I have a confession: for most of my life I’ve thought open source was stupid.

You see, I grew up struggling economically. When I discovered in high school that I enjoyed working on computers and that it paid well, I saw it as my way out of poverty. Code was my ticket to a different life.

I started using Linux in 1996, Slackware 3.0. At the time, most households only had a single phone line. It could be used to chat on the phone, or it could be used to dial up a BBS or the early internet, but not both. If you were fortunate enough to have two computers in the house, only one could use the internet at a time. Even if you had the newfangled broadband cable internet that didn’t use your phone line and was way faster.

I set up my Slackware box as a NAT router to share internet across the two computers we had (one my father had purchased for keeping track of taxes and whatnot, and one I had eventually built myself for playing games from money I’d earned doing odd jobs). Nowadays you buy a $99 router and you’re done. Back then, what I was doing was state of the art and could only be accomplished using IPTABLES on Linux for a reasonable price.

While I enjoyed using open source (kinda, what I actually enjoyed is that it didn’t cost any money) I never understood the business model. Do a bunch of work for free, give it away, then charge for doing more work? The economics never made sense to me.

And here’s the part where I risk offending people, so let me be careful: I have enormous respect for the idealism behind open source. People contribute because they believe in something larger than themselves, a commons of knowledge, freely shared. That’s genuinely noble.

But I could never square the idealism with the complaints. People give away their labor freely, then express frustration that they aren’t getting paid for it. And I’d think: what did you expect? You can’t give something away and also expect to be compensated for it. Those are contradictory desires. The business model assumes someone will pay out of gratitude or goodwill rather than necessity, and that’s just not how economics works.

This never made sense to me. That was a huge part of why I joined Microsoft after graduating college.

Fast forward to today, and there are only two times when I’ve contributed to open source and felt good about it.

The first time is when I went on Adderall for ADHD and suddenly programming felt easy compared to the struggle it had always been. It was easier for me to give away my work because it didn’t feel like I was doing something difficult, something that cost me real effort. The sacrifice was gone.

The second time is today, with the advent of AI coding. I’ve open sourced the majority of my hobby projects I made in 2025 because I didn’t suffer creating them.

And that’s when it clicked.

Between AI coding and 3D printing’s unbelievable free library of things, for the first time in my life it looks like a fully open source future might be possible. It might even be necessary.

Human beings have two tricks of economic value: muscles and brains. We mechanized muscles in the early 20th century, and it led to the Great Depression. This time we’re automating brains. We’re out of tricks. Just like open source of the past, I don’t see how we’ll be able to have the same property rights in a world where nobody can pay rent because they can’t do anything of economic value through no fault of their own.

The most common response to this is UBI. I don’t think that solves it, here’s why.

Proponents of UBI often don’t have an answer to the following question: “What stops a landlord from setting rent at or near UBI and capturing most of it?”

The only plausible answer to that is: there’s a lot of space in the United States, and if people can’t afford to live where they want to (Manhattan for example) then they’ll simply move to where they can afford. That’s the strongest steelman I can make for the UBI argument.

We’re already seeing rents flatten across the country as algorithmic pricing and institutional investors target formerly affordable markets. The gap between rural Pennsylvania and Manhattan has compressed dramatically in my lifetime.

But what if rent didn’t matter? What if the things we need to survive weren’t scarce at all?

Picture a small homestead, maybe five acres, maybe fifty, it doesn’t really matter. There’s a house, a garden, and a workshop. The workshop has 3D printers, CNC machines, and robots that know how to use them. Solar panels on the roof. A good internet connection. A few families, or maybe just one, or maybe a dozen friends who actually like each other, the people you choose to live alongside.

Robots tend the garden. They plant the seeds in spring, pull the weeds in summer, harvest the crops in fall. They bring the tomatoes inside, wash them, slice them, and prepare dinner while you’re out on a walk. They don’t get tired. They don’t resent the work. They just do it.

When Christmas comes, your kid wants a toy she saw online, something a designer in Tokyo dreamed up and shared with the world. You download the files, and by morning it’s sitting under the tree. When the tractor breaks, the workshop prints the part. When someone gets sick, an AI helps diagnose the problem and the robot pharmacy compounds what’s needed. Nobody had to clock in. Nobody had to trade their hours for someone else’s dollars to make any of it happen.

Everyone’s writing the dystopia. I want to talk about what happens if this goes right.

Here’s what I’ve learned: if the work is easy, it’s easy to give away. I never understood open source because I was thinking about it through the lens of scarcity. My labor was valuable precisely because it was hard. Giving it away felt like giving away a piece of myself. But when AI made coding feel effortless, sharing became the default. I wasn’t sacrificing anything. I was just… contributing.

Now scale that up. In a world where robots handle the hard parts (the farming, the manufacturing, the maintenance) sharing becomes the natural state of things. Ideas flow freely because nobody’s been put out by contributing them. You’re not protecting your competitive advantage when there’s nothing to compete over.

This is the fully open source future. Not just code, but everything. Designs for furniture, medicine, machines. Recipes for food and fuel. The knowledge of how to live well, shared freely, because sharing costs nothing when creation is effortless.

And here’s what that unlocks: human potential. Right now, most people spend their lives in coerced labor, not because they love the work, but because they need the money to pay for the expert services of doctors, lawyers, mechanics, and everyone else whose knowledge is kept scarce. Remove that coercion, and people are free to follow their actual interests. To create for the joy of it. To contribute because they want to, not because they have to.

