Just Willonanza: MIT for a Day

As my writing instructor pointed out, I failed to send an email last week. I sincerely apologize. I will explain why in this email and I think you will agree that the wait was worth it.

The last week was an absolute grind. I stayed in the library until 11 or 12 every night, desperately trying to finish all my homework for the week. Why the rush? I had a flight to Boston, Massachusetts on Thursday, so I had to be done by then.

I took my physics exams, which were different from all other exams I've ever taken. Half the test was online, blending multiple-choice and free-response questions. The other half was an in-person, team exam in which we basically "performed" our physics knowledge for our teacher. It was much less stressful than I expected. I had to postpone some of my business writing homework and my religion midterm, but I finished everything necessary to survive the week.

I took the FrontRunner up to the airport on Thursday with nothing but a backpack, some clothes, and a dream. I visited MIT when I was little, but this was another chance to see it in all its glory. I flew to Boston that evening, arriving around midnight. Then I went with my friends to the hotel. We slept very peacefully.

In the morning, I walked over to this mediterranean place that my aunt recommended. This was the absolute coldest I have ever felt in my life. It said it felt like -16 degrees outside. I think it was lower. The snow was piled higher than the cars were tall. After eating a delicious pita, I walked over to the Stata Center to check-in and figured out what the hack (hehehe) was going on.

The Stata Center looked like something out of Meet the Robinsons on the outside, but inside it felt like a colorful warehouse. It was very strange. We got checked in and they gave us crewnecks, socks, and name badges. Things were starting to get official. Eventually they broke everyone into workshop groups and had the companies present who they were and what they did. This was really interesting. Turns out that quantum computing is, like, actually a big deal or something? Who knew? NVIDIA was there, and they told us that they can't even design their own quantum cicuits anymore. It's getting too hard, so they just have AI do it. I will say, although the environmental and ethical concerns of AI definitely need to be addressed and I am not minimizing that problem in any way, AI is capable of some pretty amazing things. Finding drugs that lower the activation energy of enzymes by 70%? That's pretty cool.

Our team split up to hear all of the companies give their shpeals. Afterwards, we decided how to rank the problems. There were 9 companies, all of whom had brought some kind of problem for us to work on. We still had no idea what exactly those problems were; we only knew what types of companies we had to choose from. They gave us tiny hints as to what the problem might be, but nothing that could actually inform a decision. We rated NVIDIA and the world's leading quantum computing company, IonQ, the highest. We ate some good Indian food, then went home and started brainstorming the potential problems and our desired deliverables for the competition.

We woke up the next morning to the news that we had been given IonQ's problem. It was our second choice, and it involved noisy quantum states. This was a good thing, because Nishant (a research teammate) had a lot of experience with noise, so we felt both excited and anxious about the potential problem. We walked in somehow colder weather than the day before to their actual student center. They gave us Dunkin Donuts for breakfast because, well, America runs on Dunkin--everybody knows this. At 10 am, they threw all of us in some rooms and had representatives from each company stand up to introduce their problem. They hit publish on the Github Repos (things that hold code) and said, "you've got 24 hours."

It was crazy. We had no idea what we were going to do. We started reading the code they'd given us and figured out that we would essentially be playing one giant game of "Ticket to Ride" and "Risk" smashed together. Better still, game moves would require a knowledge of both quantum principles and code skills to move throughout the map. I have done quite a few projects with Git before, but my install was messed up and I couldn't download things to the right folder. Two hours went by before we had successfully opened the code on our own computers.

The game worked as follows: a website connected to every team's code and displayed a world map. This map had a lot of cities (nodes) that were connected by lines (edges). Each edge was ranked by how hard it was to take (difficulty threshold). In order to "claim" an edge and move to another city, you had to "spend" money (Bell pairs) to create a quantum circuit and "attack" an edge. Harder difficulty edges had were harder to take (more noise) and required a higher threshold to claim, which could either be met by sacrificing more Bell pairs or with clever circuit design. Such was the nature of the game.

By the time we had the code up and running, some teams already had hundreds of points. We thought we were doomed. 7 hours went by and we still had yet to claim a single node. It was so discouraging. One team had run code called "claude_test134" and had over 450 points on the board. We had 0. It was frustrating. 11 hours in, we were still researching the problem and trying to figure out what we were even supposed to do.

Then, a breakthrough.

