You’re sitting in the large glass room at the Life Sciences Institute. Computer science majors sit in the thick sound of typing. The gentle murmurs of “oh shit” and “hell yes” accent the vacuous space. A man wearing a dinosaur costume walks up to you. He silently hands you a bright pink energy drink with enough caffeine to make you worried. He daps you up and moves on to the next person. It's 3 p.m. — 21 more hours until the end of the cmd-f hackathon.
Hackathons are as computer science as electric apple pie. Each person, in groups of up to four, attempts to write a functioning computer program in 24 hours. 24 straight hours. They try to solve some problem that fits with the judges criteria, and work from noon one day until noon the next.
I had some friends in computer science who participated last year — they each pulled all-nighters to make a program that barely functioned. They said it was difficult, miserable and harrowing. They also said it was one of the most fun weekends they’d ever had.
nwPlus specializes in these hackathons. Based at UBC, but not an AMS club, they run three premier events each year. HackCamp runs in mid-November; it’s beginners-only, a gateway into the hackathon zeal. Early January sees their largest event, nwHacks. Open to the general public and hackers of any skill level, nwHacks is the self-proclaimed largest hackathon in Western Canada, sporting 734 competitors in 2025. Lastly, at the beginning of March, they run cmd-f (pronounced “command f”): a hackathon for underrepresented genders in computer science.
Tracy La, the logistics director of cmd-f, is a third-year microbiology & immunology and computer science (CS) double major. Originally a pre-med student, she took CPSC110 (an introductory programming course in computer science) and fell in love with the art of coding. She’s an alumna of cmd-f and nearly remade the event from the ground up this year, removing many of its rigid categories. “I think [the change] encompasses our Alice in Wonderland theme this year,” she said. “It's tipsy topsy turvy. Do whatever you want and only let your creativity be your limitation.”
Hackathons are the perfect place to play around with computer skills. You can’t get a bad grade at a hackathon, and none of it will affect your transcript. Tracy said participants “have a safety net so … they know that they're not ruining their whole experience because they can't figure this one little thing out.”
But most hackathon participants will still use every fibre of their being to figure those little things out. Some people will sacrifice health, sleep and comfort for the chance to solve a problem they created in their heads that morning. As someone with as much computer knowledge as Fred Flintstone, I never understood how someone could be so devoted to colon syntax and infinite loops. The devotion of computer science students to their craft is comparable to obsessive painters. Artisans more than number crunchers.
Computer science, ironically, is becoming less human as time goes on. The New York Times recently released a harrowing article about top computer programmers wrestling with their use of AI. The article ends with an ominous message — “skills that seemed the most technical and forbidding can turn out to be the ones most easily automated. Social and imaginative ones come to the fore. We will produce fewer first drafts and do more judging, while perhaps feeling uneasy about how well we can still judge. Abstraction may be coming for us all.”
Is AI going to take the soul from coding? Is there a human element inherent to the for loops and problem solving that could be erased by our use of other code? Does that mean coding is art? I went to cmd-f to ask that question, see the artists try to speed paint and be in the room with possibly some of the last humans to do a job like this.
That Saturday, I met the teams I would keep track of throughout the event. Some of the executives from UBC Women in Computer Science (WiCS) were in attendance. Selin Uz (a sixth-year CS major and WiCS co-president), Katja Radovic-Jonsson (a third-year CS major and WiCS community events director) and Maryum Chaudhry (a third-year business and CS double major and WiCS graphic designer) had all run their own hackathons or gone to previous ones. They had an idea for their project by the time I asked them their names: pulling from psychological research, they wanted an AI to help users frame certain thoughts, texts or talks with significant others.
Liza Junaidi, Nicole Schroeder (both third-year CS students) and Mahtab Zehtab (a third-year CS and chemistry combined major) came as a group. I talked with them for a while before Britney (another third-year CS student) joined in. Britney said she didn’t have many female computer science friends, and asked if she could connect herself with the established team. They accepted her with open arms. Computer science events can be communal like that.
