Seven Languages, Revisited

I've been thinking it's time for me to revisit Seven Languages in Seven Weeks. It was one of those books that sunk its claws into me. While I didn't end up deploying those languages in a professional setting, I remember the focus on the what and why for each language's capabilities.

Part of my interest in revisiting this book is to backstop the intellectual ossification that agentic first development is creating. I'm also curious if I still find as much joy in the quirks these languages bring to the table. At base though I want to use it as an exercise to shift my thinking around code generation. So far I've used agents to create code using the same platforms I would use if I were hand coding a solution. I'm sticking with what I know, both because I have experience with it working, and because I will have the ability to audit and debug any output.

If I'm going to continue down the rabbit hole of agentic flows I need to have a clear answer to how and why I might choose a platform or language. I need to know if it makes sense to cede that thinking to an agent, or if I want to maintain control but operate outside of my comfort zone because the language better suits the goal.

I don't think I'm going to be deploying anything with Ruby on Rails instead of sticking with TypeScript and its universe of frameworks. I might throw some Scala at the wall, but I'll have to be sure there's really an advantage over Go. Really it's less about the specific languages; it's about understanding tradeoffs.

The shift to AI generated code has happened so quickly that I haven't really spent the time I need rethinking my approach to aligning the possible with the profitable. Much in the same way as my evolution through Design Patterns starting with confusion, falling into obsessive application, and escaping into the understanding that it was never about the patterns - it was about thinking; I suspect I'll pass through yet another door when it comes to selecting a stack.

With Agents now porting frameworks faster than I can say why, we may well end up with agentic first languages where the syntactic sugar isn't aimed at reducing typing but at anchoring purpose. Maybe instead of camelCase we'll have contextualCase bringing back the Hungarian Notation of my youth.

Seven Languages fits squarely into my goal for this year to rethink my core approach to problem solving. It uses the joy of crafting code to give me a playground where I can reimagine what I need to be an effective software engineer.