How This Whole Thing Works

A layman's explanation. Mostly accurate. No robots were harmed, though several were mildly scolded.

The headline river

Every few hours, a small script that has never once been outside opens up eight or so Commanders websites, reads every headline, and asks itself one question: "does this mention the Commanders." If yes, it keeps it. If it's actually about the Lions and the Commanders got mentioned in passing, it throws it out, because this script has standards. It is not a journalist. It has never broken a story in its life. It just really, really likes making lists.

The ticker

That scrolling strip near the top is basically this site eavesdropping on beat reporters' social media, through a side door, because the front door started charging rent a while back. It refreshes every couple hours normally, and every 15 minutes during a game, so you can watch the press box lose its mind in close to real time without opening a single app.

The Blog (weekly editions)

Once a week, an AI model that lives on this computer (not the cloud, not anyone else's business) reads everything that happened and writes a proper column about it, complete with quotes and, if it's feeling spicy, a pun. Here's the important part: it doesn't get to publish anything itself. A human reads the draft first and has to physically click "approve." This rule exists because an early test model once confidently reported that a player was on this team who has, in fact, never once put on a Commanders helmet. Robots lie sometimes. Not on purpose. They just really want the sentence to sound good.

The Blog (game days, new and slightly reckless)

During an actual game, there's no time for a human to review anything between the 2nd and 3rd quarter, so this part runs completely unsupervised. Every 15 minutes, a different (faster, cloud-based) AI checks the score, and if a quarter just ended, it writes a quick recap using the real play-by-play plus whatever beat reporters and fans are posting live. It publishes itself. Nobody reads it first. The one leash it's on: it is not allowed to decide on its own that a catch was "amazing." It can only say that if an actual person online already said it first, with their name attached. Left to its own devices, it would apparently just say "outstanding" about everything, so this rule keeps it honest.

The betting line

The point spread and over/under next to the schedule come straight from a real sportsbook, by way of ESPN's own public data, the same feed that powers the little widget on espn.com. Nothing here is a tip, a lock, or financial advice from a website that also aggregates blog posts about a rookie kicker.

Podcasts, schedule, fantasy, etc.

The schedule is scraped from the team's own site. The podcasts are just real Spotify players, embedded, doing what Spotify players do. There is no fourth thing hiding here that secretly runs on a hamster wheel. Probably.

The fine print, but funnier

This is a hobby project built by one person and several well-behaved scripts, not a newsroom. If something here is wrong, it's almost certainly the robots' fault, and they have already been informed of this in writing.