Notes — Any Austin on the Hermeneutics of Video Games

Notes on Any Austin in conversation with Tyler Cowen — Conversations with Tyler (https://conversationswithtyler.com/episodes/any-austin/), 11 June 2025.

Four questions [Adler frame]

Q1 — What is it about as a whole? A conversation about how to look — using video games as the worked example. Austin, the YouTuber who analyses the infrastructure, hydrology, and populated backgrounds of game worlds as though they were real places, treats close attention as the actual subject; games are the ‘veneer’. The talk radiates from his niche outward to reality’s possible glitches, AI art, whether an artwork’s meaning depends on its history, the aesthetics of nuclear cooling towers, whether games (or any medium) improve at all, and finally to Austin turning the tables and interviewing Cowen.

Q2 — How is it argued? Improvisationally and by analogy rather than by system — Austin repeatedly flags that they are ‘making this up as we go, exploring this idea live’. He reasons from concrete cases (utility poles in Liberty City, Morrowind’s hydrology, replaying Resident Evil 4 in VR, Atari-era games) to general claims, and habitually reframes Cowen’s binary or declinist questions (‘are games getting better?’, ‘is reality a simulation?’) by rejecting the frame and offering a subtler version. Cowen supplies the counter-pressure — the productivity case for AI art, the role-in-history case for meaning, the secondhand ‘many people say games are in decline’ claim.

Q3 — Is it true, in whole or part? Much of it is aesthetic judgement and self-report rather than falsifiable claim, and both men treat it that way. The strongest, checkable observations — that real infrastructure rarely matches its ideal form, that gaming’s headline revenue figures are inflated by an AAA and competitive-title bubble, that a large share of self-described ‘gamers’ buy no game in a year — are asserted from Austin’s domain familiarity, not sourced here. [?] His ‘12 per cent of self-described gamers will buy a game this year’ figure is quoted from memory of ‘something that came out recently’ and is not independently verified in the transcript. The AI-art and games-do-not-improve claims are explicitly predictions and preferences, hedged by Austin himself (‘if you project long enough, all of these things will come undone’).

Q4 — What of it? A transferable stance on attention and taste: meaning lives in the interaction between viewer and object, technical fandom often optimises for the wrong (second-order) things, and ‘learning to look’ is a portable discipline well beyond games. It also supplies a sharp contrarian read on AI creativity — that competence and interestingness can move in opposite directions — and a vivid case study in owning a niche so specific it sounds strange, which Cowen frames as the modern route to doing very well.

Glossary

Hermeneutics — the practice of interpretation; reading a thing closely for meaning. Here, applied playfully to video games — treating a game’s rivers, wiring, and background NPCs as texts to be interpreted rather than mere scenery. [§ Opening — niche, the algorithm, and noticing]

Ideal form — the imagined perfect version of an object (a ‘correct’ utility pole, a ‘correct’ river) against which people judge a game’s realism; Austin’s point is that the real-world object rarely matches its ideal form either, so the mismatch is not a mistake but the norm. [§ On infrastructure and the ideal form]

Anthropic principle — the observation that any universe we find ourselves observing must be one whose rules cohere well enough to permit observers; Austin reaches for it (Cowen names it) to explain why reality’s rules seem to hold — a survivor’s-eye view rather than proof of design. [§ On glitches IRL]

AAA game — an industry term for the biggest-budget, mass-market titles (and, by extension here, large competitive games like Fortnite and Counter-Strike); Austin argues these inflate gaming’s headline size and are culturally a different thing from story-driven games — ‘like grouping football in with theater’. [§ On the trajectory of video game quality]

Second-order vs first-order (in commentary) — Austin’s distinction: first-order is noticing what a thing is and thinking about it; second-order is the received technical value system (frame rates, spec comparisons) that games commentary fixates on. His project is to move attention back to the first order. [§ On the growth in commentary]

Key claims by section

Opening — niche, the algorithm, and noticing [§ Opening]

Being simultaneously the best in the world at something and the only person doing it is, for Austin, an ideal position; Cowen generalises it — in the modern world, value accrues to a unique, hard-to-name thing that sets you apart, and if you cannot outline what you are best at, doing very well is hard.
Austin credits ~90 per cent of his discoverability to the YouTube algorithm, which he rates as uniquely effective at matching content to interested viewers; other channels (cross-posting, word of mouth) mostly ‘come out in the wash’.
He describes a lifelong habit of noticing and tracing how ordinary things connect — possibly seeded by a childhood of replaying tiny stretches of games under strict time limits, learning to find detail in small spaces.

