As computer science instructors, people often ask us if this or that GPT (Generative Pre-trained Transformer) is going to take away all the programing jobs. We know that some people are worried, but human programmers are not going away anytime soon. Human creativity is not so easily replaceable.
GPT models are great at generating text, but they can’t come close to the creativity and problem-solving abilities of humans. Computer programming requires imaginative thinking, finding solutions and translating ideas into code. These are skills the models can’t do alone — but they can make a programmer much more productive.
Spreadsheets have auto-completion and auto-calculation routines built in, and they’ve replaced hand-calculating long list of numbers — but they can’t replace the person who looks at the spreadsheet and decides what to do with those numbers.
Weighing Costs and Benefits
Programming is more than just writing lines of code. It involves making complex decisions. While GPT models can offer suggestions or assist with specific aspects, the buck stops with human programmers and their judgment. A model can suggest solutions, but it’s up to you to select the best solution.
Adaptability is a Human Trait
Programming languages change over time. Skilled programmers adapt to these changes, continuously learning and updating their knowledge. The most important thing we try to teach our students is to “learn how to learn” because the hottest thing today won’t be the hottest thing tomorrow.
GPT models, on the other hand, rely on existing data and patterns. They can’t learn new skills or adapt to changing circumstances. That’s up to us.
Collaboration Between Humans and Models
Instead of seeing GPT models as adversaries, we tell our students to look at them as tools. They can offer suggestions, provide code examples, and much more. By working alongside GPT models, programmers can leverage their capabilities while contributing their unique insights and expertise.
In fact, we’re using these kinds of models to help our novice programmers, not just by letting them ask questions about “How do I write programs?” but also exploring the model to figure out what are the right questions to ask.
We found that when students knew what kind of web page they wanted to make, for example, they would keep asking the model for help until they came up with a web page that matched what they had imagined. It truly was a group effort between programmer and model.
The field of computer programming
GPT models have made remarkable strides in generating text and supporting specific tasks. However, they aren’t positioned to replace human programmers. The field of computer programming thrives on human creativity, problem-solving abilities, critical thinking and adaptability to evolving technologies.
By recognizing the collaborative potential of GPT models, programmers can harness their strengths and prove that humans have to be the ones doing the thinking. We invite you to learn more about computer programming at FTCC. We’re registering students now for fall semester and would love to serve you. For more information visit https://www.faytechcc.edu/academics/computer-information-technology-programs/
Editor's note: David Teter and Andrew Norris are Information Technology faculty members at Fayetteville Technical Community College.