I help developers advance their tech careers and achieve their professional goals. With years of experience as a lead developer, I’ve guided both individuals and teams to reach their full potential. My passion lies in discovering and implementing methods that foster personal growth and technical expertise.
When people talk about AI and learning, the conversation usually focuses on AI answering questions—kids cheating on their maths homework or language students using AI as a translation service. We naturally gravitate towards the easiest path, and AI makes that path more accessible and tempting than ever.
The problem with taking the easiest path isn't just that you learn less. Learning is built on effort: wrestling with ideas, making mistakes, and connecting concepts for yourself. If we consistently replace that process with immediate answers, we risk building a habit of outsourcing our thinking instead of developing it.
This is often where the critics step in. Every new technology attracts predictions that it's making us less intelligent, less creative, or less capable. But I think that gives the technology too much agency. Tools don't determine behaviour – people do. AI can become a crutch if we let it, or a catalyst if we use it intentionally.
Personally, I've always believed that the best learning happens when it's driven by curiosity rather than obligation. It's the engagement between the learner and the material that really makes it sink in. To that effect, I’ve always leaned toward a self-directed approach—following curiosity first rather than a fixed syllabus, and building understanding by working through problems as they arise, ensuring that what I learn is practical and provable. Over time, that’s shaped how I learn almost everything, and it’s a pattern I’ve continued to return to.
That way of learning hasn’t changed with AI. I still start with curiosity, still learn by doing, still care about whether I can actually apply what I’m learning. What’s changed is the friction around it. AI doesn’t replace the learning process, it amplifies it. The uncertainty of where to start, what to learn next, or whether I’m missing something fundamental is still there, but it’s easier to navigate. It removes some of the noise around the process, so there’s more focus for actual learning.
Lately I’ve found myself thinking of it as AI-directed learning. It’s not necessarily a framework or a method (yet), just a way of describing how AI fits into my existing learning style. I’m still self-directed, still curiosity-led, but now I've included AI to help structure that curiosity a bit more, without taking it over.
In practice, AI-directed learning starts the same as self-directed learning, by asking myself: What do I want to learn? What do I want to achieve? What problem am I trying to solve? From there, things diverge. Rather than spending hours digging through Google searches and Stack Overflow threads, I use AI to map out the domain space. I can ask it to break down the core concepts involved, how the major components relate to each other, and where to best start my learning journey.
The important part is that I’m still doing the learning. If I don’t understand something, I can dig into it, ask for a different explanation, or work through examples until it clicks. AI isn’t replacing the process of working things out, it’s helping reduce the time spent being lost at the edges of it. It reduces the time spent working out the “unknown unknowns”, surfacing what I don’t yet know and making it easier to get unstuck when I inevitably hit a wall. It feels less like being taught and more like having a guide.
This idea of a guide shows up in fairly simple ways. I can ask for reading recommendations that actually match my current level of understanding, rather than just the most commonly cited resources. I can get concepts broken down into smaller parts, or ask for different explanations when something feels too dense. It can turn a set of notes into quick quizzes to check whether I’ve actually understood what I’ve read. I can even ask for project or capstone-style ideas that force me to apply what I’m learning, rather than just passively consuming information. For those who learn socially, it can also surface people worth following in a given space.
Recently, I’ve been planning on using this approach more intentionally to learn about LLMs themselves—topics like prompt engineering, RAG systems, and how models actually work under the hood. It feels like a good test case for this way of learning, where understanding is built through direction and iteration rather than passive consumption.
I’ll share what I learn as I go, and reflect on how well this idea of AI-directed learning holds up over time. I’d be interested to hear how others are approaching it as well—whether AI is acting more as a replacement for parts of the learning process, or something that helps amplify it instead.
