#30 The Future of Coding: ChatGPT and the Rise of Code Reviewers
Throughout my writing, I've explored a wide range of principles, processes, and patterns that software developers have both loved and hated. One of the most contentious among these is code reviews. Some developers despise having their code reviewed unnecessarily, while others relish the opportunity to flex their skills and even potentially punish others through the review process. Meanwhile, some lead developers and architects find code reviews tedious, while others approach it with a sense of duty, doing what needs to be done without strong feelings either way.
The role of code review is going to change. No, it's not going away. What's going away is coding (good question to ask here who was the baby and who bathwater between coding and code review). Now with ChatGPT and Generative AI, no one will write code from scratch. No one will even debug code, at least not the entire length of it.
Every developer will now onwards be promoted to the role of code reviewer. Those who are not capable of reviewing code and only did copy-paste, need to either upskill or find another profession.
But, what about the process of code review? Will it also change? The answer is yes. With the emergence of AI-powered DevOps platforms like Roost.ai, code review will become more efficient and streamlined than ever before.
领英推荐
Roost.ai uses ChatGPT and Generative AI to automatically generate test cases and create ephemeral DevOps environments. This makes the code review process faster and more effective, freeing up developers to focus on higher-level tasks.
By leveraging Roost.ai's capabilities, developers can be assured that their code is being reviewed thoroughly and comprehensively. The days of manually reviewing code line by line are coming to an end, and the role of code reviewer is set to become even more crucial in the software development process.
To learn more about how Roost.ai can help streamline your DevOps process and improve code review, visit us at https://roost.ai.
Principal Consultant / NITI's AIM/ATL Mentor
7 个月with llm generated code and used in product, how will license part work. unit testing will be important with llm generated code. long ago i worked for a product which was 25 year old, the first level of review was on indentation, formatting of the code. with auto generated code, i think this should become zero unless company defined coding standards are different. some day prompt engineer should be able to provide the template and naming convention which llm need to follow to generate the code. in second level a person who knew better than developer would review and before merging into main the one who has worked on product since beginning will review. it were the days when there was no devops, no branching. what will be the role of sonarqube kind of tools doing static code analysis, llm itself can take care of it, like cyclic conditions, breaking of funciton into smaller size.
{views expressed are personal}
1 年You should talk to AICTE (the ex warden of Nilgiri is now heading it ??) about including such tools in the engineering curriculum. I guess it could really help students with work life balance, when they start their careers.
Interesting post! Thanks!