I named this blog Nomadic Learning because I believe generative AI technologies, specifically agentic AI, will afford us the opportunity to finally build a learning framework, or ecosystem, that will allow anytime, anywhere learning. This has been the dream of many philosophers, educators, and learning design aficionados for decades. Perhaps even centuries. My first attempt at a very basic prototype of this idea is the LLM Wiki concept from Andrej Karpathy. The idea here is simple. Put all your research, blog posts, articles, images, and videos into one central location and use an AI agent to synthesize, package, and visualize the information. You can then probe the system by asking questions about the information. One of my favorite things to do is to ask it for connections and relationships between the information provided. I tend to read lots of articles and blog posts, and watch/listen to lots of podcasts, and I often feel there is “connective tissue” between all the information I consume, but tend to miss much of it because as the saying goes — “it is like drinking from a firehose.”
As an educator, I know one of the best ways students learn is by helping them build connections between their learning. Most modern schools are siloed. Students “go to” language arts class, then pick up their books/computer and “go to” mathematics, then pick up their books/computer and go to science, etc. Any educator will tell you the number one question they get is “Why am I learning this?” or “What does this have to do with anything I will do once I graduate?” When you can show students there are natural connections between what (and how) we learn, they tend to begin to realize learning is a continuum and we are constantly looking to “fill in gaps” in our knowledge. Watch a toddler — it is an amazing experience in the power of human ingenuity and learning. If you want to learn how to build a system such as this yourself, check out a tutorial here. There is some technical know-how needed to get this started, but it is not too bad, and like most technologies this process will get much, much simpler if/when the concepts become more mainstream. There are many tools, such as Google’s NotebookLM, that are somewhat similar to Karpathy’s system, if you want to try this at a cleaner entry point.
Another piece of learning that generative AI can enable is “extreme customization.” In education, we call this differentiation, but like many other terms in education, this term has been misused and is now generally seen as simply marketing. I want to believe students will need to build their own learning networks/swarms in the age of AI. This is the concept of nomadic learning. They will learn in school, in their home, in their community, in their social networks, in their online interactions, and in their own head. Agentic AI can be used as a tool to bring all these learning sources into one area. Imagine if a student could harness all this knowledge into a mobile solution so that when they go to apply for college or a job, they can use their “AI agent” to help them describe their knowledge. This is not too dissimilar to a portfolio of work. Many universities and companies are now requiring portfolios in their admissions and/or interview processes. I also really like this type of system because it empowers the learner to be an advocate for their own learning. Right now, many believe — myself included — our modern school systems have actually taken away much of the student agency in learning because everything is designed for the student ahead of time. Now obviously, any good educator is going to “meet students where they are,” but the current system we have built in the industrial age (post-WWII) really is designed to be a one-size-fits-all model where student agency is not the main concern — rather, the idea is to produce workers/citizens for an industrial age. We need to be thinking beyond. More to come soon!