Latest videos
A Code Centric Journey Into the Gleam Language
Where AI Meets Code
Optimizing Cloud Native Java
What's up with Rust?
Building Modern Software at Scale: Architectural Principles Two Decades in the Making
Ant Farm Entropy: Sugar Powered Encryption
Java's Hidden Gems: Tools and Libraries
Prompt Engineering for Generative AI
Vanilla Web: You Don't Need that Library
37:15

CodeConcise: A New Era for Legacy Modernisation
So far, most attention to Gen AI in software development is on generating code. But there is as much, if not more, value in understanding existing code - particularly long-lived, large, and complex legacy systems. At Thoughtworks we have been experimenting with using this technology with our clients, embodied in a tool called CodeConcise, which combines an LLM with a knowledge graph derived from the abstract syntax trees of a codebase. We have found this useful both to draw out low-level requirements and to build a high-level explanation of a system. This talk will show what we have learned and how it could change the age old challenge of wrangling legacy at speed, iteratively and with zero downtime.