One of the most fascinating but also on the other hand most difficult part of software engineering is the programming of the software, app or program. This is where most of the expertise is needed and is the most vital step. It often takes quite some time to acquire the skills of programming and also once learned, one of the most challenging steps is debugging the code. This is where ChatGPT and other generative AI algorithms can help. At this moment, ChatGPT can solve simple to medium tasks efficiently, however, struggles when problems become more complex (Nascimento et al., 2023).
Talking from personal experience, I was always quite fond and interested in programming, however very often struggled getting the right code and implementing a bug free code, however, since I have discovered ChatGPT early 2023, this part has simplified my experience with coding and has led to the improvement of not only coding, but also debugging and providing documentation. ChatGPT is in my opinion nowadays indispensable in software development and engineering, as it can free up resources, which would have otherwise not been used efficiently. ChatGPT has revolutionized the way we create software.
Personally, I did also have experience with ChatGPT, especially the plug-in Noteable, where I could let ChatGPT create an algorithmic trading application, that gave the user notifications if there were buy or sell signs. Even though the algorithm was relatively simple and did not use high complex mathematical or logical formulas, it could, nevertheless, show what the capabilities of such a tool on the financial or fintech world could set upon in the future, where even more datapoints are connected. However, when looking and utilizing code, many underestimate the bias the code could entail. This has be viewed as a concern for ethical reasons (Hörnemalm et al., 2023).
Nevertheless, considering what the future holds, I don’t think that there is a replacement for software developers happening, as the problem solving and critical thinking processes cannot be replicated until these technologies reach the Turing test. This then leads to the question whether software developers in the future, also after the Turing test completion, are a thing of the past?
Hörnemalm, A., Norberg, O., & Mejtoft, T. (2023). ChatGPT as a Software Development Tool The Future of Development. https://www.diva-portal.org/smash/get/diva2:1768339/FULLTEXT01.pdf
Nascimento, N., Cheriton, D., Alencar, P., & Cowan, D. (2023). COMPARING SOFTWARE DEVELOPERS WITH CHATGPT: AN EMPIRICAL INVESTIGATION A PREPRINT. https://arxiv.org/pdf/2305.11837.pdf
Generative AI, the end of software developers?
14
September
2023
Interesting article Mr. Schally, I think ChatGPT still has a long way to go for coding! But indeed already shows promising results. I think In case of adaptability it would definelty change the way developers work as well if AI becomes so smart that it can even give an answer for code which is very novel but also needs a lot of thinking power, if that makes sense.