This week a games journalist I follow, Jamin Warren, published a piece on Kill Screen arguing that capitalism corrupts play. He used Tottenham Hotspur getting relegated as a case study and cited Roger Caillois's 1958 book Man, Play, and Games.
My immediate reaction: wait, doesn't KmikeyM test the exact opposite hypothesis?
So I did what anyone with an Executive Producer agent would do… I told Thalberg.
The rabbit hole
Thalberg is my AI production agent. I gave it the Kill Screen article and my half-formed thesis: what if financial structure doesn't have to corrupt play? What if you build it inside the magic circle?
Within a few minutes we had an outline. Six sections. But the outline kept referencing
Caillois's arguments secondhand from Warren's article, from Wikipedia, from general knowledge. I wanted the real thing.
I already had a PDF of Man, Play, and Games in my eBook reader. I pointed Thalberg at it.
200 pages in 5 minutes
As we discussed the outline I asked: "Does Caillois have anything to say about poker?" Thalberg read the PDF and came back with four specific pages where Caillois discusses poker…
page 18 where he calls it an agôn-alea hybrid
pages in Chapter IV where he defines corruption
page 54 with his classification table.
Direct quotes, page numbers, context.
That would have taken me an afternoon of reading. Maybe a full day if I kept getting pulled into other chapters (which I would have, it's a fascinating book). Instead, I had the exact passages I needed while my argument was still hot (let’s be honest, I was procrastinating).
The key find: Caillois doesn't say money kills play. He says the corruption happens when money stops respecting the rules of the game. He explicitly treats poker, real money, real stakes, as legitimate play. This completely reframed my essay. Warren was using Caillois to argue that capitalism corrupts play, but Caillois himself is more nuanced than that.
I would not have found this without “reading” the actual book. And I would not have read the actual book today without Thalberg.
The back-and-forth
Here's what AI-assisted writing actually looks like for me. It's not "write me an essay about play theory." It's a conversation where the ideas develop between us:
Thalberg drafted Section I. I read it and said "this is a bit generous" about the opening line. We cut it. I rewrote, Thalberg gave me notes.
I suggested poker staking as a bridge between Spurs and KmikeyM. Thalberg structured the gradient.
I asked: "since money isn't real, couldn't we argue that capitalism is itself a magic circle?" Thalberg ran it against Caillois's six criteria and found capitalism scores 3 out of 6. KmikeyM scores 6. That became the structural spine of the essay.
The ENIC/pizza metaphor came from me pushing back on Thalberg's framing. Thalberg said ENIC were "outsiders corrupting the game." I said no… they're more like players who got distracted. Like someone playing Monopoly when the pizza arrives. Thalberg rebuilt the section around that.
The ending with the line "not a contagion but an inoculation," emerged from me saying the essay shouldn't end with victory or with "the experiment continues." I wanted something about how I can't always tell what's real and what's performance. The inoculation metaphor came out of that tension.
None of these ideas came from a prompt. They came from a conversation. The AI's job wasn't to be creative. It was to be fast enough that my creativity didn't have to wait.
The speed
Outline to published blog post, distributed across Twitter, Bluesky, Threads, and LinkedIn all in one session. That includes reading the source material, pulling direct quotes from a 200-page book, drafting six sections, running a copyedit pass against Strunk & White (a skill i grabbed that someone else made), and writing social copy for four platforms.
I would not have done this in a day without Thalberg. I definitely couldn't have done it while hit with allergies and recoving from a cold.
The takeaway
AI didn't write this essay. It sounds like me and the refrences are mine. The XCOM example, the "rage quit" aside, the pizza metaphor. None of that is AI.
But I wouldn't have written it without AI. Not today. Maybe not ever. The barrier to writing a serious response to a serious piece of games journalism is the research: reading the book, finding the relevant passages, getting the quotes right. Thalberg eliminated that barrier and we had a fun argument about the essay.
The best description I have for what happened: Thalberg did the research at a speed that let me stay in the creative flow instead of breaking it to go hunt for quotes. By the time I might have gotten distracted or lost momentum, the post was done, and it changed significantly from the first draft
Read the full essay: The Contagion of Reality
