Approximately 200 students gathered in Lobdell Dining Hall last Saturday to participate in the first phase of MakeMIT, a hardware hackathon organized by TechX. While the past year has seen college hackathons (including TechX’s very own HackMIT) increase in both scale and number, most of the emphasis has been on software, with few options for non-computer science students to get in on the action.
“We saw there were hackathons happening basically everywhere across the U.S. Hackathons provide a great environment for Course 6 people,” said Thuan D. Doan ’15, one of the event organizers. “We said, why isn’t there something like that for hardware? For MechE’s and for 6-1’s? One of the reasons why is because there’s such a high barrier to entry. We thought this would be a great opportunity to fix that.”
MakeMIT debuted this year with about 50 teams of three to five students each. Registration opened in January.
“MakeMIT provided me with an excuse to ditch my daily monotonous life, and build something just for the fun of it,” said Emma M. Steinhardt ’16, a student in Course 2 and first-time hackathon participant. “I haven’t done a hackathon before, because they are all for software-related things, and I’m into designing mechanical things.” In fact, mechanical engineering was the best represented course at MakeMIT, the major of 46 percent of the hackers. Another 36 percent were from EECS, with the remaining participants coming from a variety of majors and skill backgrounds.
While a number of the big college hackathons provide support for hardware, it’s often tacked on as a sideshow to the software projects. Skyler E. Adams ’16, one of Steinhardt’s teammates, recalled his experience working on a hardware project during HackMIT. “Doing something hardware-oriented costs money, and if you order materials it means your idea is somewhat pre-meditated. [My group] and I built these motor systems from spare parts we had lying around, but the people that looked at our project thought we had started much earlier.”
At MakeMIT, all tools and materials were provided, including 3D printers, Kinect sensors, microcontrollers, sensors and input mechanisms, motors and actuators, and more. In order to level the playing field, no outside materials were allowed. “We spent a lot of time just thinking, what is the perfect amount of materials? What materials do we get to allow teams to do everything they want to without costing too much? Having a team of ten people try to speak for 200 people is not very easy. It’s a huge risk.”
The event organizers consulted with professors and makers to create a comprehensive list of materials, but some teams were unsatisfied. “I think next time they should get some input on what materials to stock,” said Adams. His team built a physical arcade-style version of the late mobile game Flappy Bird, featuring a 3D-printed bird avoiding pipes on a scrolling LED matrix. They went in expecting individually addressable LED strips, but ended up having to hand-solder the array themselves. At a hackathon — where time is the most valuable resource — this was a huge productivity sink for the team.
Another bottleneck came when teams were allocated only a limited amount of material to laser cut. Many teams opted to 3D print their larger parts instead, which “took ages and almost DoS’d them,” said Adams, comparing the lines to a denial-of-service attack. Some projects had to be cut due to insufficient time for 3D printing and laser cutting near the end of the hackathon.
In fact, time was even more valuable than usual for Saturday’s hackers. Typically, hackathons range from 24 to 48 hours in length, but MakeMIT was instead broken up into two one-day phases, due in part to safety concerns related to hosting an overnight event. “If we have a drilling or Dremel section, having hackers that are sleepy and tired operating those is obviously a risk to safety,” said Rachel S. Wang ‘16, one of MakeMIT’s co-directors. “We really did try to push for an overnight hackathon, but [since this was] year one, we saw that it wasn’t feasible.”
At the end of Saturday’s event, teams presented a diverse range of projects, ranging from Wakey Wakey, a silent alarm clock, to ShotBot, a robotic bartender. Other projects included a relay baton that tracks split times based on handoffs and a oscilloscope probe built on a budget of under $50. One team, which included one of the developers of Tidbit, created another Bitcoin-related hack, exposing a design flaw in the official Bitcoin wallet.
These hacks were judged based on three criteria — functionality (how successful they were that day), potential for success (how much further the project could be developed), and hack factor (resourcefulness and creativity). Only the top ten teams from phase one are being invited back next weekend for the second round of MakeMIT, where they will further iterate upon their prototypes.
Taking first place and $2000 in prize money was a guitar-playing robot capable of both strumming and fretting. In second place was GoPro DataPac, an attachment for GoPro cameras that records data about action sports, such as velocity, altitude, rotation, and acceleration. LexoGlove, an exoskeleton glove that teaches the deaf-blind how to perform American Sign Language fingerings, took third place in Saturday’s competition.
Elizabeth Zhang ’16, one of the hackers who built LexoGlove, appreciated the fact that the hackathon was split over two days. “I don’t really have the stamina for [an overnight event.] I like sunlight and fresh air.” Her teammate Julia C. Canning ‘16 agreed, saying that the extra week gave the team “time to refine [their] design more and time to order materials.”
The teams advancing to phase two of MakeMIT now have an opportunity to request a bill of materials. “Obviously, now they have more of a sense of what materials they want,” said Wang, “so we’re going to get those for them so we can be prepared.” Additionally, teams will have access to a machine shop and mentors, giving them a chance to develop their prototypes into more complete products.
LexoGlove currently uses a servo-based underactuated mechanism to pull the wearer’s fingers in and out to show them the sign corresponding to a particular word. This approach was sufficient for a proof of concept, but servos can be bulky. Before next Saturday, Zhang, Canning, and their third teammate, Edwin H. Zhang ’16, will think through the next iteration to make it thinner and lighter, perhaps by switching to a linear motor.
“At this point, it’s a bit less competitive,” noted Doan. “Everyone’s kind of made it, and now it’s more of a we-want-to-make-this-work kind of thing.” And the same can also be said of MakeMIT itself — despite the current trend among hackathons to “grow big,” scaling is far more difficult for a hardware-based event, where materials are paramount.
Unlike HackMIT, MakeMIT wasn’t trying to reach hackers beyond the Boston area. Nevertheless, 30 percent of participants were from outside MIT, showing a growing demand for hardware-based opportunities. “I definitely want MIT and other schools to do more hardware hackathons,” said Zhang. “I keep telling all my friends who do software, there’s no virtual without the physical.”
Despite a few logistical hiccups, the feedback from sponsors and participants was positive. In the near future, MakeMIT may accept a few more teams, but the organizers don’t want to expand at the cost of quality. Instead, they believe MakeMIT can inspire other schools to organize hardware hackathons. “Our hope was that MIT can lead the charge on this thing and show that it is possible. Hopefully, across the country, we’ll start seeing more of it.”