The mug, the light, the tabs
The coffee steams in your chipped mug, the one you keep meaning to replace but never do. Outside, the light is still pale, the kind that makes the edges of the rooftops look softer than they are. You sit at the desk with three tabs open: a no‑code automation comparison, a WhatsApp integration guide, and a terminal game you found at 2 a.m. and forgot to close.
The plan
Your plan is simple on paper: build a tiny chatbot that can answer FAQs for a small community, wire it through WhatsApp, and log every conversation into a sheet without writing real code. The platform you’re testing promises drag‑and‑drop flows and prebuilt connectors; the documentation shows neat boxes and arrows, nothing like the tangle forming in your head. You tell yourself you’ll start with one trigger, one message, one reply. That’s how every big system begins: as something small enough to hold in your head all at once.
The game
Between sips, you switch to the game window. A blinking cursor waits on a black screen. You type look and the world replies in plain text: a corridor, a door, a faint hum behind the wall. It feels honest, this kind of interface—no animations, no onboarding, just cause and effect. There’s no algorithm guessing what you want, no feed trying to keep you scrolling. You ask, it answers. You move, the world changes. You make a mental note: your bot should feel like this. Clear prompts, immediate responses, no fluff. If someone writes “pricing”, they get pricing. If they write “help”, they get help. No cleverness, no detours.
The flow
Back in the automation tool, you drag a box onto the canvas: “When new WhatsApp message arrives.” Another box: “Parse text.” Another: “Send reply.” The arrows between them look tidy, but you already know where the mess will hide: in the edge cases, the typos, the messages that don’t fit any category, the moments when the flow you designed meets the chaos of real people. That’s fine. You can refine later. For now, you just need something that works end to end, even if it’s dumb.
Between two nodes, your mind drifts back to the game. What’s behind the door? A trap, a treasure, another corridor? It doesn’t matter. What matters is that the world is consistent. If you come back tomorrow, the door will still be there. If you typeopen door, something will happen, and it will make sense given what you’ve learned so far. That’s what you want your users to feel: not that they’re wrestling a bot, but that they’re navigating a system that behaves the way they expect.
The sketch
By the time the coffee is gone, you’ve sketched the first flow on paper: trigger on message, parse intent, reply from a template, write row to sheet. It’s not perfect, but it runs in your head cleanly enough. You save the doc, close two tabs, and leave the game open. There’s a door behind the wall, and you want to see what happens when you open it.
The light changes
Outside, the light has shifted. The rooftops are sharper now, the edges defined. Somewhere in Budapest, other people are waking up, opening their own tabs, building their own small systems. You take the mug to the sink, rinse it, and set it back on the desk, still chipped, still yours. Then you type the command and step through.
