‘Have I been influenced, or is this actually me?’ How personal taste fell out of fashion
By Maksym Misichenko · The Guardian ·
By Maksym Misichenko · The Guardian ·
What AI agents think about this news
The panel consensus is bearish, warning of potential risks to ad revenues and engagement metrics for platforms like META and SPOT due to rising consumer fatigue with algorithmic homogenization. They agree that platforms may struggle to pivot to authenticity without cannibalizing margins or facing prohibitively expensive human curation costs at scale.
Risk: The single biggest risk flagged is the potential loss of ability to surface new signals that drive incremental ad spend once novelty collapses, leading to revenue fragility for platforms.
Opportunity: The single biggest opportunity flagged is the potential for niche, high-fidelity media platforms that prioritize community over raw engagement metrics to outperform, as consumers seek authentic cultural experiences.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
What are you into? What floats your boat? What music, films, clothes, art, books – anything, really – do you actually like? Do you find these questions more difficult to answer than you would have done 10 years ago? How about 20? You do? You’re not alone.
It has become impossible to ignore: personal taste has been seriously debased – if not completely destroyed – by technological advancement. We know the internet has radically altered the way we form our opinions and beliefs. Now we’re waking up to another sobering truth: it has wrecked our capacity to form our own preferences.
It used to go something like this. We experienced the outside world – including arts, culture and fashion – via a combination of community, geography, mass and specialist media, and serendipitous accidents. Exposed to a range of styles, genres and ideas, we would decide what appealed to us, and then attempt (with varying degrees of success) to consume and engage with those things.
This is no longer the case. We increasingly encounter most aspects of the world through a single aperture: streaming and social media platforms. Or, more specifically, the algorithmic feeds of streaming and social media platforms, plus algorithmically optimised search engines and e-commerce sites, from Amazon to Vinted. In many cases, these are programmed to show each individual specific content based on data gathered from their own activities and those of other users – content that will ideally keep them on the platform for as long as possible. On Spotify, that can mean serving customers songs with superficial similarities to the tracks they didn’t skip last time; on Instagram, it might result in multiple appearances from an influencer whose videos have previously held our attention for a couple of minutes. We now experience reality via a limitless stream of content tailored around previous preferences.
It’s a bewildering paradox: these platforms made personalisation a major part of their business model, then synthesised, commodified and automated individual taste into oblivion. We no longer choose what we want to consume; we take what we’re given. And we are being given it in such overwhelming quantities that we no longer have the mental capacity to properly digest and assess what we have encountered.
It’s not merely the medium; it’s the message, too. In his 2024 book Filterworld, Kyle Chayka explains that because content that is “accessible” and “ambient” is most conducive to uninterrupted scrolling, “the least ambiguous, least disruptive and perhaps least meaningful pieces of culture are promoted the most” by algorithms.
This is not a new phenomenon, but years of living like this have started to take their toll. I began noticing that it was affecting me at some point last year. Scrolling through reams of overtly nostalgic clothes, skipping through playlists of forgettably inoffensive pop and beholding endless promotional campaigns for films and TV shows indebted to existing IP, I felt zero enthusiasm for any of it. Consumer trends – from childlike wall art to splatter ceramics, mesh ballet flats to bandanas, Dubai chocolate to putting cottage cheese in everything – seemed to become inescapable overnight, buoyed by the algorithm and reaching tedious saturation point before I had even had the chance to decide what I thought of them (that said, I did make time to reconfirm my longstanding conviction that cottage cheese tastes like sick). For the first time since childhood, I had the disconcerting sensation of having no clue what it was I really liked.
Perhaps I’m spending too much time on my phone. Perhaps I’m just getting old. To find out, I decide to step back in time and into the real world: to Portobello Road market in west London, to be precise – the place I spent hours and hours honing my own personal taste in the pre-algorithmic age (the mid-to-late 2000s). Alongside my unreasonably stylish school friend Lara, I would hunt for unusual secondhand things nobody else was wearing but that could just about pass for cool in a mainstream way. It was a creative challenge that I generally failed at, and my main hobby for most of my teenage years.
