Notions

There is a particular kind of understanding that develops through making. A concept artist blocking out a scene, a procedural artist building the logic of a growing plant, or a designer adding a few lines to a sketch so everyone in the room can see the same idea. Each draws on years of practice, connecting the possibilities of a tool with the needs of a particular problem. Much of that knowledge is invisible in the finished work. For a long time, I understood my role through that relationship with the tools. Learning the software, developing a workflow, and finding ways around its limitations were central to how I became a designer. Increasingly, I see those skills as the setting in which a broader practice developed. Agentic AI introduces a different relationship with software. As these systems become more capable, more of the operation could be delegated: navigating an application, managing versions, coordinating tasks, or carrying an idea between different stages of production. Our attention can move towards defining the intention and guiding the outcome. The interface changes with it, supporting both the person directing the work and the agent carrying it out. For anyone who has spent years developing fluency in particular tools, that shift can feel bittersweet. There is satisfaction in knowing how to build something, and a sense of identity in the skills we have worked to acquire. Seeing parts of that process automated also opens possibilities for a broader practice. Work that once required several separate specialisations becomes more approachable, giving designers room to explore across disciplines. That possibility raises an important question about how we learn. Operating a tool has always involved decisions. Understanding why a camera angle works, how a material responds to heat, or which operation needs to happen first often comes through doing the work. Repetition develops perception alongside technique. We learn to recognise problems because we have encountered them, worked through them, and seen the consequences of our choices. When an automated system takes over those decisions, we need ways to preserve that understanding. If it selects a manufacturing setting without explaining the relevant conditions, the person approving the result may struggle to assess whether it is appropriate. Someone entering the profession may also miss the practical experience through which that knowledge once developed. Access to more capabilities brings value when we can understand their limits and recognise when deeper expertise is needed. This gives interface design a wider responsibility. As agents act on our behalf, the software needs to make clear what they can decide independently, where human input is required, and which assumptions are shaping the work. We need ways to inspect a decision, question it, and change its direction. Responsibility for an outcome requires enough visibility into the process to make an informed judgement. In the earlier posts, I explored the growing importance of direction, taste, and vision, guided by ethical judgement. Putting those qualities into practice requires its own set of skills. A clear brief establishes what matters, what can change, and what a successful outcome should achieve. Careful evaluation connects the results back to that intention. As tools make it easier to generate variations, choosing where to invest further attention becomes a larger part of the work. We also need to recognise when a convincing result is supported by sound reasoning, and when we lack the knowledge to assess it. In my own practice, pairing Houdini with generative tools has given me more room to explore worlds and moving images in service of a story. I can test atmospheres, compositions, and visual directions that might previously have exceeded the time available. I can stay with an idea longer before committing, or develop a version I would otherwise have left unexplored. That additional space matters to me. It makes a broader creative practice feel possible, with more freedom to move between disciplines and follow an idea beyond familiar methods. Its value grows with the understanding I bring to it, and with what I am willing to learn along the way. There is something of a renaissance in that possibility, provided curiosity extends to how the work is made. The knowledge developed through making remains part of this future. It shapes our eye, informs our judgement, and often leads us towards ideas we could not have reached through intention alone. As more execution becomes automated, we need to keep creating opportunities for that kind of learning. I want to spend some of the time these tools give back on closer observation, practical experimentation, and the ideas I have yet to try. That includes understanding what a system is deciding on my behalf and whether I know enough to stand behind the result. A broader practice asks us to keep developing the knowledge that gives our choices meaning.