ChatGPT has been praised for making work easier for many and among these is its ability to execute well-delineated coding tasks. The output usually needs a little tweaking but it is great to achieve tasks in a domain. It can also save developers a lot of time from reading documentation. Interestingly, when coupled with GitHub copilot (also powered by GPT), some of the problems in the generated solution become apparent fixes themselves when actually writing the suggested code in an editor. But will this AI tool replace software developers in the near future?
ALSO READ: GPT-4: 10 things you can do with the new ChatGPT
No, at least not soon since ChatGPT or any other AI-based system is not expected to make programming languages like C#, Python, PHP, and JavaScript obsolete. Programming languages are tools used to build software applications, and they serve a specific purpose. While AI-based systems like ChatGPT can help automate certain tasks and improve development workflows, they do not replace the need for programming languages. Moreover, programming languages are constantly evolving and adapting to new technologies and demands, which means they will continue to be essential in software development for the foreseeable future.
ChatGPT uses examples from the web. If you ask the AI tool to make code, then Google for a line of that code, it is most likely that you will find the same answer on Stack Overflow or somewhere like that. When I requested it to make me some C# to do something, then Googled for a line of it, and found the code on Microsoft’s C# forums.
The AI is not creating new code, it’s copying/pasting from the web. But the beauty about it is that it’s able to aggregate examples from several sources from the web, not just use a single example verbatim though. This is interesting and impressive, but it’s not programming. So for software developers, it’s a big sigh of relief.
Flipping the dev-ChatGPT coin
On the other hand, What generative AI can do is churn out complete programs integrated together from references in the language that you want to target. Meaning: all of these languages, indeed every programming language, will become obsolete at some point but generative AI won’t make them obsolete…this can be easily done by the industry itself.
Eversendapp developer Frank Ondokara recently said in a tweet that “ChatGPT just made my code significantly more readable, & improved error handling too, by suggesting a while loop over my recursion approach. Looks like I owe @Kasendwa an apology for tearing down his claims that it improves productivity. Helped me resolve an annoying bug”. So this is a clear testament that developers need to embrace and adopt the use of these new generative AI tools but with caution.
In most cases, Generative AI might produce an API server in PHP using your favorite framework but it’s unlikely to have the capability to verify that the code produced doesn’t also contain the flaws that were present in the code that the AI referenced. Not only that, it’s unlikely to produce the most efficient, most maintainable, most secure code. But this is also the same for humans writing code.
On the other side of the coin however, in an industry where low cost and speed are paramount and offshore development by companies in the West is often sought to control budgets, I think we are very likely to see that code produced by tools like ChatGPT, Code Pilot, etc., entering production warts and all since it is a comparable proxy to hiring less skilled workers in the first place. This could hurt the employment of junior software developers. It could also see the job of a Senior team member transforming into some sort of verification expert.
