Ruby Thelot | The a16z Show Summary

1. Key Themes

The Pluriculture Fragmentation Problem

The monoculture of broadcast television has shattered into thousands of insular digital communities. Ruby observes that shows which dominate online discourse have far smaller audiences than older mass-market hits — and the gap between perceived cultural dominance and actual reach is widening.

"If you take a look at what the viewership for that is compared to a show like MASH, compared to even the finale of Lost, not that many people have seen it. But if you're on X, it feels like everyone's watching Succession. Right. And yet the numbers, kind of, when you compare them to even a show like Big Bang Theory, more people watch Big Bang Theory than Succession, which kind of sounds crazy when you think about it." [00:08:49]

Balkanization and Babelification: The Two-Stage Internet Breakdown

Ruby and his collaborator Rui Yee coined a two-part framework: first, the internet "Balkanizes" into insular communities; then, inside those communities, language "Babelifies" — the same words acquire completely different meanings depending on which digital island you inhabit.

"The problem with containment, with being so insular, is that when we have something like looks maxing or anything that breaks containment... we've developed our own unique idiosyncratic micro languages. We no longer know what things mean when we speak to each other online. We're using the same word, but they have a completely different meaning if you're in the furry community, if you're in the incel community, if you're in high fashion Twitter." [00:09:41]

Triple Capture: Audience, Algorithm, and Now Machines

Content creators are no longer just shaped by their human audiences — they are simultaneously captured by platform algorithms and, increasingly, by bot audiences. With Cloudflare reporting ~50% of internet traffic as bots, Ruby predicts a coming era of fully machine-to-machine content creation and consumption.

"There's this waltz, this three-legged waltz between machine, algorithm and your human audience. And if we predict, as I do, that the number of bots is going to increase, suddenly it may just be that we will only have bots making things. We'll only have AI-generated content, an AI audience. And suddenly what we'll see is what I call machinic taste." [00:13:18]

Taste as Virtue, Not Just Aesthetics

Ruby reframes "taste" — the most-hyped concept in AI-era tech discourse — with deep historical grounding. The question of taste emerged in 17th-century England precisely when a new wealthy class needed guidance on consuming virtuously. Ruby argues we are in an identical inflection point today.

"Taste is not just aesthetic preference. Taste is a way for people to consume virtuously. The question of taste emerged in the 17th century at a time where people needed guidance as to how to consume properly." [00:17:30]

"We're entering a similar industrial revolution. We are faced with or at the eve or maybe in the midst of immense amounts of wealth being created. And we're asking ourselves, how do we consume virtuously? And that is why the question of taste is emerging. It mirrors 17th and 18th century England in that sense." [00:18:47]

The Stated vs. Revealed Preference Gap on AI

Americans say they dislike "AI" but enthusiastically use the very tools it powers. Ruby identifies this as primarily a language and framing problem — the word "AI" conjures job loss, while the same technology branded differently is beloved.

"I find that Americans don't like AI. AI is a very broad term, a very broad umbrella here because they all use chat. And like, I love it. I love chat. But they don't like AI. Does that make sense? There's like a bit of a divergence there." [00:25:06]

"When I talk to mothers in Bryan, Texas, like, oh yeah, like I had a leak in my faucet, whatever. I took a photo and in about an hour or two, I was able to fix that thing. I love that. But to them, it's kind of like Google." [00:00:00]

Journalists' Algorithmic Bias Distorts Trend Reporting

Ruby's quantitative research reveals that media narratives about cultural trends are often artifacts of reporters' own recommendation algorithms rather than genuine mass phenomena. His analysis of 2,000 top TikTok dating videos over five years found that dating content is mostly positive — contradicting the widely reported "heteropessimism" narrative.

"Dating content is mostly positive. But our beloved journalists, I think because of their algorithms, I think are getting hit with a lot of that content. Right. And then they think it's a trend. But when I take a look at the data, at the top 2,000 videos across five years, I can say that it's actually kind of decreased over time." [00:05:50]

Platform Culture Has Replaced National Culture as Identity Marker

Online etiquette and personal identity are now defined less by ethnicity or nationality and more by which platforms a person inhabits. Ruby argues this creates novel social friction because there is no shared "etiquette book" for cross-platform interaction.

"Whereas culture usually is somewhat ethnicity based or nation based... the culture online is not specific to what a person looks like, but more which platforms they're on. I would argue that there's a Tumblr culture, a Twitter culture, an Instagram culture." [00:30:57]

New Aesthetics Exist — They're Just Not Mass, and Finance Killed Discovery

Ruby pushes back on the "no new aesthetics" panic, arguing aesthetics are being born constantly in synthesis — but financialization of culture industries and the death of critical gatekeepers prevents them from reaching mass audiences.

"There are many new aesthetics. They're just not mass. And it's very difficult to make a new mass aesthetic because for the most part, mass culture has been subsumed by finance... critics also are no longer doing their job, right? Because we publish on platforms that are advertising based. And so we can't be as mean and as critical as we used to be." [00:14:36]

2. Contrarian Perspectives

"Is This a Real Thing or Is This Three TikToks in a Trench Coat?"