Nishant successfully claimed one edge. We were on the board! We were in last place, but it didn't matter. We'd done something. We came together as a team to figure out what we'd figured out in 12 hours. I had been working on making algorithms that could traverse the map once we'd figured out how to attack a node. A couple of my teammates had been working on figuring out how to attack each difficulty level. We pushed forward with our respective ideas. My teammates were able to perfectly reverse engineer the noise present in difficulty 1 edges (the easiest), which meant that our attack circuits for those edges would be the theoretical perfect ones. We had found something. Difficulty 2 proved to be similar, so we found that one pretty quickly as well. D3 took a little longer, but we eventually got it.

A couple hours later, they kicked us out of the MIT student center and sent us on a midnight walk to a nearby hotel. They opened their conference room for us and we sat down and got back to work. I completed complex algorithms for path searching and optimization, which would allow our bot to enter the game and take over automatically. A couple hours later, we cracked D4.

At around this time, we got a message from the sponsors. At 1:30 am, they messaged everyone with a new piece of information: they had completely wiped the database and everyone had to start over. Turns out there was a bug in the setup that allowed teams to repeatedly attack an edge without having the proper fidelity, and since quantum is probabilistic, they would eventually succeed. They were also violating the non-locality principle we were supposed to follow, but we had ruled it out of our solution space long ago because the rules forbid it.

About 30 minutes later, Kai sent me the attack circuits for difficulties 1-4. I plugged them into my algorithm, started a new session, and pressed play. We watched the bot claim city after city, with small messages telling us that it had succeeded. Our team shot into first place. Nobody else had claimed anything. We were ecstatic. We went from being completely confused and in the dust to having perfect attacks and bots to run them. We started designing other types of algorithms that would spread across the map using different philosophies, like greedy, return on investment, and cluster identification. Eventually, we combined them all into a one hybrid model that would switch between modes based on map and organism conditions. It was pretty sweet.

Kai also made a playable version of the game on his computer so that he could just click and play manually. This was also a big deal, because it let us test out our perfect attacks on anything we wanted to. By the end of our algorithm design, we were actually starting to run into our own bots, which was a problem, because those bots had perfect attacks and therefore couldn't be beat. That meant we couldn't test our algorithms any further. It was annoying yet hilarious. We had almost done too well.

Somewhere in the next 6 hours, we wrote a paper, made a slide deck, built a website, and ported Kai's game to that website (with formatting for mobile). It was crazy. I made some Python animations with Manim to illustrate how our algorithms worked. We packaged it all up into a Github Repo and sent it in for judging. 10 minutes and 2 bagels later, we walked to another building to present our solution to the company representatives. I think I forgot to mention... we and all the other teams stayed up all night long to work on the problem. We didn't sleep at all. It was insane.

We were the last group to present. It was crazy. One team named "Quih" *dying rose emoji* was a sight to behold. They were five clones of each other, each doing Fortnite dances and saying things like, "it's not that shrimple." We got to see everyone else's solution and see how smart they were for what they did. There were kids from MIT, Yale, and Algeria. There were kids in CS, Physics, and Math. There were PhD and grad students too, all competing against us. I wasn't nervous, I was just excited to show off what we did. It was a really cool problem and I felt like we had a cool solution. We gave our 8 minute presentation answered some questions, then went back to the Stata Center for lunch.

We were slightly late to the award ceremony so we didn't sit together. Kai and I stood in the very back of the giant lecture hall. They announced the winners the QuEra challenge, and we were just happy to see it all happen. IonQ was next. They asked for a drumroll and turned the slide. It said "Q-Gars". I didn't even recognize that it was our name for about 3 seconds. I couldn't believe it.

We won MIT's international quantum hackathon! We took first place!

We were beyond excited. We walked down to the bottom of the hall where they gave us each our certificates, a water bottle, and a Nintendo Switch 2. It feels like a dream. Maybe because I hadn't slept in 30 hours. Anyway, we took a couple pictures, then we had to rush off to the airport to fly home! I got back at 12 am that night. I was so tired.

Upon arriving home, we also learned that our Quantum Information Student Association had finally been approved. I am acting as co-president with a couple friends from research. We are really excited about this and already have a lot of students and companies who want to participate. We also want to start something like iQuHACK in the west at BYU, since there are only 3 quantum hackathons in the US. The MIT one is the only one that is very popular. I'm excited to see how that goes.

I hear that there may be some interesting articles to read about our experiences. I'll send more information on those once I have it.

It is so late and I have school in the morning, but there's your big update on what's happening in my world. I'd love to hear from all of you. Email me!

-will

PS: for those who are interested and can view websites, here's the one we built
https://qgarsiong.vercel.app/iQuhackweb.html#game

And for the first time in a while, here are some photos. I'll send more later when I get them back.

Presenters Extraordinaires