What this covers
Any Austin, a YouTube gaming commentator, sits down with economist Tyler Cowen to argue that video games have not improved qualitatively since the mid-1980s, and that the most rewarding way to encounter art—games included—is through direct observation rather than technical analysis or historical knowledge. Austin contends that while modern games like Red Dead Redemption II are vastly more complex and expensive than The Legend of Zelda, the emotional payoff is different in tone but not materially greater. The conversation weaves together media criticism, philosophy, and infrastructure analysis, with Austin treating video game design as a lens for understanding how constraint and imperfection shape all human systems.
The two explore several distinct territories. Austin argues that Art as Interaction means meaning lives between viewer and work, not within it—a sculpture moves without historical context, and that context adds a separate enjoyment but does not alter intrinsic value. He contends the YouTube algorithm is the most effective matching system ever built, that AAA games experienced a 15-year revenue bubble now deflating, and that competitive multiplayer titles occupy a fundamentally different cultural space from narrative games. On AI, Austin predicts image generation will become less artistically interesting as it grows more competent, mirroring gaming's failed chase for Realism Frontier Fallacy. Cowen counters that historical and contextual knowledge is constitutive to art's value, citing Michelangelo and Art in Dialogue. He also diagnoses the NBA's decline through Regression Toward the Mean: as a sport becomes popular, the product frays. The two disagree sharply on whether context matters and on virtual reality's promise, with Austin dismissing VR as merely more expensive, not better.
Any Austin argues that art across mediums—including video games—does not get qualitatively better over time, and that the most valuable way to engage with games (and the world) is a mindfulness-like practice of simply observing what things are rather than fixating on technical specs, narrative, or historical context.
- Video games have been basically as good and impactful since the mid-1980s NES era; they are now more complex and expensive but not more meaningful.
- Mainstream gaming commentary shares a flawed value system fixated on technical details and comparative analysis rather than felt experience and observation.
- The first-order goal is teaching people to look at things, think analytically, and use their brains, with infrastructure analysis as merely a veneer.
Media and art derive meaning from viewer experience and emotional resonance, not inherent properties.
- The meaning of art is not in the object itself but in the interaction between the viewer and the work—the value 'lives somewhere in between'—so a sculpture can be deeply moving with no historical knowledge, and knowing its history adds a separate kind of enjoyment that lives in a different part of the brain rather than changing the work's intrinsic meaning.
“The meaning of the art . . . is it’s not about the art. If the cup is the art . . . this is art, and this is me, and the value is here somewhere. It’s thisinteraction.”
- Video games have not gotten qualitatively better over time; once Nintendo revitalized the industry in the mid-1980s after the Atari crash, games have been consistently good and fun, and although modern games like Red Dead Redemption II are vastly more complex and expensive, the value you get in your heart is different in tone but, in the aggregate, not much more impactful than playing The Legend of Zelda in 1984.
“Itotallyreject thenotion that video games used to be better. Itotallyreject the notion that they used to be worse.”
- Austin's video game infrastructure analysis is fundamentally a second-order veneer; the first-order purpose is a mindfulness exercise teaching people that they are capable of looking at things, using their brains, and being analytical—where getting the interpretation 'right' doesn't matter, because an equally good argument can often be made for the opposite conclusion.
“it’s kind of a mindfulness exercise in a way. Let’s look at it. What is it? Let’s talk about it. . . . that’s all second order. First order is, learn to look at things and use your brain”
- Despite many commentators and low fixed costs creating apparent variety, gaming (and Marvel/comic) commentary shares one homogeneous value system that over-weights technical details like frame rates and settings and obsessive comparative analysis ('this game plus this game'), none of which actually matter, when we should instead engage games the way we engage a sculpture—asking what it made us feel and, first and foremost, what it literally is.
“the actual value system of all these commentators, the things that they care about are all exactly the same. There’s a real over-importance or whatever tied to technical things.”
AI and emerging technologies will be economically significant but artistically limited to commodity functions.
- We can't reliably distinguish authentic human art from inauthentic art even now—works we assume were made with sincere passion were often made under duress or chiefly for money—so it is very unlikely we will be able to distinguish AI-generated art from human art, undermining the common assumption that 'I'll be able to tell if it was made by a human.'
“it turns out that they made it under duress, and they didn’t really want to. They didn’t care about it. All of it is made up, whatever. To me, that makes it very obvious that we can’t even tell the difference between what we perceive as authentic human art versus inauthentic”
- AI image generation has become less artistically interesting as it has grown more competent, trending toward cleanliness and realism; this mirrors the 20-year struggle in video games to chase perfect realism, which has never been where the frontier of interesting art lies, so AI art's value will be in cheaply servicing low-level artistic needs rather than producing culturally enduring work.
“The art AI, I think, has gottenlessinteresting as it’s gotten more competent, whereas the inverse has happened with LLMs and stuff.”
- The real economic impact of AI image generation will be to absorb the work that would otherwise go to low-paid stock photographers or sketch artists—letting people get a competent picture of something rather than pay $50 for it—rather than producing the cool surreal art that humans actually treasure.
