Week 7
Through my research, I've explored how recommendation algorithms shape behavior, voice, and taste. I've done that by looking at social media closely, writing a taste inventory, making a zine, coding p5.js sketches, building my Side Eye extension, and collecting images from artists and films.
I've noticed that the loop isn't really about other people anymore. Bhandari and Bimo found that TikTok users mostly interact with the algorithm itself. They like and comment to train their feed, not to connect with anyone. Who I am online gets shaped by a system more than by a community.
Building Side Eye helped me realize how much data misses. It can track my time, visits, and scrolling, but not how any of it felt. A great hour and a wasted hour look the same in the numbers. That gap is where my thesis lives.
I also discovered that artists were asking these questions long before algorithms. Cindy Sherman and Lynn Hershman Leeson were both working with performed selves and constructed identities in the 1970s.
What surprised me was how much I want to be "gotten." People in the study describe the moment TikTok understood them almost like a relationship, and I recognized that feeling in myself. I started this research sure that I knew my own taste apart from the feed. Now I'm not sure I can tell where mine ends and the algorithm's begins.
Week 5
So far I've been looking at the loop from a few angles. The readings, the artists, and my own scrolling all point to the same thing. The loop is not one thing. It is made of smaller parts like behavior, voice, and taste, and each part can pull me in a good, bad, or neutral direction.
What I've noticed:
- When I collected screenshots for my poster, the same faces, gestures, and trends kept repeating. I didn't search for any of it. The feed kept handing it back to me until it felt normal.
- Bhandari and Bimo describe the "algorithmized self." On TikTok, people mostly interact with the algorithm, not with other people. We shape ourselves by reacting to versions of ourselves that the feed reflects back.
- Artists like Hito Steyerl and Zach Blas respond by hiding from the system. Nam June Paik and Lauren McCarthy respond by watching it closely. That range helped me see that resistance is not the only option.
What I've realized:
- Crit helped me separate my terms. Influence is the force. The loop is the dynamic, a continuous action that keeps repeating. Thriving is my objective.
- The loop is composed of parts like behavior, voice, and taste. Because the action repeats, it slowly shapes each one. What I watch changes how I act, and over time it changes my voice and my taste too.
- I don't want my thesis to be anti-algorithm. My position is closer to pro. The loop is always going to be there, so the more useful question is how to use it without losing yourself.
What surprised me:
- The TikTok design pushes people to create, but most participants in the study saw themselves as viewers. People don't always do what the platform wants.
- Users knew their data was being collected and saw it as a fair trade for a good feed. They only felt uneasy when it became "too accurate" and started showing people who looked and acted like them.
- People use likes and comments less to connect and more to steer the algorithm. That made me realize we already negotiate with the loop every day, often without thinking about it.
Where I'm going next:
- Looking closer at the continuous action and breaking it apart into what it is made of.
- Mapping what is good, bad, or neutral about each part and what it influences. I want to ask what is positive inside the good and the bad, and build a scale instead of treating it as one or the other.
- Defining what thriving means for me. Right now it is still thrive = ?, and I want to fill that in through my own experience.
- Exploring the pros and looking at all the possibilities through case studies.
- Backing up why my position is pro. That means looking at the full range, from people who resist the loop to people who fully embrace it, and showing where thriving fits.
Week 4
So far I've explored how recommendation systems shape identity, starting with my own TikTok feed. My first print, A Cycle of Repetition, arranged screenshots into a loop, and seeing it laid out flat made the feed look like one person repeating. Bhandari and Bimo's idea of the algorithmized self gave me language for this, since they argue that people engage less with each other and more with a filtered version of themselves. I collected precedents from Nosedive and Nam June Paik to Hito Steyerl and Zach Blas, and what surprised me most was that almost all of them are about escape. My life doesn't work that way, because I can't leave these platforms when I work on them. The users in the study didn't leave either. They used likes and comments to steer the algorithm, which shifted my question from how to escape the loop to how to thrive in it without disappearing. Writing truisms this week also showed me that words I had treated as neutral, like the loop and thriving, already carry positions, which tells me I'm still defining the phenomenon rather than describing it.
