IBM Generative AI for Code with watsonx Code Assistant for Z

I was amazed with IBM’s new Generative AI watsonx Code Assistant tool for Z (Mainframe Application Modernization), and I was able to explore three key features like application discovery, auto refactoring, and Visual Studio Code (converting COBOL to Java).

It doesn’t require to modernize entire application as it optimizes the existing COBOL code to perform better and provides state-of-the art visual flow on understanding the old code and what it does. The selective modernization approach would benefit the customers to save the cost.

It ignited me to look back at Mainframe as I have 11+ years of previous Mainframe technical experience (Assembler, COBOL, CICS & Web Service, DB2, SQL, MQ, VSAM, Java, HTML, XML, JCL, PROC, Ezytrieve, REXX, IDMS/DB, CA-Datacom, and IMS/DB).

I need to know your thoughts on ‘‘IBM Generative AI for Code with Watsonx Code Assistant for Z’’ tool.

I feel as others might have already shared: the technology is new and exciting yet not mature enough to make the impact that AI hypes it to be today (e.g. replacing programmers/developers). That being said, it will improve with time and additional training models/data so it’s good to see a path for converting COBOL to Java - but I wonder if we have more COBOL programmers, wouldn’t it be best to continue with the original language?