Can you buy taste?
How tasteslop makes you spend more, buy worse, and never figure out what you actually want
Last week, a startup called Taste Labs came out of stealth with $18.5 million in seed funding. Their mission, in their words, is to “end AI slop” by building “the data and infrastructure layer to give AI models and agents taste.” Their founder says they want to measure, judge, and codify what “great” means in subjective domains — starting with design — so that AI can eventually tell you what’s good.
This is an objectively stupid concept (and based on the amount of replies, an opinion shared by most of the internet) because taste is not something you can automate or scale. You can’t extract it, codify it, or train a model on it, because taste isn’t a pattern you can reverse-engineer from a dataset of things that have been labeled “good.”
This attempt to automate and commodify taste has released a flood of algorithmically-generated aesthetic content that looks like taste lacks the substance. In the words of writer Emily Segal, it’s tasteslop.
And this flood of slop is making it harder for people to develop the kind of independent judgment that would actually help them spend better, buy less, and stop cycling through consumer identities every 18 months.
Taste is having a fun little moment in Silicon Valley
When AI can seemingly generate most things, what human output is left? The differentiator, says the tech bros, is judgement, or taste, writes Kyle Chayka in the New Yorker.
This is ironically the same industry that has spend decades telling everyone humanities degrees were worthless and that creativity was a nice-to-have in a world that really needed engineers. Our society leans towards tangible, quantifiable metrics of success like profit or utility over creativity and art, which are harder to measure. This is the same logic that powers optimization culture or the focus on income as a sole measure of success.
The more sinister reason behind the Silicon Valley’s desire for taste, Chayka writes: People hate AI and see it as a direct threat to their creativeity and livelihoods, so giving “anti-humanist technologies a veneer of liberal humanism” helps AI tools seem more palatable to the masses.
So enter tasteslop, or algorithmically generated content that mashes together things that are considered tasteful without any understanding of why those things were considered tasteful in the first place. It’s a veneer, writes Emily Segal, the writer who coined the term.
So what is taste, really?
Taste is discernment. It is your ability to perceive quality, to tell between things that work and things that don’t, and to know why, based on quality and your personal preference (both of which are subjective). Taste is an internal experience, and is cultivated over a long period of time.
Having taste isn’t the same thing as having style, necessarily. You can hire a stylist, you can copy a mood board, you can assemble a nice-looking outfit without having any understanding of why it works. And having taste definitely isn’t always having wealth. We’ve all seen rich people with terrible taste.
Many people misconstrue taste with aesthetic, which I see as the flashy, materialistic, superficial cousin. An aesthetic is a visual or sensory wrapper — a look, a vibe, a style. You can easily adopt an aesthetic simply because it looks good. But your taste is how you make that aesthetic your own. In other words, you can have aesthetic without taste.
Segal defines taste as requiring three things: discernment (can you tell things apart?), pattern recognition (do you know enough history to sense when something is new vs. derivative?), and idiosyncrasy (is your taste connected to specific, personal experiences that can’t be easily replicated?). Taste, she argues, is not a property of objects. It’s a relationship between objects, people, histories, and timing. And it needs to be socially validated (AKA nothing in a vacuum).
Tasteslop is the uncanny valley of taste. It has the right objects, the right color scheme, the right aesthetic — but it just feels a bit hollow, because it is. You see someone’s apartment or outfit and you read it as taste, but what you’re actually seeing is a set of signals that were algorithmically assembled. In other words, you can perform taste without actually having it.
Aesthetic is rooted in consumerism, taste inherently is not. You can buy an aesthetic. You literally cannot buy taste. But our consumer economy is structured to convince you that they’re the same thing.
The reason tasteslop can never actually be taste (no matter how much money you throw at it) is the copying and pasting things that look tasteful is OG the hallmark of bad taste. The whole point is that you figured it out yourself, through a process that can’t be replicated or scaled.
Building taste takes time, which is antithetical to how algorithms work. So what happens is you get a bunch of things that look like taste, which then becomes a commodified, which then becomes oversaturated as thousands of people adopt it as aesthetic, which causes it to lose its original meaning.Tasteslop exacerbates this culture “sameness” Chayka writes about, from the clothes we wear to the buildings we live in to the media we consume.