Week 3
Through my research, I've explored how recommendation systems shape the way designers see their own taste and work. Bhandari and Bimo's idea of the "algorithmized self" gave me language for it. On TikTok, the self forms less through social connection and more through repeated engagement with a system that reflects us back to ourselves.
I noticed this in my own process. When I wrote my 10 taste words before scrolling, the list read like a brand description. That made me realize the way I describe my own voice may already be shaped by platforms.
What surprised me most was that accuracy felt worse than error. When a system described my work in a way that felt right, it was more unsettling than when it got me wrong. It made me question how specific my voice really is.
I've also realized that escaping the system isn't the answer. Hiding from it can mean disappearing from people too. The more useful question is how to stay inside the loop without letting it decide who I am.
If you have real results from any of the prototypes, like your Blind Sort score or the Lens labels, add one line with that detail. A specific finding will make this stronger.
Week 2
What came out of this week was a definition more than a finding, which is that the loop isn't social connection but repeated engagement with a system, and the self takes shape through that repetition. Having that sentence changed where I look. I've been treating beauty as the clearest case, since the standard arrives through a system tuned to me specifically, and I can't tell anymore whether the algorithm produced the standard or just found one already there and narrowed the room around it. Spotify and AI are still unresolved but sharper, since Spotify recommends off what I already did while AI takes instruction and lets me talk back, and I can't tell whether that's more agency or just a tighter attachment. The thing that actually surprised me is smaller than all of this. I went looking for people caught in loops and almost nobody describes themselves that way, they like what they get, so if the position is thriving without disappearing then most people already believe they're thriving and the word I'm on the hook to define is disappearing. I don't have it yet, which is why the next step is the phenomenon and not the application.
Week 1
What came out of this week was a definition more than a finding, which is that the loop isn't social connection but repeated engagement with a system, and the self takes shape through that repetition. Having that sentence changed where I look i've been treating beauty as the clearest case, since the standard arrives through a system tuned to me specifically, and I can't tell anymore whether the algorithm produced the standard or just found one that was already there and narrowed the room around it. The p5.js sketch got at the same problem from the designer's side, because I had to write every rule it ran on and the output only looked inevitable to me because I'm the one who typed those numbers in, which is about as close as I've come to seeing my own position in this. Spotify and AI are still unresolved but sharper now, since Spotify recommends off what I already did while AI takes instruction and lets me talk back, and I genuinely can't tell whether that's more agency or just a tighter attachment. The thing that actually surprised me is smaller than any of this. I went looking for people caught in loops and almost nobody describes themselves that way, they like what they get, so if the position is thriving without disappearing then most people already believe they're thriving and the word I'm really on the hook to define is disappearing. I don't have it yet, which is why the next step is the phenomenon and not the application.
Weel 6: Provotype
This week I explored the loop through making instead of reading. I made a parody ad for a fake product called TASTE™ and an interactive p5.js sketch called The Loop. In the sketch, the algorithm ranks you, gets bored of you, forgets you, and judges your dance moves.
The biggest thing I noticed is that there is no neutral move inside the loop. When I built the skip button, I wanted it to feel like resistance. But the system still learned from it. Skipping became data too. That changed how I think about my definition. You can't step outside the loop by refusing it. Refusal is just another input.
I also realized that sameness is easier to show than to explain. In the sketch, the more predictable you get, the more the screen turns beige and every post becomes the same post. Seeing it happen made the idea clearer than any sentence I've written about it.
What surprised me most was humor. I expected the work to feel heavy, since it's about surveillance and losing your taste. But making it funny made it easier to sit with. A fake pharma ad or a stamp that says "CERTIFIED NPC" lets people laugh first and then recognize themselves. I think humor might be one way to thrive in the loop without disappearing into it.
I'm still not sure where the line is between the loop shaping me and me shaping it. That feels like the question to keep pushing.
Open reference