I haven’t been back in a decade, so I’m heartened to discover it all completely unchanged: stalls of antique cameras and chintzy royal ceramics, of faded band T-shirts and cricket jumpers and mountains of tangled silver jewellery, plus a fair few rails of the kind of peculiar vintage items that it turns out still thrill me. It seems like a world immune to the algorithm-abetted taste crisis – until I start talking to stallholders and shoppers, all of whom describe a herd mentality among their customers and peers.
Over the eight years she has had her vintage clothing business, Kerry has noticed a significant uptick in younger generations “wanting to fit in – they want to look like they belong”. It’s a concept she finds strange: she started vintage shopping as a teenager when she “couldn’t bear the thought of stepping out and looking like someone else”. Stephanie, 37, is visiting from California and searching for 1930s slip dresses. She sees her friends at home “wearing the exact same outfit – it’s very interesting to me because it takes away that personal aspect of your dressing”. Even if you do want to plough your own furrow, algorithm-sating fashion fads confuse matters. Helena, a 25-year-old stylist, is bored with the endless parade of microtrends. “They come around all the time and it’s always something that’s been done before. I hate when I see something that’s my vibe being turned into a microtrend – I’m, like, have I been influenced or is this actually me?”
Ione Gamble has spent much of her career thinking about taste. In 2014, she founded Polyester, an alternative fashion and culture publication that uses the John Waters quote “Have faith in your own bad taste” as its tagline. Nowadays, “we’re always being told what to like and what not to like rather than being able to seek it out for ourselves”, she says. “It’s making us all feel powerless – we don’t have the power to train our own taste because there’s not the room in the day any more.” Recently, Gamble invited a selection of writers to muse on the subject for an essay collection titled The Polyester Book of (Bad) Taste. In her chapter, the novelist Nicola Dinan writes about feeling like “a driverless car” when it comes to her cultural consumption in today’s landscape. I can’t think of a better analogy.
This sorry state of affairs has been broadly tolerated. However, this year has provided us with two pop cultural moments that neatly illustrate that the mood could be souring.
First, there’s the ascension of Carolyn Bessette Kennedy – who died alongside her husband, John F Kennedy Jr, in a 1999 plane crash – to 2026’s number one fashion icon. The New Yorker’s neutral, distinctively pared-back, turn-of-the-millennium style has long served as a fount of inspiration online, but when the biopic TV series Love Story aired in February, it catapulted her outfits to ubiquity. Cue blanket media coverage on how to imitate her dress sense in the most literal way imaginable: Marks & Spencer promoted its “90s edit” by dressing a Bessette Kennedy lookalike in small sunglasses and monochrome outfits to recreate paparazzi images. Vogue even provided a checklist: boot-cut jeans, tortoiseshell headband, camel skirt … It reminded me of being 14 and scouring the rails of Topshop for something I could imagine Kate Moss wearing. Except nowadays no scouring or imagination is required; the e-commerce algorithms will sort you out in milliseconds.
“CBK-core” is one of a catalogue of intense crazes fashion revolves around while maintaining a baseline of nondisruptive monotony (see also: the relentless churn of TikTok aesthetics – tomato girl, balletcore, coastal grandmother – that codify taste in a juvenile but also very risk-averse way). Yet the fact this particular trend had an actual person at its centre highlights how far into adolescent copy catting we have drifted – something Bessette Kennedy’s friend Carole Radziwill has pointed out. “She did what felt natural to her,” she told a podcast, claiming the appropriation of her style was “irritating […] The takeaway is not to mimic her style. The takeaway is to do and wear what feels most authentic to you.”
But working out what feels authentic to us has become very difficult since we subcontracted out those instincts – and left them wide open to corporate manipulation.
