” “How do I change it?”
SignUpGenius has millions of users but hasn’t meaningfully improved in years. Their business model is advertising, which makes the incentive structure exactly backwards: more friction means more page views means more ad impressions. Your frustration is a feature, not a bug.
I wrote down three questions: What if signing up took fifteen seconds? What if the page had zero ads and loaded instantly? What if you didn’t need an account to claim a spot?
That became a spec. The spec became a repository called SignUppity.
Why I’m Terrified of the Tools I’m Using to Build It
I lead an engineering practice at my day job. I’ve watched teams adopt Claude, Copilot, Cursor, Codex. I’ve written about AI-assisted development professionally. I have opinions. And I am anxious about it — not the sky-is-falling variety that makes for good LinkedIn engagement. Something quieter.
I’m anxious because these tools are genuinely good at something I’ve spent years learning how to do. I can scaffold a full-stack application with Claude Code in an afternoon — auth, database, API routes, frontend. Work that used to take a week or two. The first time I did it, my reaction wasn’t excitement. It was a low-grade existential vertigo. If the machine can do the thing I’m good at, what am I actually good at?
They’ve started calling this cognitive surrender — the moment you stop pretending you’ll hand-write that database migration and let the AI draft it. It’s not giving up. It’s more like accepting a promotion you didn’t apply for. You used to write the code. Now you shape the code, decide what code should exist. Your value moved up the stack, whether you wanted it to or not.
The only way through the anxiety is to build something real. Not theoretically. Not in a sandbox. Something where the AI tools are load-bearing, not decorative. SignUppity is that something.
Being Junior Again (On Purpose)
I’ve spent most of my career in the enterprise world — big systems, big teams, careful consensus-driven building. It’s the right approach for that context. But it has a side effect: you forget what it feels like to not know things.
AI tools have made me junior again, in the liberating sense. I’m building with a stack I chose because it’s modern and fast, not because a committee approved it. Astro, Supabase, Cloudflare, TypeScript. I know some of these tools well and some barely at all. The AI fills the gaps — and teaches me things I wouldn’t have learned on my own, because I wouldn’t have known to ask.
The enterprise world treats AI adoption like procurement: buy seats, distribute licenses, measure usage. The indie world treats it like craft: build with the tools, develop intuition through repetition, share what works. I’m betting the indie approach teaches us more. And I’m going to find out publicly.
The Plan
Vague build-in-public projects are the content equivalent of “we should get coffee sometime.” Here’s the actual plan:
Create a public place. You’re reading it. sahd.dev documents every decision, mistake, and interesting line of code.
Build something real. SignUppity — ad-free, mobile-first sign-up platform. Time slots, event registration, magic-link auth. MVP ships in two weeks. Not “starts development.” Ships.
Learn as I go. Claude Code writes the first draft. I review, shape, and ship. When it breaks, I’ll say so. When I have no idea what I’m doing, I’ll be honest about it.
Share all the results. Revenue, costs, user counts, server bills. No rounding up. If the project fails, that’s a story too.
Fling an AI-powered rock at Goliath. SignUpGenius has millions of users and years of momentum. I have a laptop, a Claude subscription, and a very specific grudge. The odds are absurd. That’s fine.
Do it while the kid is asleep. This is a nights-and-weekends project. The constraint is the point — if SignUppity can only be built by someone with unlimited time and a runway, the tools aren’t as good as advertised. I’m betting they are.
What’s Next
Tomorrow I start the landing page at signuppity.com — Astro, TailwindCSS, Cloudflare Pages, a waitlist backed by Supabase. Then the hard part: time-slot signups with magic-link auth. Then event registration.
Two weeks from now, either SignUppity exists and people can use it, or I’ve learned something important about why it doesn’t. Both outcomes get documented.
If you’re a parent currently fighting a SignUpGenius link on your phone in a school parking lot: I see you. Help is coming.
