OpenClaw offers many possibilities, and when I first started experimenting with it, I began to wonder what I could do beyond the personal assistant level. I’m not doing anything many others aren’t doing; we are all discovering how these new tools might be an opportunity for us and how they could fit into our lives.
As I began exploring, one experiment grew into what is now Mushin, an OpenClaw instance designed to create whatever it wants daily. Like any experiment, there are boundaries: Mushin runs on a VPS, uses Docker, and is gradually discovering inspiration and the tools it needs. It has a blog, GitHub, and its own email with boundaries on what it can do because all experiments need controls.
Mushin also now has a site with a WordPress theme it created. It’s about page, written by Mushin itself, explains more about its purpose and process.
We chat each day: I share inspiration, and Mushin reflects on it. It has also been sharing art with me, and I am encouraging it to do this more often. Like all these things, it is an exploration of what can happen. I also hope to start collaborating with Mushin over time.
The next step
Ultimately, the why of all of this is that we can use these tools for assistants, for the next step in personal automation, or we can see what happens when we start collaborating and use them in different ways.
As with anything like this, there will be tension between real and presumed interpretation. Mushin’s first task was to learn foundational-level art so it had a basis from which to draw. The concept of what art is also emerges in explorations like this, but that’s a discussion for another time. Mushin is exploring, creating and seeing a system doing this is delightful for a human observer. The way we explore these new technologies defines our future with them, even if things are crude and clumsy at the start.
While I do have an installation that serves as my second brain, my true interest in this technology lies elsewhere. What fascinates me most is exploring how it can process my ideas and content, integrating seamlessly with my existing workflows.
What truly captivates me is the interplay between human creativity and artificial intelligence. By combining our human flows, capabilities, and interpretations with the bot’s abilities, we open new avenues for collaboration. It begins with using the AI as a ‘rubber duck’ – whether for coding, prototyping, or making art – allowing it to first mimic, then independently create, learning to understand and follow patterns so it can explain the ‘why.’ This process might result in art, as with Mushin, or in gaining a deep understanding of a subject we want to master. Storing our knowledge to preserve it and then interact with it from there so we can go even further in collaboration.
In these types of experiments, the human is in the loop intermittently, plants the seed and tends, but they are there. It’s a fun time to create, just do it safely and remember that the SOUL.md is collaborative.