Matthew Stanciu The first major project I ever worked on, and what first got me into coding, was a Minecraft server called Extrillius, which I built in middle school. I wanted to create a minigame server whose player experience was endlessly customizable and games were built by & for the community. I sketched out an idea that allowed players to build a parkour map in the Creative server and “export” it to the Parkour server. One of my ideas—“parties” where small groups of people can filter players and chat to them & play games together—was later implemented by Hypixel. My first GitHub repository is the first plugin I wrote for Extrillius: a swear filter toggleable per person.
What I really wanted to build was a community where people made friends, built things together, and felt a shared sense of ownership; Minecraft was just the platform on which I was building it.
But Minecraft servers are boring: probably 90% of all Minecraft servers in history were started by a middle schooler who taped together popular plugins and paid $50 to advertise on Minecraft Server List before shutting it down after a month. Starting a Minecraft server is easy, and Minecraft is a game, so there’s a low ceiling to how impressive it is to start one.
All the Minecraft servers I played on in middle school—even the more legitimate ones, like Hypixel—felt rigid and untrusting: the game modes, maps, chat structure, rules, and ranks were all created by the server owners, and players had no control over them. Moreover, the culture was such that moderators and admins were gods who enforced rules as they wished, didn’t appreciate feedback, and viewed players as malicious by default. I found this all quite strange for a game whose core tenets are exploration, creativity, and collaboration. Even the APIs that allowed their and every other popular server to exist—Bukkit and Spigot—are open-source projects created by & for the community.
I wanted Extrillius to take the form of a Minecraft server despite their perceived shortcomings, though, because I felt like there were hidden gems in the format that could make them uniquely amazing. Why couldn’t a Minecraft server allow players to create their own game maps, or control what they see in their chat, or control who they see & play with, or add functionality to the game? Why couldn’t a Minecraft server be a medium-trust environment instead of a low-trust one? Why couldn’t a Minecraft server be a community?
On September 16th, 2022, Purdue Hackers began an experiment called Hack Night: every Friday night, a group of people gather and make projects together. Hack Night made two bets:
- People don’t make things because it’s hard and their classes drain their energy—but if they were given a period of uninterrupted time to make things with other people, they would make unimaginably creative things.
- Many people don’t actually have something to do on Friday night and are looking for a group of friendly and interesting people to hang out with every week.
These were not obvious bets. Some pushed back, fearing nobody would come to an event on Friday night. But at the third event, something strange happened: a few hours in, a group of people spontaneously gathered together and created a new way to measure time.
It’s called Lightning Time, and its creation was incredibly eye-opening for me. For one, it validated Hack Night’s first bet—something I thought would take much longer. But even more surprising was the month after Lightning Time was created: we created a Raycast extension, an npm package which we later used on the Hack Night dashboard, and an explainer with custom, interactive React components—the post for which this blog was initially created.
We created a new thing for people to use, built tooling, wrote documentation, marketed it, and shipped it. We basically shipped a product. Sure, the product is an esoteric art project that’s useful to almost no one, and we have $0 in MRR—but we shipped a product! It was exhilarating.
Barely a month later, we did it again when we shipped a new landing page—a marvel of creativity and attention to detail that continues to make me smile every time I visit it five months later.
We shipped an online Macroquad code editor. We shipped a website for viewing upcoming & past events. We shipped a public API. We shipped an Edge Config-based link shortener with an optimistic UI dashboard. We’ve been shipping! And all of it is real software intended to be used by real people.
One lens through which you can view this is that Purdue Hackers operates somewhat like a startup: we run three core programs, and we build software to support these programs, which serve real users. But unlike a startup, we’re not trying to make money. All our software can suck: we can have show-stopping errors, downtime, terrible performance. We can build and rebuild all our software, entirely open-source. We can experiment with new & unstable technologies. Churn doesn’t matter because thousands of new potential “customers” come in every Fall. Most of our operating costs come in the form of food for events and don’t total more than a couple hundred dollars per month. Companies like Vercel are happy to sponsor our hosting costs. And other companies are knocking on our door to sponsor us, because all this creates people with real engineering skills and product sense who they want to get in front of.