Concerningly, the online ecosystem could have become so polluted — so full of algorithmically generated content — that it may be eliminating our ability to build taste at all. We’re basically getting incepted 24/7. This, essentially, is the basis for the dead internet theory.
What tasteslop does to your money
Think about how much time you spend on your phone. Now think about what your phone is doing to your sense of what’s good, what’s desirable, what you “need.” Most apps that we interact with have some sort of algorithm, and those algorithms are pushing a constant stream of content (often aspirational, aesthetically-pleasing content) that is influencing our preferences, whether we realize it or not.
What’s insidious about this is that it makes you feel like you’re developing discernment when really the point of view was built personally, for you, by a system that studied your engagement patterns and served you more of what kept you scrolling. It eliminates our ability to develop independent value about what we really value, which is the whole baseline for how we set financial goals and spend our money.
Having taste helps reduce impulsive spending
I’ve written about how so much of our spending is identity spending — buying things not because we need them but because they reinforce a image we have of ourselves or who we want to be. Taste often helps provide a filter to help tell when you actually want an item or when you’re being influenced. When you know what you actually like, you buy things that serve your real life. When you don’t, someone else decides for you. That someone else is often the algorithm.
When you haven’t done the work of figuring out what you actually value, what genuinely interests you, you are extremely easy to sell to. The economy runs on these trend cycles, which essentially depend on people not knowing what they want, whose preferences are responsive to external signals rather than internal ones.
That’s how we get a world where everyone’s apartment looks the same, everyone’s wearing the same five brands, and every six months the whole aesthetic resets and you’re supposed to buy your way into the new one.
People with genuine, independent taste don’t need a trend to validate a purchase and they tend to buy fewer things and keep them longer.
Friction is how you develop taste (and make better financial decisions)
The antidote to tasteslop is friction, or the slow process of wandering, of serendipity of mulling over, of forming opinions, of being wrong sometimes. It develops on a long walk, or in a conversation with a friend, in the middle of a book, during a movie, when you’re bored. Chayka made this point in his New Yorker piece — it’s vital now to actively think about your own taste and separate it from the influence of digital machinery. But that separation requires time, discomfort, and a willingness to not know what you think yet.
Friction also produces better financial decisions. So much of our system is designed to make spending frictionless. Introducing friction back into your financial life is one of the most effective things you can do.
Some of this is practical: deleting saved payment info from your computer or phone, unfollowing accounts that consistently make you want things you didn’t want before you saw them, building a 48-hour rule for any non-essential purchase over a certain amount.
It’s also about how you spend your time. A huge part of developing taste is by spending unstructured time in the world. This is the opposite of optimization, which is why it feels unproductive. Taste requires time that doesn’t have a measurable outcome. And in an economy that has trained us to believe every hour should be monetized or optimized, that kind of time is often difficult to find. You can’t scroll your way into knowing what you like.
So — do you know what you like?
I am certainly not immune to any of this. For one, I make content for algorithms for a living, so I’m constantly navigating the gap between what I’d choose for myself and what I know will perform. I notice the sameness in my own consumption and I participate in it, sometimes consciously and sometimes not. Taste is something you have to keep working at, and the algorithmic environment we live in makes that work harder every day.
Do you know what you like? Can you say why? And did you figure it out yourself or did a feed decide for you?
If you can’t answer that, that just makes you a normal person living inside a system that profits from your uncertainty. But it does mean your spending is being shaped by someone else’s incentives, and you will keep cycling through identities and aesthetics without fully understanding who you are.
The thing that would actually help (slowing down, building friction) is free. It just takes time and friction and the willingness to be bored for a while. Which, in an economy that monetizes every second of your attention, is pretty expensive.
It is a little bit of a bitter irony that after years of being ridiculed and decried as useless, the AI world has gone crawling to humanities to understand how to generate preference, discernment, and the ability to read all kinds of text.
This was SO good thank you!