A minute of company history

Survive the AI eras. First chatbots. Then agents. Then whatever the agents hire.

What you do all day

The office empty during lockdown, with everyone on a video call

People, not headcount

Everyone has a desk, a mood and a breaking point. Automate the work they were proud of and they coast, burn out, then leave.

Yak in #random: an office-on-fire image captioned SEV-1. Everything is fine, with the replies We are mid-outage and Love this, will laugh once prod is back.

Talk back in Yak

Answer your team's threads, or post a pep talk, a meme or "who broke prod?" yourself. They react, they reply, and they argue back. A meme during an outage gets the reaction it deserves.

The Squads tab in Staff: Core posted to a product, Growth on an update, and faces marked on loan

Put people in squads

Name a team, give it a lead and post it to a project. The longer a squad works together, the faster it ships. When the product launches, the people who know it best stay on to keep it running.

Launch day reviews from the parody tech press

Ship it, then read the reviews

Pick a category and an angle, build, launch, and let TechCrunchy decide how you feel about it. Some combos are brilliant. Some are AI-powered spreadsheet number 47.

Call the outage a stress test

Rogue agents, breaches and outages at 3am. Red lights, people running, and a call to make before the customers notice.

Decisions you can see

Big calls show up in the office: a novelty cheque the size of a door, a whiteboard that says the market has spoken, the founders hiding while a stranger tries the product.

Automate it, and live with it

Agents are cheaper than people, right up until the cloud bill develops feelings. Hand the work to the machines and watch the rack run hot, the pull requests pile up overnight, tech debt piles up wherever reviews slip, and the people who used to do it find something else to care about.

The Waffle Party is real

Reward your team with waffles. And a mandatory dance break. Nobody asked for the dance break.

From the office

One career, eight eras

An era start picks the year your company opens. Choose Classic and you open in 2019, as always. Choose another and the office, the staff and the street outside are dressed for that decade, the chat app and the market change with it, and the game plays on through everything that follows. Start with boxed software in a garage, ride the dot-com boom and its bust, or go straight to the chatbots, the agents, the consolidation or the plateau. Earlier starts score a little less than the full Classic run, so the long career stays the prestige one.

Two founders in the Garage, 35 people at HQ, then fewer as the agents take over.
A garage office with rows of beige CRT computers and dot-matrix printers, a corkboard, a whiteboard and a green wall display
Pre-internet, 1990. Boxed software, mailed patches and a LAN noticeboard for chat.
A dot-com garage office with a chart and a digital clock on the wall, and staff in cream shirts working at CRT monitors
Dot-com. A BOOM/BUST chart on the wall, an IPO offer on the table, and a bust to survive.
A digital wall clock reading 23:59 marked Y2K, beside a falling BUST chart, above a row of CRT monitors and dot-matrix printers
Y2K. The millennium clock counts down on the wall as the year runs out.
Web 2.0 staff at modern desks beside a wall of green dashboards, with tall city buildings outside the windows
Web 2.0. Old browsers to support, an ops wall of dashboards, and a city growing up around the office.
Three staff in period clothes: one in a yellow Y2K graphic shirt, two in cream shirts, with oversized glasses
Dressed for the decade. Every era has its own wardrobe, from tucked shirts and oversized glasses to fleece vests, graphic tees and hoodies.

Then the modern eras

2019

Classic SaaS

Web apps, mobile apps, APIs, and a lot of pricing pages. Nobody has heard of a large language model.

about 2022

The ChatGBT Moment

A chatbot writes a sonnet about your churn rate and the whole industry loses its mind.

about 2025

Agents

The software now does the job instead of helping with it. Someone has to watch it do the job.

about 2029

Consolidation

Everyone is buying everyone. The vendors change their prices, their models, and their minds.

about 2034

The Plateau

Everyone has the same models now. What is left is people, taste, and whether anyone trusts you.

For anyone who's worked in an office

The press has opinions

TechCrunchy

"Twenty years of the software industry in one evening. We recognised every meeting. We had been in most of them."

The Vergence

"Automate the support team and the numbers go up. So does the number of people quietly updating their LinkedOut."

Wired-ish

"The first management sim where the most dangerous thing in the office is the shared printer."

Hacker Olds

"Built mostly by AI agents. The commit log is scarier than the ransomware event."

Human in the Loop. Someone has to be.

Play free in your browser