Many of our new software projects have been spearheaded by organizers, but Hack Night has become a model for what open-source collaboration looks like. Today, 30 people gather at Purdue’s makerspace every week to make projects together—including generative art, an RC car, a website that lets you draw by holding your finger up to your webcam, and a NeoPixel Lightning Time clock.
(Update: as of December 2023, Hack Night now has an average weekly attendance of 80.)
This is the hidden gem of the college club format. CS (and adjacent) college clubs can operate similarly to startups, shipping real software for real users in a high-energy environment—but they don’t have to face any of the real-world constraints that startups face, such as significant operating costs, delivering business value, being cash flow-positive, and ensuring their products are reliable and bug-free. Instead, they can channel the high-energy environment of startups toward creating & contributing to open-source projects, making art, and learning new technologies.
Unbounded by real-world constraints and energized by those around them, college club members can quickly build portfolios of impressive projects. In addition to being personally creatively fulfilling, these portfolios also build marketable skills and increase their self-confidence, which also make them more employable.
This arrangement is uniquely enabled by the environment of a university. I struggle to think of another space in the world where something like what I’m describing is possible. The closest thing to it that I can think of is the Recurse Center—but the Recurse Center’s retreat only lasts 6-12 weeks, whereas a college club lasts up to 4 years; and attending the Recurse Center comes at a cost to attendees in the form of taking time off work and living in New York City, whereas college clubs simply support the goals that students are already pursuing by being in college.
Purdue Hackers has become a standout success case for this model over the 2022-2023 school year. One year ago, we had no community and no culture of shipping; today, in addition to everything I mentioned earlier, community members are working on a Purdue Hackers flag, a two-foot-tall LED-lit Conway glider powered by a custom PCB, an experiment in program synthesis, Conway’s Game of Life over SSH, an electronic drum kit, and many more incredible projects.
Every workshop we’ve run this year focused on shipping a creative project, including a gravity simulator, a shuriken-catching game in Rust, a Discord bot, and sound-based art. By the end of all these workshops, everyone walked out with a finished project.
A strong community identity has also formed. Projects like the LED Conway glider are built by & for community members; the SSH Game of Life app was initially built for a workshop run by a community member; multiple community members have told me that Purdue Hackers has become their primary social group.
All this would have been unfathomable a year ago, before we unlocked the hidden gem of college clubs.
I get the impression that nearly every CS college club is guided primarily by helping its members get jobs. For example, they recognize that employers want to hire people with marketable skills, so they run workshops on “MERN Stack” and “iOS & Android Development”. Or, their members feel their job applications would look better with extracurriculars, so they retain membership by providing dozens of leadership positions that look good on a resumé. I understand why they do this—the uncertainty of getting your first job in tech feels daunting, and it’s a tough hiring market right now—but when I see this, I’m reminded of the Minecraft servers that inspired me to start Extrillius in middle school.
I found those servers so frustrating because I felt like they could all be way better if they simply leaned into the collaborative spirit of Minecraft—but instead of taking the low-hanging fruit, each new server looked to the last as a benchmark for success, copying their bad decisions without thinking about why they were doing it in an infinite Telephone-like race to the bottom.
Similarly, I think optimizing directly for employment is the wrong approach for college clubs. I don’t deny that those that do it are successful in helping their members land jobs—but as a result of pursuing employment directly, they lose the soul, community, and radical personal growth that they could be uniquely capturing via the college club format.
There’s a better way, and Purdue Hackers is what it looks like. We’ve created a vibrant community where people are making friends, building creative projects together, and growing radically as programmers and people. And we’re not sacrificing employability; in fact, we’re more employable because of it.
Something this year has made clear to me is that college students are yearning for spaces where they can make things with friends. Nobody asked for Hack Night, yet as I’m writing this, some community members who are staying on campus this summer are planning unofficial summer Hack Nights. If there’s this much energy at Purdue, I’m sure there there’s as much or even more potential energy at every other major university—but today, most of that energy remains untapped.
I’m inspired to build Purdue Hackers because I know I’ll never be able to do anything else like it in my life. I have a short three-year window to create something uniquely special; to settle for anything less than that would be deeply disappointing for everyone. As we approach the long summer break, I encourage other club leaders to think about why you’re running your club and what you can do to make it uniquely special.