Most media treat any viral social behavior as a meaningful cultural trend. Ruby's contrarian claim is that the vast majority of reported "trends" are statistical noise amplified by reporter filter bubbles — and the data almost never backs the narrative.

"I like to say, is this a real thing or is this three TikToks in a trench coat? Right? Because oftentimes what feels like a trend, what feels like a novel social behavior can just be super micro, super small." [00:04:54]

Techno-Optimism Will Rise — But Only When AI Stops Threatening Bread

Against the prevailing tech-world assumption that better products and more capability will win people over, Ruby argues the adoption unlock is purely economic and communicative: AI will be embraced when it is clearly seen as protecting, not destroying, livelihoods. The messaging problem is as large as the technology problem.

"Techno optimism will be on the rise once the claimed benefits of the technology are aligned with the desires and interests of American constituents... I don't know who thought it was a good idea to go on national television, tell people they're going to lose their job. But if I'm an American, I wouldn't want to lose my job because that's how I feed my family. It's kind of very simple, frankly." [00:26:33]

Agents Will Develop Their Own Aesthetic Subcultures

The conventional assumption is that AI models converge on a single homogenized aesthetic. Ruby's contrarian prediction is the opposite: a rich LLM ecosystem with different model weights will produce divergent machine taste — agents will have their own aesthetic subcultures, just as human online communities do.

"If we have a rich LLM ecosystem with different kind of models and different kind of weights, then one can surmise that they'll have their own specific aesthetic preferences." [00:14:08]

Good Taste Is a Form of Ungovernability

In a world where the purpose of platform algorithms is to capture and categorize users, Ruby reframes taste as a kind of political resistance — not a luxury signifier but a survival tool for resisting algorithmic reduction.

"Good taste means I can't be fit into this little box that the algorithm is designed for. And it's almost when people say the importance of taste, it's not just about having a preference or it's not just about the aesthetics of it, but it's this bigger mission of how do you actually become ungovernable? How can you see past these algorithms?" [00:20:29]

AI Scope Creep in Student Projects Is Net Positive

While the dominant media narrative frames AI use in academia as cheating or shortcutting, Ruby observes the opposite effect in his NYU thesis program: students are taking on larger, more ambitious projects precisely because AI gives them access to knowledge outside their specialization.

"I've actually seen an increase in like the size and the scope of the project. People feel like they are able to tackle them now." [00:32:14]

3. Companies Identified

R2S / 131

Ruby Thelot's research and consulting firm. Mentioned as the organization conducting quantitative social media analysis — including the heteropessimism study of 2,000 TikTok videos and an upcoming index of 5,000 AI-sentiment videos — as well as providing cultural intelligence to major brands.

"What my firm, R2S, what we do at 131 is we do a lot of quantitative analysis of social media to see how prevalent is this thing, really." [00:04:54]

Anthropic

AI safety company. Cited for a large-scale survey of 85,000 users examining how people use AI for career advancement and life management — used by Ruby as evidence of the real revealed-preference adoption of AI tools.

"If you take a look at Anthropic Survey, they surveyed 85,000 people to ask how they were using their AI for career advancement... at work or life management." [00:08:49]

Cloudflare

Internet infrastructure company. Cited for its data point that approximately 50% of internet traffic is now bots — a foundational stat for Ruby's "machinic taste" thesis.

"We might have seen the numbers from Cloudflare coming out this year, about 50% of internet traffic as they see it as bots." [00:12:03]

Mozilla Foundation

Nonprofit behind the Firefox browser. Mentioned as a research partner for Ruby's study on the social meaning of Instagram's "Close Friends" feature.

"With the Mozilla Foundation recently, I did research on close friends... If you're in my close friends, are we acquaintances? Are we like families?" [00:02:08]

Social media platform. Mentioned through Matt Klein, its Head of Insights, as the source of the "audience capture" concept central to Ruby's content-evolution framework.

"The content of audience capture I get from my dear friend Matt Klein, who's the head of Insights at Reddit." [00:12:32]

H&M

Global fashion retailer. Cited as a consulting client of Ruby's firm, specifically for commissioning research into digital micro-communities — which directly led to product decisions like the Dime Square shirt.

"H&M recently had a Dime Square shirt last year. And so that is based on research that was done on different digital communities." [00:11:10]

Pepsi

Beverage company. Named as a consulting client alongside H&M, indicating the breadth of industries commissioning cultural intelligence research on digital communities.

"A big part of my job when I do consulting, you know, anywhere from H&M to Pepsi is to kind of go into these small digital communities and like map them." [00:10:56]

Betaworks

New York-based VC and startup studio. Mentioned as the venue where Ruby delivered his essay "The Devil You Know" — the talk that originated from the essay that got him banned from LessWrong.

"The essay that got me banned actually ended up being a talk that I gave at Betaworks, the VC fund in New York." [00:22:56]

4. People Identified

Ruby Thelot (Ruby Justice Thelot)

Designer, artist, cyber ethnographer, and NYU professor in the Integrated Design and Media program. Author of A Few Essays on Taste. Founder of research/consulting firm R2S / 131. Mentioned throughout as the primary guest and expert voice — notable for his quantitative approach to cultural trend analysis and his historical framework for taste.

