Next Steps
Yesterday was my last day at Geolonia. I had been working at Geolonia since the beginning of 2021, making it just over 3 years. During my time there, I worked on many interesting projects, but I felt that it was time to make a change.
In these 3 years, a lot has changed. The world has seen the rise of LLM-based AI solutions appear and grow rapidly. There are many questions about its effectiveness and ethics, but it’s too late for that now. It’s time to really think about how AI technologies can be used to advance society.
I’ve been living in Japan for over 10 years, and recently I moved to a relatively rural area. The ratio of people in their 60s and 70s to the kids in school is completely out of sync. In Tokyo, it’s not as apparent, because it’s hard to get a feeling for what’s outside your social bubble. But in rural Japan, it’s around you constantly. This isn’t a new problem – population decline has been happening for decades. Government policies and politics have been focused on raising the birth rate, something I thought was possible, if implemented properly.
I don’t think that’s a viable solution anymore. I probably should have realized this a long time ago, but similarly to how it’s impossible to cost-cut your way to being a billionaire, the only way forward is to be more productive.
In software engineering, there’s the concept of the “10x engineer”. A software engineer who can do 10 times the output of a “regular” engineer. I’ve seen firsthand how a single passionate developer can use some insight, ingenuity, or previous experience that guides them to create solutions that would normally take a full team of engineers and managers to do. Just like a master carpenter intimately understands their tools, “10x engineers” leverage their expertise to create a significant advantage. The “10x engineer” concept can be applied to any endeavor.
In the coming years, AI will be table stakes. We’ve gone from an interesting chat demo, to the “Copilot” model of assisting humans, and we’re already seeing autonomous agents. Companies today are structured around hiring, managing, and aggregating the outputs of humans to create value. The large companies of today are working on replacing humans with AI agents, but I don’t think that’s enough. The companies of the future will be born in a world where a single human could be managing thousands, even millions of autonomous agents, and we’ll need new structures to achieve this. Imagine a startup with a lean human team orchestrating millions of AI agents – each specialized in a narrow skill – to tackle complex tasks. This radically different model calls for new management frameworks and new ways to measure success.
I’m currently planning on starting with geospatial solutions – I’m especially fond of geospatial data because it’s a concrete bridge between the physical world and the virtual world. For example, the Ministry of Land, Infrastructure, Transport and Tourism has been working on creating digital models of Japanese cities for the purposes of creating a digital twin, in Project Plateau. When developing, the rate of progress is governed by the speed of feedback. A high resolution digital twin will enable development in the physical world much like development in the virtual world.
If you’re interested, please contact me.
This marks the start of a new journey where I’ll focus on a world where AI is ubiquitous, and humans are wildly more productive than ever before in history. I’m excited for the future – if you are, too, let’s connect.