High Agency in the Agentic Era

This past month, I’ve been working super hard to accelerate new feature releases. One thing I’m truly grateful for is the opportunity to build on Nostr and Lightning, experiment with new tools, and grow alongside them.

Voice-driven workflows and agentic automation have expanded what I’m able to build and contribute. They are not merely productivity tools—they can change who gets to participate, and how much they can do.

In truth, learning to rethink development around agentic tools is a bit of a clusterfuck. Things that used to be done one way can now be done in completely different—and often far more efficient—ways.

It’s a good idea not to listen too closely to the naysayers. Start with a fresh, clean slate and throw out your assumptions.

Understanding how to use AI harnesses and agentic tools in the ecosystem is still a steep climb. Many of these tools are frontier technology; there is no established playbook. High agency is required to figure out what they actually mean in practice and how to apply them effectively.

It is real R&D. In my experience, making any meaningful headway requires being self-driven rather than directed by an authority. It requires an innate desire to learn, experiment, and understand.

People who chase labels, checkmarks, or titles will likely find this transition especially difficult over the next few years. Every day and every week over the past three years has brought another breakthrough. It feels like drinking from multiple fire hoses that constantly change direction.

What I’m delivering now certainly didn’t come out of nowhere. It is the product of thousands of hours of silent work in the background: a ton of trial and error, plenty of missteps, and a lot of lessons learned.

But here’s the thing: R&D doesn’t pay in the early stages. Eventually, though, the unpaid hard work pays off—in spades.

At the end of the day, nobody can take away what I’ve learned and experienced. To me, that’s the real meaning of being self-sovereign.

Engineers solving problems with agentic tools are now making 200x performance gains compared with those who haven’t yet learned to harness them. And this is just the beginning.see this note

Recursive self-improvement is compounding exponentially. Those who decide to wait a little longer until things are “stable,” or until someone is ready to hand-hold them through it, may find that it’s too late to catch up.

Seth Godin recently put it this way:

“This is the biggest change in our world since electricity.

Either you work for AI, or the AI works for you.

There were workers in 1890 who said they would wait for electricity to go away. That did not work very well. AI is like this, but much, much faster. If you work for an AI, it’s not going to treat you very well.”

Watch the clip here.

Thanks for taking a moment to read my thoughts. I’m quite sure there is much more to come, and it’s going to be an exciting—but chaotic—decade.

Cheers.