I don’t know if we’ll ever get warp drive. But this future? The homestead with robots and a good internet connection and a community of people who chose each other? I can see it from here. And for the first time in my life, I understand why someone would give their work away for free.

I’ve been a Kanban believer for years. Not the buzzword version—the real thing. The system that transforms chaotic work into visible, manageable flow. The methodology that helps teams stop starting and start finishing.

But here’s the problem: most “Kanban” tools aren’t really Kanban at all. They’re glorified task lists with columns.

The Gap That Bothered Me

If you’ve ever used Trello, Asana, or any of the dozens of board-style apps out there, you might have walked away thinking Kanban is just… accounting. Moving cards around. Busy work that pulls you away from actual work.

I get it. I’ve felt that frustration too.

Where were the WIP limits that force you to finish before starting something new? Where was the real-time collaboration that didn’t require a page refresh to see what your teammate just did? Why did every tool feel like it was designed by someone who’d never actually used it during a stressful sprint?

So I spent the last few months building what I wished existed.

Meet Taskleef

Taskleef.com is a Kanban board that takes the methodology seriously—without taking itself too seriously.

Work is stressful enough. Your tools shouldn’t add to that. Taskleef wraps powerful features in a calm, focused design meant to help you find flow, not fight your software.

What Makes It Different

Real-time by default. When your teammate moves a card or adds a comment, you see it instantly. No refresh dance. No “someone else is editing” warnings. Just seamless collaboration powered by SignalR.

AI that actually does things. This isn’t another chatbot that just talks at you. The AI assistant in Taskleef can create subtasks, reorganize your cards, suggest tags, and move work through your pipeline based on conversation. And you choose your provider—OpenAI, Claude, or Gemini—and swap between them whenever you want. Bring your own API key, keep your costs transparent.

Subtasks that grow up. Sometimes a subtask turns out to be bigger than you thought. In Taskleef, you can promote it to a full card on the board while keeping its connection to the parent. Context preserved, complexity managed.

Smart recurring tasks. Not just “every Friday”—though we do that. Taskleef also supports completion-based recurrence: “4 hours after I finish this.” Two distinct models for two different workflows.

Real WIP limits. Actual constraints on work-in-progress, per column. The feature that makes Kanban Kanban, finally implemented.

Everything Else You’d Expect (And Some You Wouldn’t)

• Quick-add capture with inline tag parsing ([tagname])
• Unlimited nested subtasks
• Due dates and defer dates (for when something shouldn’t surface until later)
• Sub-column states—mark cards as blocked or done without moving them
• GTD-style inbox for rapid capture
• Comments, activity logs, and card subscriptions
• Project grouping for organizing multiple boards
• List view when you need a different perspective

Watch Your AI Work

Here’s something I haven’t seen anywhere else: you can watch AI make progress on your board in real-time.

Most AI integrations are black boxes. You ask for help, wait, and eventually get a result. Maybe it worked. Maybe it didn’t. You have no idea what happened in between.

Taskleef is different. Because the board updates in real-time for everyone—teammates and AI agents—you can literally sit back and watch as an AI breaks down a complex card into subtasks, reorganizes your backlog, or moves completed work through your pipeline.

It’s not magic happening behind a curtain. It’s visible work.

Ask the built-in assistant to break down a feature into implementation steps, and you’ll see the subtasks appear one by one. Connect an external agent through the API, and watch it triage your inbox while you grab coffee. There’s something deeply satisfying about seeing work get organized without lifting a finger—and something reassuring about being able to see exactly what’s happening.

This isn’t just a gimmick. Visibility builds trust. When you can observe your AI’s decisions as they happen, you catch mistakes early, understand its reasoning, and stay in control. The same principle that makes Kanban powerful for teams—making work visible—makes AI assistants more useful too.

Built for Developers and AI Agents

Taskleef isn’t a walled garden. If you want to build on top of it, integrate it into your workflow, or let your AI agents manage your tasks—you can.

Fully documented API. Every feature in Taskleef is accessible programmatically. Create cards from scripts, sync with other tools, or build your own integrations. Full documentation lives at taskleef.com/docs.

Open source examples. I’ve published the tools I actually use while developing Taskleef, including a simple CLI for quick task capture and a REPL loop for interactive board management. Grab them, fork them, or use them as a starting point for your own integrations: github.com/Xatter/taskleef

AI-agent ready. With a clean API and structured responses, Taskleef plays nicely with AI coding assistants, automation scripts, and agent workflows. Point Claude, GPT, or any agentic system at your board and give it work to do. The real-time sync means you’re never wondering what state things are in—you’re watching it happen.

Built for Focus

Every design decision in Taskleef asked the same question: does this help someone get into flow, or does it interrupt them?

The result is a tool that stays out of your way when you’re working and surfaces exactly what you need when you need it. No visual clutter. No notification overload. No enterprise complexity masquerading as “features.”

Just a clean board, your work made visible, and the confidence that you’re actually practicing Kanban—not just shuffling cards.

Try It

Taskleef is live at taskleef.com.

If you’ve bounced between capture apps, Kanban boards, and team tools looking for something that combines all three without the bloat—this is what I built it for.

I’d love to hear what you think.

Questions, feedback, or feature requests? Hit me up on Threads.