Our episode on ChatGPT and Seniors intrigued the old codgers. We’re not certain how many folks in our audience are tracking the developments surrounding ChatGPT, but we’re are still trying to understand what it means to senior citizens in their golden years as well as the impacts to the broader world. In this post, Richard writes about his surprising conversations about the unintended consequences of ChatGPT with some of his peers at Caltech.
I’ve been playing around with ChatGPT for a couple of weeks now. I have a lot of friends doing the same. We’ve mostly concluded that it’s clunky, makes numerous mistakes, and, even when it’s clever, it’s B or B+ clever. No one has seen A-level work yet. I even heard about some unintended consequences of ChatGPT on the very profession that created the beast.
Well, that’s not quite fair, and this story I’m about to tell potentially upsets the apple cart.
I went to the Caltech campus recently to take a look at my assigned classroom for a journalism course I teach each spring. I peeked into a nearby office to see if a colleague was there. Indeed he was. French professor Trevor Merrill and I hadn’t seen each other for almost a year, so we chatted away, about our courses, the pandemic fallout, all sorts of stuff.
Before leaving, I just had to ask him about ChatGPT. Had he heard about it? He certainly had, he said. It was all over the news and the campus. Had he tried it? No, he didn’t know there was a way to play around with it.
I helped him find the site and Trevor tried a few interesting questions, like asking ChatGPT to write a short essay comparing stories by Balzac and Stendhal. There were problems with the response, but he was amazed at how quickly that response came.
So, I asked, if you received a longer essay from a student with these as the main points, what grade would you give it? He laughed. Given grade inflation, I’d probably give it a B, he said.
Then he told me a story. A friend of his, a professor in the Computer Science department, had been wrestling with a coding problem for several weeks. On a whim, he decided to offer up the problem to ChatGPT and see what would show up. No more than a few minutes later, ChatGPT spit out the code. His friend went through it and checked it. The code was correctly written. It had done in 5 minutes what the professor had been unable to do in a couple of weeks.
A few weeks earlier, I’d heard a similar story from a Caltech seismic researcher. When he asked, ChatGPT had written most of the code he’d been working on. It didn’t finish the job, he’d said, but what it had done was correct, and he suspected if he asks ChatGPT again in a few weeks or months it will be able to deliver the whole code.
These episodes bring up a very interesting scenario, one that goes against expectations.
The New Yorker magazine recently published a story about the death of the humanities on America’s college campuses. For years, we’ve been consumed by STEM, STEM, STEM. If students aren’t studying Science, Technology, Engineering, or Math, they should expect to settle for second-tier jobs with low remuneration. And even those humanities-related jobs are likely to be replaced by machines or other new technologies.
Reflecting on that new reality, students at Caltech have been convinced that the most exciting opportunities lie in the world of Computer Science, or CS. Over the past half-decade, scores of Caltech CS graduates have garnered great-paying jobs at the so-called Big 5 — Amazon, Google (Alphabet), Apple, Facebook (Meta), and Microsoft. The largest major at Caltech is no longer engineering or physics. In 2021, students studying Computer Science constituted almost 40 percent of the declared majors on campus, followed by Engineering at less than 30 percent and Physical Sciences at 18 percent.
Yet here we have this new technology everyone is raving about — a product of Computer Science — and among its first effects may be the elimination of jobs in computer programming. This a rather interesting case of the unintended consequences of ChatGPT’s emergence.
Why wouldn’t a company ask ChatGPT to code something before it hired a new programmer? Why wouldn’t a company ask ChatGPT all kinds of coding questions and do away with its current programming staff?
So here we are, left with the irony that the people creating this new technology, this new tool, are engaged in inventing their own unemployment.
Will it also result in the unemployment of French professors and Journalism teachers? Left to be seen.
Don’t forget to listen to the old codgers on your favorite podcast app or on the Camp Codger podcast website.