Their program was solving a very direct problem: library study room booking is a pain in the processor. You have to go to each individual website to see if something is free at your specified time. They wanted to create a place that would consolidate that.
Evan Sun, Paul Xu and Ryan Liu (all second-year CS students) wanted to work with a very large scope. Computer science students have a presenting problem — they have no idea how to do it in front of crowds or hiring managers. Their program would give an AI presentation coaching session based on a video you submit of yourself. It would pick up on where you talk too fast, use too many ‘ums’ and ‘ahs’ and tell you how you could improve your communication. They said their scale was “ambitious, for sure, but that's what's exciting.”
Lastly, Charlotte Jacques (a third-year CS student) and Manya Goel (a fourth-year CS student) had both hopped on the ferry from UVic for this event. The world is an onslaught of bad news, and sometimes people want a respite. Their idea was to have users select what type of good news they want to see — science, art, and so on — and Claude (Anthropic’s AI model) would try to find an article that matches their interests. The scope seemed manageable, and if things went south, they would have time to fix it up.
Each team used large language models (LLMs) in some capacity to create and manage their program. Even if they were wrangling the AI like the WiCS members, or using Claude for its ethical considerations like those from UVic, LLMs were ever-present here. There was an organizational push to use it: sponsors awarded prizes to the best use of their LLM. I’m someone who treats AI like a monster under his bed — don’t look at it, don’t touch it and it won’t affect you. cmd-f was my immersion therapy. Here, it wasn’t something driving creative bankruptcy; it was a tool. Tracy called it “a sidekick instead of the hero.”
As the afternoon waned, the scope of the projects became real. The WiCs team saw basically no issues. They were a well-oiled machine, sitting in a line, talking frequently and helping each other where they could. AI had helped make their original website design, and now they were humanizing the whole thing. I watched the experts trim the rosebush of their project. As an arts critic, technically, Chaudhry’s work was gorgeous. Simple, elegant, as clean as the code they were working on.
The two from Victoria wanted to show me their newly made UI. It had worked two seconds prior. They turned the screen and it was frost white. This error took them two hours to fix. They did it with a smile.
The three second-years working on the AI coaching program were in hell. Each was locked into their own headset. They would detach once an hour to freak out with each other. They tried to explain the AI to me, but it went way over my head. Obviously, something wasn’t working. They all knew that they were going to stay late. They were thinking in the ballpark of 4 a.m. They desperately tried to hold the reins of a project spiraling out of control.
Dinner hit, and not a soul moved. Each person in that room was in the pilot seat. I stood up, alone, and picked up some of the provided poke. Sponsors covered every meal given out in this event. From the outside, it seemed like they had an infinite amount of free food. Coders hesitantly meandered towards the grub, slinking back to their chairs for silent dinner and YouTube. It was like staying at a sleepover where no one knew each other — quiet eyes and tired typing.
The night started to rear its head. The sky was cloudy, but the stars poked through and bathed the tired souls around me. People began to hunker down. Happy chatter turned to quiet typing.
My UVic group went home around 10 p.m. to try to do more work. For the library people, bugs started to pop up that they couldn’t control. By 11, Zehtab and Britney left to work from home, leaving Schroeder and Junaidi desperately trying to fix a deteriorating front end. The three in hell had managed to train the AI. This was hour 11 or 12 of the hackathon, and they could finally start coding. The WiCS team had a finished product, adding little touches as bugs started to rear their heads. They went home at 11:30.
Nighttime at a hackathon is a surreal experience. A limbo space where anything said is both lost to the aether and cemented in time. People were both locked in and losing their goddamn minds.
The organizers put on karaoke at 10:30 p.m. I would say five or so hackers went to this, standing on the edges of the crowd while Schroeder and I scream-sang “Call Me Maybe” in a random LSI lecture hall.
Surrounded by late-night zeal, I had the first energy drink of my young life. Here’s my review: don’t drink the Alani Nu pink energy drinks unless you truly hate yourself. I could feel my heart in my teeth until I called it a night at 1 a.m.
