Part II Kanvana is born the Swim Lane Complexity Grows

 

Read part one here Part I the AI Hybrid Build Release 1 , heads up and shout out to my AI co-auther's Gemini and ChatGPT for this Part II in my now series on AI Builder Builds.

AI is not replacing software engineering today. But it is changing how I build, how fast I release, and where I spend my time.

I just released v1.4.0 of Kanvana (Kanban + Nirvana) my Personal Kanban and soon to be AI-Agent board ( https://github.com/mdiener21/kanvana), and this one is a big one: Swim Lanes, Sub-Tasks, Relations. But honestly, this release is also part of something bigger for me.

I’m actively pushing to learn the real edge of hybrid development with AI tools, agents, and models: where the actual productivity gains are today, where the limits still hit hard, and where architecture, security, quality, and cost decisions can spiral fast if you are not careful.

There is a lot of hype right now. And yes, we are clearly on the path toward far more autonomous software delivery. No question. But we are also definitely not there yet, my best guess is at this pace we are 1 to 2 years away that being very conservative estimate.

This is exactly why I like building in public with projects like this. It gives me a practical playground to test what works, what breaks, what still needs strong engineering judgment, and what should absolutely not be left unchecked.

I've now released v1.4.0 of Kanvana the Personal Kanban now with Swim Lanes, Sub-Tasks, Relations

The complexity level rises and so does the challenge to get things "correct". I use swim lanes to add a second dimension to my board so I can instantly organize and manage work by team (label-group), label, or workstream across the same workflow without losing clarity. The new feature:

🏊 Swim Lanes

• Group by label, label group, or priority

• Drag tasks across columns, lanes, or both

• Collapse cells, rows, and columns independently

• Add tasks directly into any swim lane cell

• Keep lanes compact while still supporting drag-and-drop into Done

• Sticky headers, mobile-friendly layout, persisted board state

Also added

• Quick swim lane toggle

• Drag-and-drop lane ordering

• AI Agent Ops starter template

• Stronger test architecture with Vitest, Testing Library, MSW, and Playwright

• Better import safety checks

• Security hardening and stricter CSP

• Better spec and contributor workflow

For me, this is the real work right now: not blindly chasing AI hype, but figuring out how to combine AI speed with real engineering discipline.

That’s where the value is. And that’s where I think the future gets built.

Take a look, try it out, and tell me what you think. Star the Github Repository Where are you seeing real gains with AI in development today, and where is it still falling short?

#buildinpublic #ai #softwareengineering #productivity #kanban #opensource #webdevelopment #engineering #testing #agents

Be a builder, look around you opportunities are everywhere go build it!

Whats next ? I think I will now dabble towards offline syncing across devices with a simple backend with pocketbase.io a Go powered BaaS (Backend As A Service) it uses SQlite. This will be of course optional for you to activate, enjoy the next post.