On infrastructure and the ideal form [§ On infrastructure]

People judge in-game infrastructure against an imagined ideal form and flag anything non-ideal as an error; but real-world infrastructure is ‘usually wrong’ too, being the product of colliding contractors, coordination, and accumulated effort.
The insight is appreciative rather than technical: the coordination behind something as mundane as a paved road with a utility pole reaching a house is vast and under-appreciated.

On glitches IRL [§ On glitches IRL]

Austin sets aside both the simulation-hypothesis and the flashy Matrix-style glitch as unlikely framings; the interesting version is that reality, like a game, runs on rules that even when codified sometimes interact strangely — a possible real-world analogue of ‘hacking physics’.
Cowen adds the physics parallel (quantum mechanics and general relativity are not obviously consistent) and a many-universes thought — some possible worlds have glitches and merely work ‘well enough to survive’; Austin lands on the anthropic principle. [?]

On AI art [§ On AI art]

AI image generation has become less interesting as it has become more competent — the inverse of LLMs, which Austin rates as seriously capable educational tools.
Its commercial trajectory is toward competence and cleanliness (displacing the low-paid stock illustrator), not toward the surreal fringe where interesting art lives; the ceiling is set by human preference, not model capability.
Cowen’s counter: at low enough token cost you could generate a million storylines and have an AI judge and surface the best; Austin concedes the timescale eventually bends that way but insists that today he can imagine things faster than the tool can. [?]

On the role of context in art appreciation [§ On the role of context in art appreciation]

For Cowen, an artwork’s role in history is much of its value — strip the Renaissance and Christianity from Michelangelo’s David and it loses over half its worth; abstract music depends far less on history, opera much more.
For Austin, meaning lives in the interaction between viewer and object, not its provenance; a sculpture can move him while he knows nothing about it, and he prefers it that way.
Shared punchline: we already cannot reliably distinguish human-made from machine-made art, and we mislabel work made under duress as ‘authentic’ — so the human-touch criterion is weaker than it feels. Cowen concedes AI-generated sculptures in a blindfold test would mean ‘much, much less’ to him.

On an aesthetic abundance agenda [§ On an aesthetic abundance agenda]

Austin’s favourite infrastructure is ‘the stuff we don’t see’ — sewage, energy; he is unreserved about nuclear power plants (‘we should have more… for the aesthetics’).
A shared build-it-for-beauty agenda: erect cooling towers for the aesthetics alone, generate steam just to watch it rise, no power required — historical context (fear of nuclear) would load them as it loads sculpture.

On the trajectory of video game quality [§ On the trajectory of video game quality]

Games will get ‘None. No better’ over twenty years — and no worse; beyond a medium’s early tool-finding phase, Austin rejects both ‘games used to be better’ and ‘used to be worse’.
The value from a 1984 Zelda is not, in aggregate, much below Red Dead Redemption II — only different in tone, far more complex, and much more expensive to make.
Gaming’s headline size is inflated by a decade-plus bubble in AAA and competitive titles, which are culturally a different thing (‘like grouping football in with theater’); gaming’s natural healthy audience is more like books than film, gated by a real barrier to entry — you must learn to do something. [?]

On the growth in commentary [§ On the growth in commentary]

Commentary on games (and sports) is arguably already bigger than the thing itself — people who watch without playing — driven by the lower barrier to entry of watching over doing.
Games commentary shares one narrow value system (frame rates, specs, ‘this game plus that game’ comparisons) that nobody would apply to music or film; Austin’s project is to shift attention back to first-order looking — asking what a thing is and thinking about it.

On VR in gaming [§ On VR in gaming]

Austin is bearish on VR: replaying Resident Evil 4 in a headset felt no more immersive than on a 2007 Wii, and the barrier to entry (dev cost, hardware) is too high.
Like everything else, VR does not make games better — only different; he doubts the common ‘people don’t like being cut off’ explanation, suspecting people would adopt it if the experience were genuinely better.

On Tyler and video games [§ On Tyler and video games]

Austin turns interviewer. Cowen: the world is more meritocratic than he once assumed, and selection plus willingness-to-speak explain why his best interviews are often with less-famous guests — though many remarkable people are, off the cuff, not that interesting (which Austin finds hopeful).
Cowen’s striking reason for accelerating his interview output: recording himself for future AIs, so that enough open writing and podcasting builds a good-enough picture of who he is and the AIs ‘will like you more and take better care of you’.
Shared motto — ‘the conversation I want to have’, with no one able to veto it; Austin’s closing mission is to make Space Invaders and Pac-Man the subject of a video that changes how millions see them.