Starting in the Middle

Nothing begins cleanly. I remember approaching AI wondering what I even wanted to know. Did I… do a cute search with it? Did I ask it trivia questions? Did I let it help me finish my sentence? I don’t remember what I ended up throwing into the void. Or even what I got back… It was just…. okay? This was it? It was like going to the zoo and the animals are just sleeping. And if you look really closely, you might see a tail flick. I wasn’t excited. I kept reading about AI everywhere, but I didn’t approach it really again. I didn’t see any reason to. Sure, I may bump into it. Alright, I would deal with it then.

Last fall, I took a course where our professor wrote in bright bold red in our syllabus that we were ENCOURAGED to use ChatGPT to complete our assignments as much as we could for completing python codes that would work through engineering experiments. He circled the paragraph during the orientation lecture on his projected screen and then cut over to his python environment where he demonstrated how ChatGPT couldn’t correctly write a program for calculating square roots. He was challenging us to find there was a limit, to show us that true engineering couldn’t be replaced. And I excitedly took that challenge. I wanted to find the limit too, and do less homework from scratch. Right?

I dove in. And I watched the videos and I read the articles. And they helped a bit, but I saw where the work was skipped. I saw confusion. I saw the inaccuracy. And I kept coming back. Because I saw something else. Not perfection, but progress. If I adjusted, ChatGPT adjusted. And we kept bunny hopping until we found a rhythm that worked. And I had codes that worked. And when I went in to ask other how they were dealing with certain interactions with ChatGPT, I found out that a lot of my classmates had found their limit. I kept talking, convincing no one to try again at that time. Meanwhile, my ChatGPT was not only coding with me but helping me choose new perfumes based on ones I already liked and laying out what ended up being a wonderful itinerary for a Hawaii trip with my family… and I couldn’t figure out what it was I was managing to do any differently.

It wasn't until I read a headline about the next version of ChatGPT that the answer finally clicked. Developers were promising an exponentially better model because it would have "more data," but I realized they were singing off-key. The problem wasn't a lack of data; the problem was a lack of context. People don't need AIs with bigger brains; they need AIs that know how to meet them where they are.

While I may not be able to reach the developers, I can certainly reach others who have hit their own limits with AI. The goal of this blog is to help you build your own bridge into a better, more contextual experience with these tools.