Add a draft for The Joy of Hackathons

Author: dguo

astro.config.mjs

@@ -22,6 +22,7 @@ export default defineConfig({
         mdx(),
         sitemap({
             filter: (page) =>
+                !page.includes("blog/the-joy-of-hackathons") &&
                 !page.includes("blog/my-indoor-air-quality-setup") &&
                 !page.includes("blog/the-black-hole-problem") &&
                 !page.includes("blog/the-power-of-the-link") &&

src/pages/blog/the-joy-of-hackathons.md

@@ -0,0 +1,227 @@
+---
+layout: ../../layouts/BlogPostLayout.astro
+categories:
+  - programming
+date: "2025-03-01"
+unlisted: true
+title: The Joy of Hackathons
+---
+
+I love [hackathons](https://en.wikipedia.org/wiki/Hackathon). They remind me of
+the pure joy that can come from programming, from making something
+[fun](https://www.jollycode.org/) or [interesting](https://neal.fun/) or useful
+just by typing the right incantation of code. Or these days, by giving the right
+prompt to a [LLM](https://en.wikipedia.org/wiki/Large_language_model).
+
+When programming for a job, it can be easy to lose some of that joy. Real world
+software engineering involves so much more than the core act of creating
+something that sort of works. But most hackathons are all about that spirit.
+
+## Freedom and Constraints
+
+The magic of hackathons comes from two seemingly contradictory factors: freedom and
+constraints.
+
+The freedom comes from being able to decide what you want to work on and how you
+want to do it. For once, you aren't limited to the projects that your company
+deems important. You don't have to particularly care about business priorities
+unless you want to. Even at a technological level, you are hopefully free to use
+things that aren't officially supported. You can experiment. You can make a
+proof of concept for something that you know is a good idea. You can challenge
+how things are typically done.
+
+But hackathons are also about constraints. The main one is time. Online
+hackathons can ran for a while, but most in-person ones limit you to a few days
+at most. This limits the scope of what you can do, which forces you to focus and
+act quickly.
+
+Hackathons can also have themes. And they might revolve around a few
+technologies or concepts. For example, the annual
+[Js13kGames](https://js13kgames.com/) competition is a challenge to create the
+best web game that fits in a 13 kilobyte ZIP, and each year has a different
+theme. This can help you come up with a good idea, since your brain doesn't have
+to start from an overwhelmingly infinite set of possibilities.
+
+Combining freedom with constraints in this way means that you have agency but
+are also forced by circumstance to actually create something.
+
+## Benefits
+
+Why participate in a hackathon? It can certainly be a lot of work. I've pulled
+almost all-nighters for some hackathons. And a cynical take is that company
+hackathons are just a way for companies to get employees to do a bunch of extra
+work.
+
+But hackathons are fun! I love working on something that I know can't drag on
+for months and can't get interrupted by shifting business priorities.
+
+Hackathons are also a great way to learn something new and to try out new
+technologies in a low-stakes environment. Sure, many hackathons have judging,
+awards, and prizes. But besides a bit of time, you don't really have anything to
+lose by participating.
+
+There's also always the chance that your hackathon proof of concept does become
+something bigger. Maybe it becomes a widely used tool, feature, or product.
+Maybe you can start a company with it. Or maybe you'll just get the satisfaction
+of making something cool.
+
+I was drawn to programming because of the short feedback loops of writing some
+code and quickly getting a result. Hackathons are a version of that. You come up
+with an idea, try to turn it into reality, and almost immediately see how it
+turns out and how others react to it.
+
+## Advice
+
+For those who are convinced that hackathons are worth trying, I have a few tips:
+
+If you don't know of any hackathons to participate in, try looking on [Devpost](https://devpost.com/) or
+[Major League Hacking](https://mlh.io/).
+
+Keep teams small. You want to bias towards speed, and working with too many
+people on a hackathon project introduces too much communication and coordination
+overhead.
+
+Forget many best practices. Your goal is to make something that sort of works.
+For once, you can be wholly justified in not writing any tests, not caring about
+documentation, writing ugly code, ignoring scalibility or security issues, etc.
+The only thing that usually matters is the end result and how well you can
+present it. This is also part of what makes hackathons fun! You have good reason
+to loosen your standards and create something as fast as you can.
+
+## My Experiences
+
+Here's a log of my own hackathon experiences.
+
+### HackPrinceton in 2012
+
+My first hackathon was [HackPrinceton](https://www.hackprinceton.com/) in 2012.
+I was a senior in college, and I created [my GitHub
+account](https://github.com/dguo) because of this hackathon. I made a movie
+recommendation website called [MovieSeer](https://github.com/dguo/movieseer).
+The [code makes me cringe
+now](https://www.dannyguo.com/blog/a-code-review-of-my-earliest-projects#movieseer),
+but it was still fun to create something from scratch.
+
+I'll never forget it because at the end of the hackathon, I confidently
+volunteered to present first. I did a live demo, and the website hung for what
+felt like an eternity after I pressed the button to generate recommendations.
+When they finally showed up, people politely applauded, but that was my first
+lesson in the [perils of live
+demos](https://www.youtube.com/watch?v=znxQOPFg2mo). I'll always have sympathy
