







It's the golden age of Domain Driven Design in software.
I've always liked DDD, where you model software in real-world nouns and verbs, often with the help of domain experts.
In the old world, you'd spend hours arguing about names and little details of relationships. I've never seen a team able to stick to DDD over the long run.
But now you can rip an LLM across your domain language and shape a system like clay.
Once you have this clear language, it's trivial to convert it to strict code, or system diagrams that give you a visual on the design.
The best part of LLM-assisted DDD is that the code (data model 'nouns' and function 'verbs') becomes a human-readable set of docs and constraints. The natural language helps fight "Claudevoice" and other problems of AI slop.
After years of wading through mud to make changes to systems, it's a thrill to learn something about your domain and watch your LLM thread your learning through the system.


