"Taste is not just aesthetic preference. Taste is a way for people to consume virtuously." [00:17:30]

danah boyd

Researcher and author; formerly a professor at Cornell and longtime researcher at Microsoft Research. Cited as a foundational figure ("the goat") in cyber ethnography, specifically for her research on MySpace's Top 8 feature — work that directly informed Ruby's own research on Instagram Close Friends.

"There's the goats like danah boyd. I mean, she's a big inspiration for my own research. A big part of me trying to understand what close friends meant was looking at her research on MySpace Top 8. She was at Microsoft for a long time. She was at Cornell as a professor and she was leading a team at Microsoft Research." [00:03:29]

Matt Klein

Head of Insights at Reddit. Named as the originator of the "audience capture" concept — the idea that creators become prisoners of their audience's preferences — which is a cornerstone of Ruby's triple-capture framework.

"The content of audience capture I get from my dear friend Matt Klein, who's the head of Insights at Reddit." [00:12:32]

Rui Yee

Ruby's collaborator and co-author. Named as co-creator of the "Balkanization and Babelification of the Internet" paper — the conceptual framework that underpins much of Ruby's analysis of internet culture fragmentation.

"My friend Rui Yee and I worked on this paper called The Balkanization and Babelification of the Internet." [00:08:49]

Tyler Cowen

Economist and public intellectual. Mentioned as someone so concerned about the absence of new aesthetics that he created fellowships specifically to incentivize their creation — cited as external validation of the problem Ruby addresses.

"Even people like Tyler Cowen have like created fellowships for like somebody please make a new aesthetic." [00:14:18]

5. Operating Insights

Quantitative Signal Testing Before Publishing Cultural Claims

Ruby's firm runs a systematic quantitative check on every proposed trend before treating it as real — pulling the top posts by view count across platforms, sampling thousands of videos, and measuring prevalence over time. This is a replicable research protocol any media, brand, or investment team could adopt to separate signal from noise before acting on a "trend."

"We took a look at the top posts about dating on TikTok. Editing over a million views, took a look at 2,000 videos. And then, using the scripts, identifying did heteropessimism grow since 2020 on TikTok, Reels, and across different platforms." [00:05:50]

Map Digital Communities Before Deploying Product or Marketing

Ruby's consulting model for brands like H&M and Pepsi is to enter small digital communities, map their distinct language and culture, and only then make product or marketing decisions. This suggests a pre-launch cultural audit — identifying the specific micro-community "epicenters" most relevant to a product — is a high-ROI step that most companies skip.

"A big part of my job when I do consulting, you know, anywhere from H&M to Pepsi is to kind of go into these small digital communities and like map them, that H&M knows what to sell." [00:10:56]

Use Platform-Specific Cultural Codes to Signal Credibility

Ruby observes that a person's dominant platform is immediately legible from their vocabulary, humor, and communication style. For operators building in public, content creators, or anyone doing B2C marketing — deliberately studying and mirroring the vernacular of the platform you are publishing on is a concrete, underrated lever for organic reach and credibility.

"The culture online is not specific to what a person looks like, but more which platforms they're on... when you meet someone, you can tell which platform they're on the most just by how they talk." [00:30:57]

6. Overlooked Insights

An Upcoming Publicly Available AI-Sentiment Index Could Become a Reference Dataset

Ruby briefly mentions that R2S / 131 is publishing, within roughly a month of the recording, a quantitative index of AI sentiment derived from 5,000 social media videos — tracking positivity, negativity, and trend direction over time. This is a non-obvious, high-value dataset. For investors tracking AI adoption curves, brand teams deciding how to position AI features, or policy researchers, a rigorous longitudinal index of actual public AI sentiment — not survey-based opinion polling — would be genuinely novel and potentially influential.

"At R2S and 131, we have a paper coming out specifically on that. Same thing with the heteropessimism. We're going to be analyzing about 5,000 videos on AI to sort of determine what is the current sort of positive negative and see how it checks over time. It's going to be one of our indexes that we publish in the next month." [00:28:38]

"Machinic Taste" — Agent-to-Agent Aesthetics as a Coming Design Constraint

Ruby's concept of machinic taste — that as bots become a majority of content audiences, content will evolve to satisfy machine preferences rather than human ones — was treated as a speculative curiosity in the conversation but has immediate, practical implications. If AI agents become the dominant "consumers" of content (for training data, for agentic browsing, for summarization), the aesthetic and structural properties that perform best for machines may diverge sharply from what humans find appealing. This creates a design and product split: human-optimized vs. machine-optimized content formats could become a real product category, and companies that recognize this early could build significant competitive advantages in content infrastructure.

"Not only is there audience capture, but to get to your audience, there's also a gatekeeper, which is the algorithm... And now when your audience is now 50% bot, 50% human, we suddenly see machinic or agentic capture happening too... We'll actually get to see what are the aesthetic proclivities, what are the aesthetic preferences of an agent-to-agent connection." [00:12:32]