As anyone familiar with Buggy knows, the BAA is constantly looking for opportunities to better connect students, alumni, and the wider community with the sport that we all love. For example, this year the BAA has worked to improve our support for new teams by refreshing the buggy build book, expanded our history database with race details from the 1980s and 1990s, and started to provide limited video coverage of freeroll practices.
With that in mind, we wanted to find a new way for fans of the sport to engage with the vast collection of knowledge that is contained in the BAA community. As many Carnegie Mellon students and alumni know, AI is proving to be a technology that is allowing people to learn things that just a few years ago they would have never believed to be true.
Of course, we already have experience in the BAA with Artificial Intelligence through our raceday predictions provided by MechaJockey. However, we wanted to find a way for the greater community to actively engage with our data and alumni experience from the last hundred years. MechaJockey is not efficient enough to run full time interactively (and besides, who knows what MechaJockey might do to us meatbags). Thus we needed to develop a new, vibrant, and ovine-forward approach. A chatbot was really the only possible answer to meet these needs.
With that in mind, we herded multiple departments at the university including the CS department, Computing Services, and Landscaping to shepherd the necessary skills to bring such an experience to life. Together, this group of experts have developed a new form of LLM that we call the neural flock (paper to be published later this year in the respected journal “Colamborations of the ACM”).
Thus, today we are pleased to announce SchenleyAI, available for public use on our discord. As an early beta, SchenleyAI is able to answer simple yes or no questions. It has been extensively trained on all of our most active discord discussions, as well as the older forum archives, history data, photos, rolls reports, and miscellaneous other buggy content we have at our disposal. As it happens, the insights provided by our most vocal alumni proved to be the key to bring this vast knowledge together.
Today’s release is the beginning — a beta test capable of responding to your yes or no questions about buggy (we’re keeping it simple, as things always have a risk of going wrong with new Buggy Software). You can try it today in any of our public discord channels. Simply preface your question with ‘$schenley‘. For example:
Our experiences so far have been Not Baaaaaad, but we look forward to seeing what everyone else can bleat out of it. We can hardly imagine all the new concepts that AI will bring to our favorite sport. Log into our discord and give it a try today!