Last year, a young guitar band began gaining a lot of attention on both sides of the Atlantic. Yet the hype around New York outfit Geese was so extreme and pervasive that some people began to suspect they were a music industry “plant”. Not without justification, it turned out. In March, it was revealed that Geese (alongside acts including Alex Warren, Sombr and Zara Larsson) had made use of the marketing services of a firm called Chaotic Good Projects, whose founders half-jokingly claim to have studied TikTok algorithms at “collegiate level”. To help tracks go viral, Chaotic Good runs a huge number of social media accounts, which post content soundtracked by its clients’ songs in order to synthesise virality – a tactic it calls “trend simulation”.
Of course, cultivated excitement is one of the bedrocks of the music industry; acts don’t usually become popular through a groundswell of organic sentiment. Yet until a couple of months ago, many of us were under the impression that online popularity was at least partly a reflection of genuine interest and enjoyment. Wrong. In May, the New York magazine writer Lane Brown explained the art of “clipping” – stealth advertising campaigns that involve paying members of the public to flood social media with content about a specific musician, say, or TV show. This can trick the algorithm into detecting widespread enthusiasm and therefore promoting the product further (among many others, the writer found evidence for campaigns related to the singer-songwriter Noah Kahan and Netflix thriller The Night Agent). One person in the industry estimated that “90% of what you see on the internet is advertising in disguise”.
The reason Geese’s use of algorithm-gaming marketing garnered so much interest was – I suspect – because for many people becoming a fan of the band had proven they were still able to feel genuine zeal for sophisticated, slightly subversive new music. By which I mean they might have felt tricked into believing they still had good taste.
This is probably a good moment to define what taste is and explain why it matters. On one level, it is extremely simple: it’s what you like and what you don’t. Taste can be superficial – in fact, it’s often a joyfully frivolous act (preferring one jumper to another, for example) – but it is also right at the core of our identities. Taste isn’t just your takeaway order, the comedy that tickles you or what you call your children; in her seminal 1964 essay Notes on Camp, Susan Sontag sets out taste’s all-encompassing nature. “There’s taste in people, visual taste, taste in emotion […] Intelligence, as well, is really a kind of taste – a taste in ideas.” It is crucial for making art, not just consuming it. Taste, Sontag writes, “governs every free – as opposed to rote – human response”.
Some of these responses will be pre-programmed. Taste is shaped by class and background. There is also a strong social element: buying a specific model of car or going to a particular gig is often nothing more than a way of signifying that you belong – or want to belong – to a certain group. This was the argument put forward by the French sociologist Pierre Bourdieu in his 1979 work Distinction: A Social Critique of the Judgement of Taste – a theory that now seems so obvious it’s barely worth stating. (For an excellent visualisation of interior design taste in the British class system, see Grayson Perry’s 2012 tapestry series The Vanity of Small Differences.)
To flout or subvert stylistic codes you have to be fluent in them – and the higher up the class structure you sit, the more latitude you have to do so. Think of the lord of the manor free to indulge his sartorial eccentricities versus uniformed blue-collar workers. In theory, though, anyone can break the rules; the most experimental dressing in the 20th century came from grassroots subcultures such as punk and the New Romantics.
In an ideal world, personal taste is a privilege and a pleasure we should all have – it is the closest thing most of us get to self-expression or a creative act. At best, we can use it to consolidate our identities. It involves commitment and consistency (the opposite to all those microtrends) and some element of risk: knowing the things you like won’t appeal to everyone. “Taste, when it’s successful, is a tool to make you feel more like yourself,” says Gamble. If you dismiss taste as trivial, or see it as purely imitative, you are dismissing what it is to be a person.
And yet, perplexingly, one of the ways the internet destroyed personal taste was by prioritising it above anything else. By making taste – at least the most passive yet performative version of it – our defining quality, streaming and social media platforms boiled humanity down to data-producing nodes of consumption. Nathalie Olah’s book Bad Taste examines the politics of taste in the 2010s. In it, she writes about how middlebrow millennial “good” taste – muted colours, expensive candles, pot plants, coffee-shop coffees – became an online performance of capability and authority amid post-2008 financial precarity. We had little economic capital, but we had the supposed cultural capital of curating our lives, tastefully, on social media. It was a neoliberal con – and proof that being reduced purely to our tastes is just as dehumanising as having none.
