One last little rabbit hole: "The Old Stories: Moses" is being positioned as a prequel to the Wonder Project's series House of David — or, more bizarrely, as part of the "House of David cinematic universe." That accords with a certain brand of evangelicalism that is inordinately focused on Israelite kingship. It's probably not overstating it to characterize that as a vein of monarchism in modern evangelicalism. Which is interesting in its own right, but then you pair it with their particular brand of patriotism. There's a kind of parallelism, for example, between House of David and their other big production, Young Washington, both of them biopics about the ascension of national leaders to both adulthood an authority. That pretty quickly escalates into a view of American exceptionalism as a revival of Israelite monarchism. So once again we find generative AI being embraced by executives with tendencies toward authoritarianism.

That may be because, while it seems logical that evangelical pop culture is meant to evangelize, much of it functions rather as a substitute. They may tell you that the point is to convert the unbeliever, but the product is better geared toward giving evangelicals alternatives that can pass as worship while still feeling like pop culture. In that regard, it's more important that the result looks like it was meant to compete with a Nolan movie than that it does something interesting on its own merits.

For example, the beginning of the "Moses" trailer could pass for a tense, quiet drama about Moses grappling with theophany and the divine commission placed upon him. You could make that story on a soundstage, without all of the VFX, without even attempting to pass it off for a location shoot, and might end up with a movie that is both more interesting and more effective as a tool of outreach. But something in the culture or personalities of evangelical filmmakers won't let them make that movie.

Part of what I find so fascinating about all of this is how little reflections there seems to be in evangelical filmmakers' embrace of technologies like AI. There's no debate or even inquiry into how it meshes (or doesn't) with Christian values, because the embrace is driven by prerogatives that override any concern that AI might be unethical or sinful. It's similar to Christian rock in that so many choices are decided by the imperative to meet popular culture on its own terms.

Anyway, "we shot it in a week" is misdirection, in a way. Evangelical filmmakers have been shooting quickly and cheaply on sound stages for decades, because they're driven to adopt any innovation that promises blockbuster production values on arthouse budgets. The advance here is that AI enhances that approach for a kind of heightened verisimilitude — not because it's better than artist-produced VFX, but because, for a limited time only, it's relatively cheap compute: https://merveilles.town/@lrhodes/116859028925045960

And since they're big time patriots, making movies like American Underdog and Young Washington, of course they'd partner with an American AI company like… <checks note> Chinese government-owned Kling AI. Which, irony and free-speech concerns aside, may limit the "Moses" series' value as an illustration of the affordability of AI filmmaking, if sovereignty concerns lead states to start limiting where companies can get their AI services.

Okay, context. Erwin runs the Wonder Project, a "faith-based entertainment" company specializing in movies and television with Christian evangelical and patriotic American themes. Their four biggest films have reportedly grossed $150M worldwide —more or less what Freakier Friday grossed on its own last year. So they're in a relatively big niche, but still a niche playing for margins. And unlike other niches (say, Troma-style horror) Christian entertainment companies are motivated by respectability — albeit, a very Hollywood brand of respectability. They want to look like a big budget action spectacle but spend like a mumblecore art house flick. Given that dynamic, it's no surprise that an evangelical company is one of the first to go almost-all-in on AI.

And look, this stuff is obvious to me in part I have a background in film and video, but as lay audiences grow more and more familiar with it, they'll start to notice the seams as well. CG effects that blew people's minds 30 years ago look hokey as hell now because the novelty has worn off. So if generative AI is going to establish a firm foothold on this industry, filmmakers have to get smarter about how they use it. They have to make better use of it than this.

I do think that filmmakers will get more sophisticated about their use of this tech — provided that the cost doesn't balloon out of control, which is a distinct possibility. But a lot of the problems that crop up from video generation can be avoided by, say, using ML to generate static 3D models and then paying human artists to animate on top of that. I could see using that approach to create, for example, wide angle establishing shots of ancient cities. I don't like it, but I can imagine it being less obvious.

I think part of the problem here is just the process. When you have VFX artists making, say, a pillar of fire, and the effect isn't working, you can have them go in and tweak specific parts and wind up with something better. But when your whole process is "prompt the AI, check the output, try a different prompt," then even small changes mean you're basically re-generating the entire shot, with unpredictable results. And if price is tied to compute, those changes eat into your budget quickly in a way that salary work might not. And I think you can see evidence of that on screen in that way that some shots recur with variations that are pretty obviously just slightly different prompts, but that the filmmakers decided to use anyway, counting on quick editing to keep viewers from noticing how things don't quite match up.

First, let me acknowledge: Someone did a bang up job cutting the trailer for "The Old Stories: Moses." Based on that alone, you'd think AI was up to the job.

It isn't. I'm not gonna sit down and watch an episode, but I watched the "parting of the Red Sea" scene on YouTube. It's an improvement over that weird patriotic web series that Darren Aronovsky produced, but mostly because the flaws are less obvious when you keep the focus on human actors. Sometimes, that's literal: There are lots of shots where the actor is the only thing in focus, but that's most unnoticeable because shallow focus is so common in modern cinematography. When they're in focus, the generated parts are still full of continuity errors and uncanny effects, and the editing is choppy and repetitive to cover up for video generation's tendency to get sloppier the longer it goes.

Morbidly fascinated by this article that @xgranade posted yesterday: https://www.latimes.com/business/story/2026-05-13/why-house-of-david-director-thinks-ai-can-save-hollywood-jobs Obviously, the article is leaning heavily into the hype, and its framing (they filmed an epic in one week!) is highly suspect, not least of all because principle filming almost always takes less time than pre- and post-production. But there are multiple dimensions to this, and each one is a potential rabbit hole.

apparently discord's ai moderation is banning people for posting grid patterns

https://www.reddit.com/r/OutOfTheLoop/comments/1umdc67/comment/ovbceiy/

And the aforementioned edition of Second Breakfast: https://2ndbreakfast.audreywatters.com/financial-nonsense/ makes the further point that the demand that has driven actual AI use so far can be understood as demand for underpriced compute. How the companies were generating output almost doesn't matter to the economics of AI, because their users were effectively cashing in on work done below the actual cost of doing work. That makes it clear just how unsustainable the industry is, as it currently exists.

There's a really great point in this post from May: https://davekarpf.beehiiv.com/p/what-sort-of-ai-bubble-are-we-in (h/t Second Breakfast) that our estimation of the demand for AI is based almost entirely on a market that assumes the cost is effectively $0. Adjust the price to something that's actually profitable (or even just break-even), and a lot of that demand will evaporate. Which we're beginning to see now.

this serious eats explainer on toasted sugar is blowing my mind. tldr sugar doesn't melt under heat to liquify and become caramel, instead it's going through thermal decomposition. as a result you can toast sugar without melting it, just at a lower temp. https://www.seriouseats.com/dry-toasted-sugar-granulated-caramel-recipe

Learn to value your privacy more than tech companies do.

Callback to something I posted a few months ago: Libraries spent decades crafting policies and best practices to protect data about their readers because they understand how much your reading habits can reveal about you. Then we go and volunteer all of that data (and more) to for-profit companies with a financial motive to exploit everything they collect, in exchange for some pretty graphs and middling book recommendations.

It's cool that Kobo now has StoryGraph integration, I suppose, but I fundamentally don't trust cloud-based book tracking apps — especially those with an AI component. They're all data sinks, and most companies are going to do something unapproved with that data, or get bought out by a parent company that will.

⁂