Yes, I still run a CS 1.6 server. I have a Discord alert that pings me the second someone connects so I can drop what I’m doing and go play. I’m 41 years old.

This isn’t nostalgia goggles. Or maybe it is, but there’s something real underneath it.

1998-2008#

There was a window where the best PC games came from people who just loved games. Counter-Strike was a Half-Life mod made by two guys. Quake III shipped and then id Software released the engine source code under the GPL so the community could do whatever they wanted with it.

The business model was simple: make a good game, sell it for $50, ship an expansion pack if you had more ideas. No battle passes. No seasonal FOMO. No $18 weapon skins. No dark patterns designed by psychologists to keep 14-year-olds spending their parents’ money.

You bought the game. You played the game. If it was good, you played it for years. CS 1.6 came out in 2003 and people are still playing it. Not because Valve is running live events and drip-feeding content. Because the game is good.

LAN Parties#

CRT monitors, cases with the side panels off because the cooling was bad, a folding table that was definitely not rated for the weight of 6 PCs, a flat of Bawls guarana from the gas station, and two large pizzas from whatever place was still open at midnight. Playing games with your friends in the same room, yelling at each other over the sound of AWP shots and buy menu clicks.

UT2K4 was the LAN party king. Onslaught with vehicles, Bombing Run where you’re basically playing football with rocket launchers, Assault maps that played like mini campaigns. You’d haul your tower across town, spend 20 minutes running cables and arguing about who gets the good chair, and then play until 4am. The game had a server browser. You picked a server. You played. No matchmaking algorithm deciding your fun for you.

Those nights happened in someone’s basement on a Thursday, and they were better than anything a $200 million live service game has produced since.

Game Servers and the Old Internet#

The early internet was weird and personal and ugly and alive. GeoCities pages with tiled backgrounds and MIDI music. Forum communities where you recognized usernames. IRC channels where the regulars would idle all day and conversations would pick up and drop off naturally.

Game servers had the same energy. You’d find a CS server with good ping, play on it for a few weeks, get to know the regulars, and it became your spot. The admin was just some guy who paid for hosting because he liked the game. No corporate entity behind it. Just a dude with a server and a community that formed around it.

That’s mostly gone now. Discord replaced IRC, Reddit killed forums and then went public and sold everyone’s posts to AI companies, Twitter became whatever Twitter is now. The internet got bigger and more capable and turned into something you use instead of a place you visit.

The Kaplan Interview#

Jeff Kaplan on Lex Fridman’s podcast talked about how the games industry went from people who wanted to make cool things to people who wanted to extract money from the people playing them. The shift from “how do we make this fun” to “how do we maximize revenue per user.”

Live service games aren’t designed to be finished. They’re designed to never end, because ending means the revenue stops. Every reward loop, every daily login bonus exists to keep you coming back. Not because the game is compelling. Because stopping feels like losing something.

He talked about early Blizzard: arguing about whether a quest was fun enough, whether a zone felt right to explore, whether a dungeon’s pacing made you want to run it again with friends. WoW launched at $50 with a $15/month sub. That price hasn’t changed in 20 years. You paid for the game, you paid to keep the servers running, and in return they built new continents and raids every couple of years. The deal was obvious and honest. No storefront shoved in your face every login, no limited-time offers manufacturing urgency.

Now the conversations are about conversion funnels, whale spending patterns, and how to structure a battle pass so free players feel just enough pain to consider paying. The talent is still there in individual developers. But when your game needs to generate $500 million a year in microtransactions, the people making decisions are the ones who love spreadsheets, not the ones who love games.

Indies Keep the Spirit Alive#

The tools got cheap. A small team can build something real now without a publisher. Valheim, Lethal Company, Balatro, Stardew Valley. Tiny teams, sold for $10-30, no live service, no battle pass. Good games that respect your time. Same energy as the early days, just with better tools.

On the retro side, communities like Fightcade keep classic arcade fighters alive with rollback netcode. Third Strike, Marvel vs. Capcom 2, Garou. You hop on, see who’s in the lobby, challenge someone, play sets. The games are 20+ years old and the people playing them are there because the games are mechanically excellent. It feels like the old internet because it basically is.

Retro Dad Night#

Part of why I host game servers is selfish. I want to play CS 1.6 with friends when the mood strikes. I want retro dad night to be a thing where we can hop on UT2K4 or CS after the kids are in bed without dealing with some random public server’s rules or ads or pay-to-win garbage.

But the servers are also public. Anyone can join the CS 1.6 server. The Minecraft servers are open for friends and family. I’d rather run the infrastructure myself than rely on someone else’s service that might disappear, start charging, or inject ads. It runs on hardware I already own, on power I’m already paying for. The cost is basically nothing.

The full list of what’s publicly available is on the game servers page.

CS 1.6#

22-year-old game. No progression system, no unlocks, no cosmetics, no ranking anxiety. No reason to play it other than the fact that it’s fun. The skill ceiling is absurdly high. The maps are burned into my brain from thousands of hours. The movement has a rhythm to it that modern shooters don’t replicate because they’re too busy adding slide mechanics and operator abilities.

I built monitoring infrastructure that alerts me on Discord when someone connects to a game server that came out when I was in high school. Full observability stack, Grafana querying Loki logs for the string “entered the game” and firing a webhook. Enterprise-grade monitoring for a 22-year-old video game.

I also host Minecraft servers because my kids and their friends play on them. Different kind of hosting, same motivation. I’d rather run the server myself than have them on some random Realms instance or a pay-to-win public server.

I can’t recreate the LAN parties. But I can keep the servers running, and when someone joins and I get that Discord ping, for a few rounds it gets close.