Surprise upcoming changes for course selection
Last week, the administration announced that Cognomos will be retired for student course matching in the 2023–24 school year. Instead, all course matching will be conducted through a new system incorporating OpenAI’s artificial intelligence (AI) image generator DALL-E.
DALL-E made a stir in early 2022, when OpenAI announced that they trained a neural network to generate images from plaintext descriptions and provided public access to the AI system.
“We value student feedback, and we’re always looking to improve our famously arduous and unpredictable course matching process,” said the administration in their formal announcement. “We’re very proud of this decision as it aligns with our current reasoning of using award-winning algorithms that have little to no relevant application in the context of course selection.”
In an interview with Ultra Vires (UV), a member of the administration lamented that “the only downside is that DALL-E is open source, which means the Faculty will no longer be able to pay undisclosed sums for a non-functional third-party system.”
A test run by UV editors confirmed that inputting the text “600 happy U of T law students with enough credits to graduate” still generates a set of timetables with less course conflicts and better credit distributions than this year’s Cognomos results. Editors were also pleased to note that the timetables could be generated in various colours and art styles.
“I guess it can’t get any worse,” commented Sarah Study, a 3L student who had been allocated 30 credits in fall and three credits in winter—and is still missing the mandatory Legal Ethics course despite ranking it first on her “Great” list.
When asked if there would be any solutions for those currently suffering from Cognomos issues in the meantime, the administration was quick to reply, “We hope you’re happy with your allocated schedules, and good luck. And if not, there’s always movement on the waitlists!”
