I’d let this thread seep into your cells before we continue.
Ok, now you are ready.
[Only do this hands-free via Bluetooth. Safety first.]
While driving to pick up my kids, I interrupted the podcast that was playing to launch Grokbot’s voice mode to ask about TLT volatility (a random market idea popped into my head). The bot is connected to moontower.ai data via MCP and, of course, the internet and its own training data. Before I know it, I’m at the destination and never switched back to the podcast.
This felt weirdly profound.
Podcasts have an infovore following. But can anything compete with an infovore discussing topics they care about and getting to ask questions and guide their own learning? AI + voice feels like the seed of a trend that is not totally visible to everyone. A ”conversational AI custom media eats traditional media” thesis.
If this is a thing, it would leave less attention for the businesses and individuals that monetize attention. Publishers sell page views, minutes, subscriptions, which are all quantities that are produced once, then delivered at zero marginal cost. But custom, conversational media runs on tokens and compute.
Attention is zero-sum bounded by hours and my guess is individualized media means more properties are sucked into the black hole of the “middle”. The “hollowing of the middle” has been the ruthless evisceration of saccharine mass market media. You either want to own a hyper niche or be broadcasting the Super Bowl. The middle is the failed Vice magazine which outgrew an edgy audience but was not large enough to scale its fixed costs.
I’m not planting any new flags here. It’s just part of my gradual awakening to how these powerful technologies can reshape behavior, in turn, altering and replacing business models. The investing question is not “can” but “will”.
An obvious counterargument against this is that most people are consumers and lurkers. They want to scroll and channel surf. They don’t actually want to steer if entertainment is an escape as opposed to an outlet.
That’s all valid and fair. But it’s also consensus (for good reason). But weird things happen when we remove friction. I’m old enough to remember ordering a pizza on the phone before DoorDash existed.
Games are an interesting category in that they’re a hybrid of active and passive. We don’t generally make our own games, but we enjoy active participation in our couch leisure, whether it’s Catan, GTA, or D&D. In Games: Agency as Art, the philosopher C. Thi Nguyen argues that games are the art form whose medium is agency itself. The designer shapes what you can do and what you’re trying to do, but the specific experience depends heavily on what you bring to it. More examples in the same vein:
-
User-made games: Minecraft and Roblox, where players build the content.
-
Fan fiction and modding: people taking a finished story and making their own version.
-
Musical hobbies: Everything from performing an instrument or dance to curating playlists
Choose-your-own-adventure shows like Netflix’s Bandersnatch never caught on. Perhaps people like agency when it’s the point and not when it interrupts a story. But again, this is consensus.
On the other hand, if the friction to self-expression is reduced to the ease by which a thought crosses your mind, like it did when I was driving, this is a glimpse into an unimagined, non-consensus future replete with multi-order cascades that can fuel a substantial engine of investing and business opportunity.