Week 10: Reflections on Reflections

This week was a different kind of work. No new builds, no prototypes, no late-night debugging. The class was structured entirely around stepping back and looking at the semester as a whole, which sounds easier than it is. Looking at ten weeks of work and trying to make sense of what actually happened is uncomfortable in a way that building things is not, because you can see everything you did not finish as clearly as everything you did.

The Experience

The session opened with Leo presenting on his practice-led PhD, and the thing that surprised me most was finding out he had a background in cinematography, videography, and set design. That reframed everything. Here is someone who came from the same craft foundation I have been building at The Next Something, and he ended up doing a four-year doctorate on generative computer art and Lacanian psychoanalysis. The work looked completely different by the end of the research than it did at the beginning. Different technologies, different questions, different outputs. What stayed consistent was the method: start from not knowing rather than from a fixed identity. The problem teaches. That framing has been sitting with me since the session. I want to have a conversation with him specifically about the filmmaking side before the final presentation, because I think there is more there.

The group activity was with largely the same people I had done the crits with, so the conversation had a familiar rhythm. What made it useful was hearing where everyone actually was. A majority of the group were still in their experimenting phase, still iterating, still not fully resolved on a direction. That was genuinely reassuring to hear out loud. I had been carrying around a quiet assumption that I was behind, and it turned out that was not really true.

The peer review surfaced the main note I needed: more structure. The content is there across my posts, but the architecture around it needs work. I want that note to inform how I write from here, including this post.

Reflection

Looking back across the full semester from Week 1 to now is a strange thing to do while you are still inside it. The idea started wide open. I came in wanting to make a speculative product launch film for a wearable AI device built for creatives, set in a future Aotearoa. That is still what I am making. But the version of it I am building now looks nothing like the version I described in Week 1.

What changed: I started with a single direction and ended up with three, and then narrowed back down through two crits. The news publication idea, a fake 2030 media site with articles, podcast clips, and radio content set in a world where AI had reshaped journalism, did not survive the crit. The honest reason was that I had no real investment in it as a practice. I could not answer the question of why I cared about it enough to build it well. Letting it go was the right call.

The two directions I have now are the speculative launch film and the Muse interactive game, and both have genuine development behind them. V1 of Muse was too short and mechanically broken. V2 has a four-file architecture, three days of narrative, five endings mapped to a leanScore system, and UI corruption that shifts the visual environment as the player's creative autonomy transfers to MUSE. That is a meaningful change from where it started. The film still has an unsolved compositing problem, which I am not going to gloss over. Getting a wearable headset display to look convincing on camera is still the central technical wall. But the concept is clear, the references are solid, and the direction is one I genuinely want to pursue.

Theory

Leo's methodology framing connected to something Schon wrote about the difference between reflection-in-action and reflection-on-action (Schon, 1983). When I was building Muse V2 or researching wearable product references, the thinking was immediate and embedded in the making. This week the thinking was slower and more uncomfortable, because stepping back means you can see the gaps clearly. Both modes are necessary. This week was entirely the second kind, and I think the discomfort of it is actually the point.

The IRC framework I have been using since Week 4 has been useful for structuring that kind of reflection. Gibbs gave me the emotional processing in the early weeks. IRC gave me the forward-planning orientation I needed once the experiments got more complex. The crit sessions added a layer that neither framework gives you on its own: external pressure from people who do not already know what you were trying to do and have to respond to what is actually in front of them.

Skills

One thing this week clarified for me is how much I have actually built in terms of capability. I came into this semester as someone who makes commercial films. I am leaving it as someone who also knows how to set up and run Claude Code to build a functioning interactive web experience from scratch, use AI image generation as a genuine part of a storyboarding workflow rather than just as a shortcut, generate video with AI tools that I can actually use to supplement production at The Next Something, and use Claude as a prompt intermediary to get significantly better results from image generation tools than prompting directly. None of those existed in my toolkit at the start of the semester. All of them have immediate applications beyond this course.

That last one in particular is something I have already started using on personal projects. The workflow of describing what I want in natural language, passing it through Claude to generate a refined and detailed prompt, and then sending that to a generation tool produces noticeably better output than going direct. It is a small shift in process and a significant shift in results.

Preparation

The final blog is due Thursday. The final presentation is Week 12. Here is what I am actually planning to cover in that 10-15 minutes.

I want to open with who I am and what I came in wanting to do. The San Francisco trip, the doomers, the question of what it means to design an AI that assists rather than replaces creative work. That is the problem that generated everything else this semester.

From there I want to walk through the experiments in order: the three ideas, the first crit, what survived, what did not, why the news publication died, how Muse evolved from V1 to V2, what the film direction looks like now and what is still unresolved in it.

The skills section needs to be concrete. Not just "I learned things" but the specific tools and what I can now do with them. Claude Code, AI image and video generation, After Effects orientation, the prompt intermediary workflow.

Then I want to be honest about the failures: V1 was too shallow, the compositing problem is still unsolved, there are things I would have tested earlier if I had the semester again.

The final section needs to show where this goes. The speculative film is my strongest Capstone direction. Muse is a close second, more complete as a prototype but a different kind of work. The news publication is a third alternative that I let go of for the right reasons but could be picked up again in a different form. I need to be clear about which one I am pointing at and why.

DES300 is a course I completed last semester rather than this one, so the direct link that rubric asks for does not apply in the same way. What I can speak to honestly is where this work points in terms of 304: pushing speculative filmmaking into more experimental territory, understanding how to build interactive experiences that put audiences inside a question rather than explaining it at them, and coming in with a set of tools I did not have before.

Before Thursday I want to talk to Jed or Leo once more. Not to get permission for a direction, but to pressure-test whether what I have is enough to make a strong case in twelve minutes. I think it is. I am in a good position. I have two solid directions, real development behind both of them, and a clearer sense of what I am doing than I have had at any point this semester.

Schon, D. A. (1983). The reflective practitioner: How professionals think in action. Basic Books.