• Welcome to droqen's forum-shaped notebook. Please login.

Have Algorithms Destroyed Personal Taste? On fashion

Started by droqen, January 23, 2022, 11:48:29 PM

Previous topic - Next topic


January 23, 2022, 11:48:29 PM Last Edit: January 23, 2022, 11:51:24 PM by droqen

Last year, and now this year again, I've been thinking about videogames as fashion, not in terms of it being something that you wear and that you can use to express yourself, but in that it is something with trends, not wholly predictable but certainly recognizable, something that can be theorized about, responded to, and seen.

I liked this quote a lot and I'd like to think about it for a while:

QuoteWe might say that "taste" is the abstract, moralized knowledge, while "style" is its visual expression.

Fashion makes taste easily visible as style, in part because its distinctions between color or cut in clothing are so specific and yet so random ("rules which we don't even know").


Section Summaries.

The Seeing Robot
There's an app that gives you a % score for an outfit. It doesn't tell you how it comes to the decision, really, but for example it might judge an all black outfit at 73%, and an all grey one at 27%.

Theories of Taste
Taste is "knowledge through pleasure".
"We don't calculate or measure if something is tasteful to us; we simply feel it."
"This is principally founded on surprise."

But Fashion Is Already Arbitrary
Is its own summary

That Scene from The Devil Wears Prada
Claims that things are in fashion because someone decided it was in fashion.

droqen's aside: my own ruminations on this topic: when I see what games are fashion, it's a fascinating activity to predict and theorize how the culture of gaming will react to a variation of what we have now, or to a revival of something that once was. someone has to make a new game to push 'fashion' forward, but it's not a choice. it's a tactical move, bound up in limitations as much as in creative freedoms.



Data-Based Fashion
What if instead of someone deciding something is in fashion, data decided what was in fashion? The author isn't sure which one is better, or worse.

droqen's aside: data can't decide what's in fashion, and in a way data is already one of the tools in use today for exploring new avenues of fashion. data is just information. someone needs to process the information. here the author could be talking about one of two things, or both:

1. what if fashion designers used more data when making their decisions? (this is already happening, i am sure)

2. what if artificial intelligence were involved in decisionmaking? (again, ai is certainly involved at some level)


"Collapsing Dominant"
The author talks about how we are "in the midst of [a] shift in taste," from the mass-media television-driven mode to the current, driven by "Instagram likes, Twitter hashtags," etc.

"Do I understand the new or am I stuck in the old?"

The Death of Svpply
Svpply was a fashion social network with human-designed collections/feeds of fashion items.
As it grew, it found it had to rely less on human curation and more on machines.
And then it died.

"If everyone's editing Vogue, it wouldn't be Vogue."

Human- v. Machine-Curation
Machines can replicate style, but they don't have taste yet.
A friend who recommends a blue shirt will have a greater depth of interaction and reasoning than will a targeted ad. (e.g. Ok, you think I'll like this blue shirt. Why blue? What will it mean?)
The author wonders if they're just falling behind the times. Maybe the next generation will be fine with machine curation. Or the next.

Taste Optimization
Machine choices are bland and uninspired because they're driven by algorithms tuned to get attention, not to express an underlying taste. "Same same but different."

Machine-Generated Content
We now have creators creating things based on this non-existent algorithmic taste:
They do what the numbers say works.
They do what gets attention, rather than doing work to express their underlying taste.

(droqen: can't it be both? i can walk to the grocery store along my favourite path and still get there on time.)


Style in the Age of Digital Reproduction
Technology enables mass recreation of things once considered unique, original. Not just a photograph of a painting - AI can replicate the style of an artist. 3d models can stand in for real people online. All this, "cheaply and infinitely."

One Tenet of Algorithmic Culture
The author describes "the Generic Style", a style which falls naturally out of a well-explored system. The way people act in a virtual space is informed by the space itself, resulting in this "Generic Style". However, this can be a comforting thing, and it even exists in real life. Offline.

Content Luddism (Ethically Sourced Culture)
"I could abstain from algorithmic culture", writes the author.
But "effective" stylistic quirks will always be captured and become part of the "Generic Style".
It's hard to abstain.

Internet piracy a decade ago felt more like a human experience than the Netflixes of today.

Hipster Platforms & Platform Hipsters
Now there exist services and platforms which present something more human than those algorithmic feeds.
You can hire someone to make a mixtape for you instead of listening to the machine's readily-available one.

The Style of No-Style
There's a service that will just ship you mostly-generic stuff (clothes, toiletries) every month. The algorithm age seems to wear away at individual taste - why not give up the fight, just let the algorithm decide?

A Pledge for the Self-Aware
We have taste, but we're also a drop in the river that carries us along. The "Generic Style of my time".

"Augmented creativity"
How can we use algorithms as a part of the creative process?
Glitch aesthetic indicates an understanding that algorithms exist, without resigning ourselves to their dominance.

The Innate Humanity of the Algorithm
Algorithms are generally driven by mass human flows.
Remember, you always have a vote!
(droqen: You're a drop in the river that carries you downstream.)


Taste Is Over! If You Want It
We can just avoid the algorithms and do things analog.

Echo, Echo, Echo
Nothing original emerges from algorithm feedback loops.