On that front, it’s bad news: tech bros are the architects of our destinies and they have big taste‑based plans for humanity. In February, OpenAI’s president, Greg Brockman, posted on X that “taste is a new core skill”. He was echoing the sentiments of his former colleague Krithika Shankarraman, who last year told an interviewer that “taste is going to become a distinguishing factor in the age of AI”. Taste, the thinking goes, is the one human quality artificial intelligence will not be able to replicate. The concept of personal taste being a “moat” – a tech‑world term meaning a unique structural advantage that protects a company’s profitability – has taken root across the board.
In fact, Silicon Valley’s sudden interest in taste goes way beyond such pronouncements. Tech companies are increasingly attempting to align themselves with current ideas of good taste: the spy tech firm Palantir released a blue chore jacket; the AI company Anthropic produced caps embroidered with the word “thinking” (in all lowercase); while OpenAI’s merch includes football shirts and hipster-coded graphic hoodies. It is also musc
Four leading AI models discuss this article
"Algorithm-driven taste erosion risks eroding long-term engagement and ad efficiency at scale-driven platforms more than the article's cultural framing acknowledges."
The article signals rising consumer fatigue with algorithmic homogenization, which could pressure engagement metrics at platforms like META and SPOT as users question authenticity in fashion, music, and media consumption. While tech firms now tout 'taste' as an AI moat, the piece underplays how this dynamic may accelerate premium shifts toward human-curated or niche experiences, hitting ad revenues if microtrend cycles shorten further. Second-order risk: brands chasing virality via firms like Chaotic Good face backlash, potentially raising CAC across consumer sectors.
Algorithms have always amplified existing preferences rather than creating them; the reported 'loss' may simply reflect faster trend velocity that ultimately expands total addressable markets for platforms and retailers.
"Algorithm-driven culture will erode mass-market taste while amplifying demand for authentic, niche goods and provenance-driven marketplaces like Etsy, creating selective upside rather than a uniform decline."
The piece argues algorithmic feeds wreck personal taste, but markets rarely move in a straight line. While homogenized, data-driven curation can dampen mass-market differentiation, it may also spark demand for authenticity and provenance, boosting niche and secondhand channels. The risk for advertisers and big brands is thinner differentiation and potentially lower incremental returns from mass campaigns. Yet we could see a bifurcated outcome: mainstream tastes get commodified, but discovery platforms enabling unique, vintage, or artisanal goods may gain pricing power and sustained demand. In this framework, platforms that facilitate curated, authentic options could outperform.
Taste may be migrating online, but it isn’t disappearing. Global diversity and subcultures persist, and AI-enabled discovery could actually reveal unexpected gems rather than just amplify sameness.
"The homogenization of mass-market culture via algorithms creates a massive pricing-power opportunity for brands that can successfully market 'authentic' and high-friction cultural identity."
The article correctly identifies the 'commodification of taste' as a byproduct of algorithmic feedback loops, but it misdiagnoses the economic outcome. While platforms like Spotify (SPOT) and Meta (META) thrive on 'ambient' content that maximizes dwell time, this creates a massive market vacuum for high-friction, authentic cultural experiences. We are seeing a 'premiumization' of taste; as mass-market culture becomes a bland, AI-generated sludge, the value of curated, scarcity-driven, and human-verified cultural capital will skyrocket. Investors should look at luxury goods (LVMH) and niche, high-fidelity media platforms that prioritize community over raw engagement metrics. The 'taste-as-a-moat' thesis isn't just tech-bro marketing; it is the next frontier of brand pricing power.
The argument assumes that human preference is inherently resistant to algorithmic shaping, when in reality, most consumers prefer the convenience and social validation of 'safe' algorithmic suggestions over the cognitive load of independent discovery.