+when things go wrong in other people's demos, since it reminds me of the
+pounding in my heart when an entire auditorium of people was just waiting for my
+website to do something.
+
+My project actually won an award for best use of
+[MongoDB](https://en.wikipedia.org/wiki/MongoDB), which came with a $500 check
+from MongoDB. Which was hilarious to me not only because of my presentation
+issue but also because there was nothing about my project that actually
+benefited from a [document
+database](https://en.wikipedia.org/wiki/Document-oriented_database). I could
+have used [SQLite](https://en.wikipedia.org/wiki/SQLite) or
+[Postgres](https://www.postgresql.org/), and there would have been no practical
+difference. I only used MongoDB because of the potential prize. But in
+retrospect, that was the point of MongoDB sponsoring the prize.
+
+This hackathon also reminded me of the value of a good presentation. The crowd
+favorite was some beer-related mobile app. I don't even remember what the app
+did. Something to do with keeping track of how many drinks you've had. But the
+creator was by far the best presenter, getting the audience to laugh and really
+pay attention. It's not enough to just make something interesting. It's also
+important to be able to talk about it or show it in a way that engages others.
+
+### HackPrinceton in 2013
+
+In 2013, I went back to HackPrinceton because my company,
+[Bloomberg](https://en.wikipedia.org/wiki/Bloomberg_L.P.), was a sponsor. I have
+three strong memories from it.
+
+The first is one of the best things I've ever seen in any hackathon: [Piano
+Stairs](https://www.instructables.com/Piano-Stairs-with-Arduino-and-Raspberry-Pi/).
+This was a hardware hack that used light-based sensors to turn a staircase into
+a working piano! The creators used a staircase in the middle of the hackathon
+building, so it was impossible to miss, and everyone loved it.
+
+The second is [What Would I
+Say](https://thisisimportant.net/posts/what-would-i-say/), which used [Markov
+chains](https://en.wikipedia.org/wiki/Markov_chain) to generate semi-sensical
+Facebook posts based on your past posts and went viral.
+
+Lastly, I remember there being an award for "Best Fail" because the reward was
+an entire Bitcoin, worth maybe a couple hundred dollars then. I hope the team
+held on to it!
+
+### Bloomberg
+
+Bloomberg held its own hackathon a few years later. I worked with a friend on a
+UI transpiler. For context, the [Bloomberg
+Terminal](https://en.wikipedia.org/wiki/Bloomberg_Terminal) is a desktop
+application that is [rendered using server-side
+JavaScript](https://news.ycombinator.com/item?id=21821327) ([this conference
+talk](https://www.youtube.com/watch?v=ODgs0eWAIKc) also provides some details).
+
+My team made a proof of concept for a program that would take that code and
+convert it into an equivalent [React](https://react.dev/) application with the
+goal of making it possible to run the Terminal (or at least a subset of it) in a
+browser. It was an admittedly simple example, but it worked!
+
+It didn't go anywhere after the hackathon, but people reached out saying they
+thought it was cool. The ego boost was worth staying up almost all night to
+finish it.
+
+### MoMath
+
+In 2018, some friends and I attended a hackathon at the [National Museum of
+Mathematics](https://momath.org/) in NYC. We made a project called Vortex Pool
+that I [wrote about
+here](https://www.dannyguo.com/blog/momath-hackathon-2018-vortex-pool).
+
+### Firefox Quantum Extensions Challenge
+
+Later in 2018, Firefox had a hackathon, and I made a few extensions for the
+"Best Dynamic Theme" category, which used the [theme WebExtensions
+API](https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions/API/theme).
+I also wrote about [that
+experience](https://www.dannyguo.com/blog/building-dynamic-firefox-themes/).
+
+### SeedFi
+
+In 2020, the startup that I worked at, SeedFi, had a hackathon. I started it by
+fixing a slew of small issues that we already knew about, such as a set of icons
+not being the same size in Firefox. I also remember fixing an issue where iPhone
+users were limited to integers for number inputs as of iOS 13.
+
+Then I added
+[TOTP](https://en.wikipedia.org/wiki/Time-based_one-time_password)-based 2FA.
+This never made it to production, but I mainly did it as a learning experience.
+
+I had some time left over, so I also made a proof of concept for a [command
+bar](https://maggieappleton.com/command-bar) for our website.
+
+### Karmahack
+
+In 2023, [Intuit acquired
+SeedFi](https://www.creditkarma.com/about/releases/intuit-to-acquire-financial-health-startup-seedfi),
+and we joined its Credit Karma business. A few weeks after our onboarding,
+Credit Karma had its
+[Karmahack](https://engineering.creditkarma.com/karmahack-2021-recap-what-makes-a-good-hackathon)
+hackathon.
+
+I had no idea how to build anything in the Credit Karma stack/platform at that
+point, but I still wanted to participate in some capacity. So I went with making
+a browser extension that ran against the [Credit Karma
+website](https://www.creditkarma.com/), since that didn't require me to
+understand how any of it worked under the hood.
+
+I went with the command bar idea again, but I rebuilt it from scratch and added
+some fancy features, like being able to highlight a section of the UI based on
+the query. My demo example was someone looking for their checking account
+number, and the command bar would both take them to the right page and also
+highlight the account number [using
+Driver.js](https://driverjs.com/docs/simple-highlight).
+
+I won third place and got this custom medal.
+
+![front of Karmahack medal](https://i.imgur.com/cBi6Bvyl.jpg)
+
+![back of Karmahack medal](https://i.imgur.com/tOiY6Ckl.jpg)