Techsplainers by IBM

This episode of Techsplainers explores Code LLMs, specialized AI models that are transforming software development by understanding, generating, and explaining code. We examine how these models differ from general-purpose LLMs through their extensive training on programming languages, documentation, and code repositories, giving them a deeper understanding of software development concepts and patterns. The discussion covers how Code LLMs work through transformer architecture and reinforcement learning from human feedback, as well as their substantial benefits for developers, including increased productivity, enhanced learning opportunities, improved documentation, and democratized access to coding. We highlight practical applications such as code generation, bug fixing, refactoring, and test creation, while also addressing important limitations including potential security vulnerabilities, challenges with complex code, intellectual property concerns, and the risk of over-reliance. The episode concludes with insights into how these tools are reshaping the development landscape, with developers increasingly shifting toward a supervisory role focused on architecture and design decisions.

Find more information at https://www.ibm.biz/techsplainers-podcast

Narrated by Erika Russi

This episode of Techsplainers explores COBOL modernization strategies for organizations managing critical legacy systems written in this 60-year-old programming language. We examine why COBOL remains vital today—processing 95% of ATM transactions and 90% of global financial transactions—while highlighting the pressing challenges of maintaining these systems amid a growing skills gap, integration difficulties, and scalability concerns. The discussion covers various modernization approaches, from less disruptive strategies like rehosting and refactoring to more transformative methods like rearchitecting and replacement. We share key best practices for successful modernization, including thorough application documentation, business-value prioritization, preservation of essential business logic, and incremental implementation. The episode emphasizes that effective COBOL modernization isn't about wholesale replacement but thoughtful evolution that preserves decades of refined business logic while enabling future innovation.

Find more information at https://www.ibm.biz/techsplainers-podcast

Narrated by Erika Russi

This episode of Techsplainers explores legacy code—software that continues to deliver value despite being inherited, outdated, or difficult to modify. We examine how these systems, though challenging, often represent significant business value and intellectual property, running critical operations across industries from banking to government. The discussion covers the main challenges legacy code presents: knowledge gaps from departed developers, accumulated technical debt, outdated technologies, poor documentation, and integration difficulties. We explore practical approaches to managing legacy systems, including incremental modernization through refactoring, the service wrapper approach, complete rewrites when necessary, and hybrid strategies. The episode concludes with best practices for working effectively with legacy code, emphasizing the importance of documentation, testing, incremental changes, and maintaining respect for systems that have successfully powered business operations for years.

Find more information at https://www.ibm.biz/techsplainers-podcast

Narrated by Erika Russi

This episode of Techsplainers explores how artificial intelligence is transforming one of developers' most dreaded tasks: code documentation. We examine how AI-powered tools can automatically generate human-readable explanations of code, bringing consistency, efficiency, and improved maintenance to software development teams. The discussion covers key benefits of AI documentation, including time savings, comprehensive coverage, and knowledge preservation, while providing practical tips for effectively implementing these tools. We highlight the importance of using AI as a starting point rather than a complete solution, establishing clear documentation standards, and integrating tools directly into development workflows. The episode also addresses current limitations of AI documentation and the evolving landscape of tools that are making documentation more interactive and contextually relevant. Whether you're a seasoned developer or new to programming, this episode offers valuable insights into how AI is solving one of software development's persistent challenges.

Find more information at https://www.ibm.biz/techsplainers-podcast

Narrated by Erika Russi

This episode of Techsplainers explores Code LLMs, specialized AI models that are transforming software development by understanding, generating, and explaining code. We examine how these models differ from general-purpose LLMs through their extensive training on programming languages, documentation, and code repositories, giving them a deeper understanding of software development concepts and patterns. The discussion covers how Code LLMs work through transformer architecture and reinforcement learning from human feedback, as well as their substantial benefits for developers, including increased productivity, enhanced learning opportunities, improved documentation, and democratized access to coding. We highlight practical applications such as code generation, bug fixing, refactoring, and test creation, while also addressing important limitations including potential security vulnerabilities, challenges with complex code, intellectual property concerns, and the risk of over-reliance. The episode concludes with insights into how these tools are reshaping the development landscape, with developers increasingly shifting toward a supervisory role focused on architecture and design decisions.

Find more information at https://www.ibm.biz/techsplainers-podcast

Narrated by Erika Russi