Leaving a 24-hour event and then coming back is a weird thing. Sleep is like time travel — you skip ahead in these stories that you’re missing. The library group stayed until 3 a.m. working on a single bug. Uz dreamt about coding.
I came back at 10 a.m. Sun was surrounded by a nest of empty energy drinks, the wrappers of old snacks and congealed Big Way hot pot containers — the contents had solidified like butter. His team apparently coded until 6 a.m, until the other two went for a little bit of sleep. Sun had stayed the entire time — he never left, slept or stopped coding. He proudly showed me the code he had slaved away on, only for the website to crash. He was a shell of a person after that.
The deadline approached swiftly. The UVic students were happy with what they could do with limited AI experience. The group working on the library project were upset that they’d had to hardcode some of the features, but were just trying to get something done. The WiCS team had a fully finished, slick application that I think they could’ve turned in the night before. An hour before submission, I heard a “fuck yes!” and cheers erupt from the three second-years. They had a shot at getting a finished product.
Typing grew more furious, people solidified plans, all crescendoing in the closing of the submission window. The entire building let out a sigh of relief. The projects were baptized with a name and sent out for judging. The WiCS team made Rabbit Hole. The library team made Study Room Hunt. The three second-years, in the nick of time, handed in Echo, and the UVic team made “The Optimist”.
Echo won a pilot program award — if they want to still make the app that nearly killed them, they now have the funding and support to do so. Rabbit Hole ended up winning the best use of Gemini. With the wrangling of the LLM they did, I couldn’t imagine anyone else winning.
These people put their souls into this code. It was as much of a labour of love as any art piece I’ve seen. AI seems built to replace that. If coding is an art, should we leave it to the machines?
The UVic students talked about how code can be pretty. Pretty code, to them, is well documented, readable and human. “I don't think [AI] can write a good code or a pretty code.” Goel said. “[A] lot of code [can] be beautiful because it's so nice. It's so well documented. There's white space, and I don't feel like throwing up after looking at the file and everything is so well organized.”
To me, she said everything but saying there's a soul missing in AI code. A human element. Radovic-Jonsson told me “Computer science was literally designed by humans to be as intuitive as possible to other humans, and that's why I always say, like, oh, the courses are actually not that hard once you kind of get into the groove. So I would argue that computer science is the most human sciences.”
The three second-years gave me a very distinct answer. You never expect someone to start quoting an Instagram reel at you, and dread the day you have to write about it for a newspaper. One of them said that a “person who works with their hands is a labourer; a person who works with their hands and their mind is a craftsman; and a person who works with their hands, their mind and their heart is an artist. So I think for coding to be considered an art, it would just take us coders, as software engineers, to work with our hands, which we are doing, work with our mind, which we are doing, but also be able to work with our heart.” They all clapped at that. People turned and stared for a couple of seconds. Who says computer scientists can’t have nuanced artistic opinions?
The most insightful answer I got was from one of the peer mentors, Harjot Singh, a CS and interactive art and technology major at SFU. We talked deep into the night, during the quiet clacking of code being brought into the world. He said coding has always been an art form. He told me, “I think that art is very personal, and the word ‘art’, especially if you think of it in the visual sense, has to have some kind of meaning, whether that's meaning to the person who made it or the person that it's being made for. And I don't think LLM-generated code is based on feeling or emotion. It is simply based on what is the best way to do this task with this code language.” For all purposes, it lacks a soul. It’s tracing a line through the maze to get to the end faster, not looking at the beauty of understanding the riddle.
To work in computer science is to problem-solve with the soul. That’s the only reason people stay until ungodly hours, hunched over screens for ungodly amounts of time. It's where the human spirit meets the cold machine.
I’d like to think that in thirty years, computer scientists will still be able to code, because something will be lost if they’re abstracted.
Tomorrow, I think I’ll write some code. After I sleep some.
