ChatGPT makes work much easier for programmers. Nevertheless, the AI makes sense for advanced users in particular. As in any job, prospective software developers should first build up an understanding of software. It doesn’t matter whether this involves knowledge of a programming language, software architecture or a framework. Any tool, including ChatGPT, hinders the learning process and is therefore no support. Instead, beginners should gain as much practical experience as possible in order to deepen their knowledge. Another reason why newcomers should do without AI support is much more pragmatic: if you don’t have basic programming skills, you won’t be able to understand what the tool produces.

Instructions with structure and precision

Programmers who want to try out ChatGPT should also be prepared to learn how to use the AI in a similar way to a programming language. Basically, the more precisely the instructions are formulated, the better the results delivered by ChatGPT. It therefore makes sense, especially at the beginning, to divide the command into small-step sub-commands that build on each other. This allows the programmer to understand at each step how the AI implements its instruction and reformulate it if necessary. Functioning prompts – as the commands are called – should be saved as documentation.

Powerful tool in selected areas

ChatGPT also achieves particularly impressive results when it comes to identifying and improving vulnerabilities in existing code. It is also a powerful tool for code translation. This is particularly interesting when new employees join an ongoing project and have little or no experience with the programming language used.

ChatGPT has also proven to be very powerful when it comes to improving existing code. Regardless of whether you let ChatGPT search for weaknesses in the code or make the code more efficient: Chat GPT usually works very reliably. Translating code between different programming languages, for example from Kotlin to Swift and vice versa, is now also surprisingly easy. This makes a significant contribution to increasing the efficiency of programming teams, for example when they are working on Android and iPhone apps at the same time.

The ability of ChatGPT to explain code should also not be neglected. This is particularly helpful for new employees in a company who first have to familiarize themselves with the existing code base. With the help of ChatGPT, this familiarization can be supported so that the senior developer is no longer solely responsible for explaining the code base.

Junus Ergin

is Managing Director of the Developer Academy.