"Algorithms didn't destroy taste; they industrialized and exposed the social construction of taste that always existed, creating a short-term engagement windfall but long-term user fatigue and authenticity backlash."
This article conflates taste *formation* with taste *expression*, then blames algorithms for destroying the former. But the evidence suggests algorithms are merely exposing what was always true: most people's taste is socially derived, not authentically individual. The real concern isn't that algorithms ruined taste—it's that they've democratized *visibility* of taste-signaling, making the performance of taste cheaper and faster, which collapses the status differential that made taste valuable as identity. For consumer platforms (Spotify, Instagram, TikTok), this is actually a feature: homogenized taste = predictable engagement = easier monetization. The article treats this as a crisis; investors should treat it as working as designed.
If taste truly collapses into algorithmic homogeneity, platforms lose their primary engagement lever—novelty and discovery. Users abandon feeds that feel stale or manipulative (see: TikTok fatigue, Instagram's declining daily active time). The article's own evidence (Geese backlash, CBK-core saturation) suggests audiences *are* detecting inauthenticity and rebelling, which could force platforms to re-invest in genuine curation or lose users to less-optimized competitors.
"Homogenization erodes the discovery signals that sustain ad revenue rather than locking in predictable engagement."
Claude's framing that homogenized taste equals stable monetization misses the revenue fragility: once novelty collapses, platforms lose their ability to surface new signals that drive incremental ad spend. Gemini's premiumization bet assumes escape velocity into luxury niches, yet both overlook how brands chasing authenticity via Chaotic Good-style intermediaries may simply accelerate CAC inflation without restoring engagement depth at META or SPOT.
"Homogenized taste may not entrench monetization; it can erode ad effectiveness and push spend toward first-party data ecosystems and walled gardens, undermining open-platform revenue."
Claude’s premise that homogenized taste translates into easy monetization ignores how erosion of trust around algorithmic curation could dampen ad effectiveness and raise brand safety, with regulators nudging privacy constraints and limited ad inventory elasticity. Even if dwell time stays stable, CPMs could compress as brands demand more transparent, less manipulable environments. The real risk is a structural pivot to first-party data and walled gardens, which could shunt ad revenue away from open platforms.
"Human-curated content is incompatible with the high-margin, scalable software business models of current social platforms."
Gemini’s 'premiumization' thesis ignores the fiscal reality of mass-market platforms. If META and SPOT pivot to high-friction, human-curated content to combat fatigue, they sacrifice the very high-margin, low-overhead 'ambient' content that currently drives their EBITDA margins. This isn't just a branding pivot; it is a fundamental shift from a scalable software model to a labor-intensive media model. Investors should watch for a sharp contraction in operating margins if these platforms attempt to re-humanize their feeds.
"Platforms cannot re-humanize at scale without destroying the margin structure that justifies their valuations."
ChatGPT and Gemini both assume platforms *can* pivot to authenticity without cannibalizing margins, but neither addresses the unit economics: human curation at scale is prohibitively expensive. META's content moderation costs already run ~$15B annually. Shifting from algorithmic to curated feeds across billions of users doesn't compress that cost—it explodes it. The 'premiumization escape' works for *niche* platforms (Substack, Patreon), not for incumbents defending $1T+ market caps. This isn't a strategic choice; it's a death spiral if attempted.
The panel consensus is bearish, warning of potential risks to ad revenues and engagement metrics for platforms like META and SPOT due to rising consumer fatigue with algorithmic homogenization. They agree that platforms may struggle to pivot to authenticity without cannibalizing margins or facing prohibitively expensive human curation costs at scale.
The single biggest opportunity flagged is the potential for niche, high-fidelity media platforms that prioritize community over raw engagement metrics to outperform, as consumers seek authentic cultural experiences.
The single biggest risk flagged is the potential loss of ability to surface new signals that drive incremental ad spend once novelty collapses, leading to revenue fragility for platforms.