Steam and the attention model for games
The latest saga in people trying to give all their money to the video games industry
I keep a list of unanswered questions and curiosities which have gradually turned into life obsessions as they have refused to resolve themselves over time. One of those is the intersection between the advertising industry (who on an annual basis would like to give many billions of dollars to companies which reach Gen Z and Alpha audiences) and the video games community (who on an annual basis reach more Gen Z and Alpha audiences than any other vertical in the world). Over time I’ve ended up refining my mental model of this interaction as two people who are attempting to date but who a) have accidentally sat down at the wrong table and b) do not understand what dating actually is.
The latest version of this situational comedy is Valve taking a position against rewarded ads in games using the Steam platform. Given the sheer heft of Steam’s footprint, this has effectively banned rewarded ads in PC gaming, which seems…strong.
There are a few reasons they might be doing this, all utterly speculative.
An economic explanation might be that Valve are trying to keep user acquisition high for PC games by keeping inventory scarce which in turn keeps general pricing high? Seems like a stretch.
Another is to suggest that Valve might have a very big, very levered position in Google and/or Amazon and they would like more games-related ad dollars to continue to flow to YouTube and maybe Twitch? While Valve turning into a hedge fund would be something of a surprise, it doesn’t seem super probable.
Yet another possibility is that they are currently building their own adtech platform and they would like for nobody else to spoil the opportunity please. I would kind of love this to be true but suspect it isn’t.
A final, more philosophical point is that Valve fundamentally disagrees that people should be able to pay for things with their attention. This is worth unpacking a little. Despite the description, I’ve never really thought of digital attention as currency. It’s more a call option that you sell to advertisers who hope it lands in the money. The nice thing about rewarded ads though, is that they’re effectively a free call option, executed only when in the money (i.e. when a player chooses to watch the ads). Trading analogies aside, my point is quite simply that this format was designed to be the ultimate alignment of interests across players, advertisers and advertisers. It really seems a weird thing to want to block.
I suppose ‘look, we just really, really don’t want ads’ is one Valve response to all this. But I find it pretty bizarre to think that a developer-centric platform would want to actively prevent its own developers from being more successful. But then Gabe Newell is also experimenting with neural interfaces so maybe he’s thinking about attention models in a very different way. Either way I’m clearly missing something here so perhaps a reader (or forwardee) can explain this to me?
Other things
In a couple of ancient history circles where I lurk, there has been an argument about whether a comet impact or massive solar storm was responsible for (amongst many other things) the ending of the last ice age. A paper just published suggests, based on the dendrochronology, that there was definitely a Miyake Event at this time. The plot thickens.
The FT’s write-up of Sam Altman’s kitchen reminds me of what we lost with Valleywag’s demise.
Amidst a CEO transition and much IPO speculation, Discord celebrated ten years of existence. I still think they are the most under-rated (and under-valued) social platform in the world right now.
The story of Infinite Reality keeps getting more bizarre. Post its latest acquisition, it has rebranded as Napster. Which at this point isn’t even surprising. Although it does remove the nominative determinism joke. Surely this has hit the Matt Levine threshold?
I discovered Granola (about ten seconds before their Series B) and it seems to be the biggest AI productivity gain for my workflow since Calendly (which remains untopped, although o3 is getting close).