I think open source is facing a sort of existential crisis. And I don’t think we’re the only ones. We’re just ahead of the curve.
In a world where intelligence is abundant, users can absolutely vibe code anything they want without our help. Why are we useful?

I would argue that vibe coding may even be safer! A developer builds software for many users, so if that software is compromised, many users are compromised too, making it a perfect target. (Now imagine you’re working in Bitcoin, and your software is open source)

On the other hand, our merchant—let’s call him Mike—is vibe coding his own software. One that doesn’t warrant this kind of attention from attackers due to the limited reward and the closed sourceness nature of it.
Also, Mike’s “mywares” have a tiny attack surface because they don’t include any superfluous features he’ll never use.
You say: “But he can’t maintain it over time!”
I reply: “Who care about maintenance?”
Mike is the only user. He doesn’t maintain it. He throws the whole project into the fire and raises a new one from the ashes with a new prompt.

And I tell you, I’ve been a developer for more than 20 years. I’m not being ironic. Most of my experience has been about keeping software maintainable, allowing it to grow steadily without collapsing under its own weight. The rules have changed. How can I serve Mike now? Mike is all that ultimately matters to BTCPay Server. No Mike to serve, no point in continuing.

Let’s say we embrace this and make BTCPay Server plugins easier for Mike to vibe code.

Couldn’t we push this further? Why does Mike even need plugins? Why not just drop BTCPay Server altogether? Couldn’t he make us obsolete? And I tell you… he can!

And he should feel no shame in doing this, no sense of loyalty. He needs to serve his customers, not us. This is capitalism. I would be ashamed if his loyalty stood in the way of serving his customers. I wouldn’t even be sad if he didn’t need us, we could retire the project and happily focus on new things, freeing up resources for better endeavors.

Will this be the future? Everybody surrounded by their own mywares. Believe it or not, I personally like that.

But there is a fundamental gap between AI and humans, and Mike’s agent can’t replace that: the will. In other words, the ability to pick a path from a sea of endless possibilities and commit to it. An agent doesn’t have a will. Everything it does can ultimately be traced back to human will.

Picking a path, exploring it, and committing to it takes effort. You have a limited number of shots every day. This is the new bottleneck.

With AI, you can certainly explore more paths, but you have to continuously decide which ones to drop and which ones to pursue. This is exhilarating at first, but eventually leads to burnout and exhaustion. Your attention is still limited, and your will decides where to focus it.

And this is a lesson Steve Jobs allegedly understood. He only kept the same clothes in his closet because he didn’t want to waste energy deciding what to wear.

Today, you’re in his shoes. Think about it: we’re facing the same challenge as the CEO of a trillion-dollar company.

People who lead companies have, by necessity, been trained to deal with this exact problem.

In Japan, when I go to a sushi restaurant, I ask for “omakase.” This means I let the chef decide which sushi I should eat. As a result, not only do I rely on his experience to choose what’s best, but I also save energy, allowing me to focus my attention on what matters more to me.

And for Mike, the problem is the same. Mike wants to focus on his business, and choosing BTCPay Server rather than building his own solution is a way to redirect his attention to where he can best serve his customers.

By running BTCPay Server, he relies on countless decisions made by its contributors, some of which matter to him and some of which don’t.

If it matches 90% of his needs, he can fill the remaining 10% with a plugin.

Everybody using BTCPay Server, and by extension their agents, builds on a shared, precisely defined conceptual model of payment processing.

Everybody building on top of those concepts reinforces a shared understanding that future agents can build upon.

In other words, would Mike be better served by a custom, vibe-coded payment processor built from scratch? Potentially. But his business isn’t limited to payment processing, and his customers would be better served if Mike focused his attention elsewhere.

And this is why, there is still a purpose.