“It will suck up every job that would’ve otherwise been done by a low-paid stock photographer, or something like that”
- In the modern world, what sets you apart—your specific unique niche and charisma—becomes much higher value than being good at the established things other people already do, because uniqueness, not conventional competence, is what makes someone succeed going forward.
“It’s no longer that you need to be good at the thing that other people have established. It’s like your charisma, and it actually is your specific, unique niche that does set you apart. That stuff is going to be much higher value, I guess, going forward”
- Creators should deliberately avoid understanding their own inner machinations, holding only the thought that the current work will be the best they can make, because artists who learn what their 'thing' is tend to fly off a cliff—self-knowledge of one's creative process is more likely to be wrong and self-destructive than useful.
“The less I understand about my own inner machinations, the better because, number one, I’m going to get it wrong. You see this all the time with artists who learn what their thing is, and they just fly off a cliff.”
- New Orleans food is excellent for its specialties—gumbo, jambalaya, crawfish, and seafood—but the average quality of non-specialty cuisine like Thai, Chinese, or burritos there is not good.
“the average quality of food in New Orleans, in my opinion, is not good except for the things that they are a specialty in.”
Game commentary and platform analysis have become detached from actual engagement with the medium.
- Gaming's true cultural audience is more appropriately compared to books than to film or TV because the interactive barrier to entry is high—you must physically learn to do something—so the oft-repeated revenue statistic that 'games are bigger than movies combined' is a red flag that masks a much smaller player count, and the recent AAA-driven revenue surge was a 10-to-15-year bubble that is now deflating.
“it’s always been clear that, really, where video games sit probably is more appropriately compared to where books are. That’s probably the size of a healthy audience for video games.”
- Because the barrier to entry for watching commentary is far lower than for playing, we are probably already at a point where there are many people who watch others play games but have never played a game, and commentary about a medium can become bigger and more time-consuming than the product itself—paralleling someone who listens to six hours of NBA podcasts without watching a game.
“There’re a lot of people who watch people play games and have never played a game in their life. We’re probably there, would be my guess.”
- Long-lasting competitive multiplayer titles like Fortnite, VALORANT, and Counter-Strike, while technically video games, are culturally as different from narrative games like Zelda as football is from theater—both involve performance for an audience, but they occupy fundamentally distinct cultural categories—so lumping them together distorts our picture of what 'gaming' is.
“culturally, calling those video games and callingZeldavideo games would be like grouping football in with theater.”
- Virtual reality does not make games any better—replaying Resident Evil 4 in VR felt equally immersive as playing it on a flat screen years earlier—so Austin is not economically bullish on VR because its barrier to entry (cost, complex development, being cut off) is too high; people would adopt it if the experience were genuinely better or different, but it merely makes games different and more expensive, not better.
“I think about it now equally as immersive. Both cases, I felt like I was in the game. What’s the point? Why even put the thing on?”
YouTube's algorithm is extraordinarily effective; creators should prioritize it over broad cross-posting.
- The YouTube algorithm is the single most adept system ever built at matching content to the people who want to watch it, accounting for roughly 90 percent of a creator's discoverability, making the effort spent on cross-posting to other social media disproportionately low in reach and mostly washing out.
“If I was to put a number on it, I would bet that it’s 90 percent the YouTube algorithm. I don’t know that any other algorithm that’s been put into place is as adept at getting content that people want to watch to the people who want to watch it.”
- When people talk about 'the algorithm,' you should mentally replace the word with 'audience'—it is simply the audience—though this is not equally true of every algorithm, since some platforms' algorithms are better than others, and YouTube's is the best.
“If you ever talk to someone, and they go, “the algorithm, the algorithm,” replace the word “algorithm” with “audience.” It’s just the audience.”
- We are moving toward a more oral culture in which people don't like to read; if you want to get an idea across, writing works, but speaking tends to get people to listen and engage more.
“if I’m trying to get an idea across . . . you can write stuff, but if you talk stuff, people do tend to listen more.”
Imperfection and complexity in real-world systems deserve appreciation and mirror game design constraints.
- Real-world infrastructure like utility poles is usually 'wrong' relative to its ideal form because building a city involves so many colliding constraints and coordinating actors—the contractor who dug the hole, the person who must be called next—so the same imperfections people criticize in video game infrastructure actually mirror reality, and appreciating the immense coordination required is undervalued.
“The infrastructure you look at in real lifeisusually wrong. Also, some of the time, if you were to compare it to its ideal form, it wouldn’t look right.”
- Just as video games have codified rules that nonetheless go wrong, reality probably has imperfections where its rule sets interact strangely; rather than this implying a literal simulation or ghosts, it more likely reflects that across many possible universes the rules wouldn't always align cohesively, and the rules may simply hold for billions of years before breaking down.
“in video games, even though those rules are codified, things go wrong sometimes. I like this idea that is there a real-life equivalent of — this feels like a really lame way to say this — but hacking physics or whatever.”
- Nuclear cooling towers are aesthetically awesome and we should build more of them simply for their beauty—even generating steam for no functional reason—because the aesthetic value of infrastructure is itself a sufficient justification, independent of any power-generation need.
“We should build them for the aesthetics, honestly. We should just build those towers.”