I approach the end of not only this course but of my time here at OSU scratching my head, wondering where the time went and where we might be headed. It feels like only yesterday that I wrote my first fizzbuzz script; now, I’m using AI tools that were unheard of at the start of this program to help me build an entire webapp from scratch. Like many of us, I’m ambivalent about AI – excited about its promise of code-on-demand, but anxious about what that means for programmers in particular and creatives in general. Before this quarter, I had fallen into a pessimistic view of this technology. It felt like I was earning a degree in abacus studies right before the debut of the calculator. My key takeaway from this course and from the AI Coder project is that there’s a hopeful alternative to this take, one that preserves the dignity of human work and presents a model for AI collaboration moving forward.
Working on this project reminded me of the success of so-called “centaur” teams in freestyle chess tournaments. These teams consist of one or more chess players and a collection of chess engines working in tandem. The humans on the team focus on overall strategy, while their AI teammates provide them with tactical options. Similarly, each member of the AI Coder team formed a centaur-like relationship with their assigned AI assistants.
In my case, I mapped out a big picture goal – creating the PawPair animal “dating” website – while Google Bard and GitHub Copilot offered me potential paths forward. Copilot excelled at producing code in an existing codebase but had difficulty explaining how it arrived at a particular solution. Bard, on the other hand, was great at explaining concepts and guiding me step-by-step through its approach. In fact, for the more complicated SQL queries, I would take Copilot’s code and paste it into Bard to learn how each nested command worked. With all of this information in hand, I took on the role of project manager, selecting the best of myriad possibilities, delegating work to my AI teammates, and righting the ship when errors inevitably arose.
My role on this “team” hints at the role humans may play in an AI-dominated future. While we can’t ever hope to produce code as quickly as these assistants, their output is only as good as the input they receive. That input, in turn, depends on the prompter’s creativity, imagination, and ability to effectively express their ideas. Our ability to think beyond the immediate requirements and imagine what else might be possible is something these AI tools may never be able to fully replicate. It’s true, they can provide us with endless options, but human judgment will still be needed to determine which ones align with the larger vision.
We needn’t only think of AI in such serious terms though. One thing that surprised me during this project was just how fun this collaborative back-and-forth could be. In the past, I had to put so many interesting ideas aside due to constraints on time and effort; now, using Bard and Copilot, those flights of fancy are just a prompt away. And seeing almost immediate improvement to one’s code at the drop of a thought doesn’t cheapen the experience but creates a motivational snowball effect instead.
I haven’t even focused on the improvements AI made to my productivity and output. With just the guidance of my AI helpers I was able to produce a website that surpassed the aesthetic appeal and functional complexity of what our human-only team had achieved in the Databases course. Don’t get me wrong, I’m proud of the work we did, but I can only imagine how much more we could have accomplished if we had an SQL-generating AI at our disposal.
And so, perhaps the situation we aspiring computer scientists find ourselves in is less like learning how to use the abacus and more like getting a mathematics degree right as the calculator comes out. That is to say, the way in which we learned C.S. will likely be obsolete, but the fundamental skills and understanding we gained will still help us to utilize these new tools and remain relevant in the age of AI. Just as calculators, computers, and various other advancements allowed mathematicians to work more efficiently, freeing their minds to explore new frontiers, AI holds the potential to transform the nature of human work, enabling us to focus on what we do best: harnessing our creativity, intuition, and holistic thinking in service of a greater good.
Sources
Google Bard – https://bard.google.com/
Microsoft Bing Image Creator –
https://www.bing.com/images/create/google-bard-personified/6544fa257b534083897447b5986afddb?id=E0ykR%2fRkVKFhVfVA4iHe6w%3d%3d&view=detailv2&idpp=genimg&FORM=GCRIDP&mode=overlay
