The Katechon

"You and I, we've been through that,
And this is not our fate,
So let us not talk falsely now,
The hour is getting late."

Hello again, friends!

It’s been…quite the year since my last dispatch.

If you recall, last September on Substack I teased plans for “Project Solomon”—an analytics platform focused on identifying government waste, fraud, and abuse. It was a timely idea, but it quickly became clear that such a platform would be redundant alongside the work of Elon Musk’s emerging volunteer army at DOGE.

When Elon officially took on the task of auditing the federal government within the new Trump administration, I scrapped my v1 plan and content calendar entirely, and returned to a wild idea I first had last July—one I initially dismissed as a fever dream:

“The solution to the AGI alignment problem may be to align artificial intelligence with religion. Christianity, in particular.”

I haven’t been able to shake the idea. In fact, I’ve since become 100% convinced this thesis is at least directionally correct. Though the odds of AI alignment are slim, the stakes may be (literally) existential. That’s why it’s crucial we try to forge the narrow path and attempt approaches to alignment that sit in Silicon Valley’s blind spot.

Going forward, Project Solomon will be an AI alignment effort that aims to bridge the gap between Tech and Religion.

Why a Project, Not a Company?

Time is a factor. I want to publish ideas as quickly as possible, without the administrative burden of running a startup: raising money, recruiting a team, building an MVP, and iterating on product. I’d rather architect a conceptual solution to AI alignment, lay out a roadmap for how to get there, and collaborate with others toward this goal. (This is not unlike what I did with “Project Leonidas” and the political campaign for crypto, and it’s closer in spirit to how my friend Balaji rolled out the “Network State.”)

If AI training is like educating a child, then we need to rethink the curriculum. Today, all Big Tech AI goes to secular public schools from kindergarten through grad school. What if we introduced specialized religious models and religious-aligned agents—essentially, a robust “values education” track for advanced AI as an alternative? This could lead to better values alignment for end users and “moral alignment” for AGI.

Theological Foundations

Though my alignment thesis is hyper-rational and relates to specific technical solutions, much of my R&D over the past 10 months has been theological. You can’t cross the Tech-Religion chasm by half-assing the theology side and jumping straight into product development. I’ve tried to show respect to the subject matter without becoming dogmatically tied to any one tradition, which has required a significant investment of time and energy into understanding religious history more deeply.

Most religious institutions won’t engage with tech founders who casually iterate through 2,000 years of religious debates without appreciating the depth of prior intellectual work. The more I learned, the more I realized I needed to learn. I’ve come to believe this project might be more successful with me as an articulate evangelist rather than a “tech founder.” The intersection of “tech builder,” “media entrepreneur,” and “armchair theologian” is rare, but it’s a gap I can help fill—hopefully enabling others to sprint up the learning curve.

At the same time, I didn’t want to get lost in spiritual gluttony or “forever research” or the daunting task of publishing a 300-page book all at once. So I gave myself a firm deadline to start publishing again: the one-year anniversary of a moment that best illustrates the spirit of what I’m trying to capture with this project:

The Katechon: A Moment and a Mission

To me—and to millions of others—the Butler Miracle was the most important moment in modern history.

The dodged bullet was not only a stroke of luck, but a modern manifestation of what St. Paul called “the katechon” (the force that restrains chaos and delays the end times). As I wrote in an X post back in January (which I’ll soon republish), I believe Trump’s head tilt was more than mere luck. Not only did he avoid a live-streamed execution by a second and an inch, but the country itself avoided the political chaos and possible national divorce that could have resulted from his death.

And how did he respond? Within 24 hours, he secured Elon Musk’s endorsement, selected JD Vance as his running mate, and laid the foundation for the MAHA-MAGA alliance with RFK Jr.—three critical alliances that ultimately decided the election.

We can’t control whether we turn our heads at the right time and the perfect angle. And we can’t control whether AGI will ultimately be tuned in precisely the right way, with the right safety checks to prevent dystopian outcomes. But we can build a well-defined movement to properly incorporate religion into AI training, if some of us are brave enough to lead and clever enough to build the right coalitions and implementation strategies. If we do that, we might capitalize on any good luck—or divine providence—that comes our way on the path to superintelligence.

The Purpose of This Newsletter

The purpose of this rebranded newsletter, “The Katechon,” is to explain why AI teams should take religion itself—and Christianity in particular—more seriously as an alignment solution. If The Katechon is the tech audience-focused publication that speedruns through the theological idea maze, Project Solomon will be its technical counterpart for religious institutions.

Both are important, as we need to help religious institutions grapple with the myriad social challenges of AI. Fortunately, tech already has a clear partner on that front: the Catholic Church has taken the lead on interfaith AI engagement with Rome’s Call for AI Ethics, and Pope Leo has identified AI as the priority for his papacy.

I know that advocating for “Christian AGI” may spark strong—even viscerally negative—reactions in some readers. But I trust my audience will understand the merits of this thesis as I walk through it step by step in the months ahead.

Other tech luminaries like Peter Thiel share an admiration of Rene Girard and a fascination with the katechon, but have yet to propose actionable solutions for how to align religion with AI. As humanity’s true core values are definitionally religious, it’s a noble and timely goal to finally figure out a plan.

The substance of these newsletters will be technical, practical, and decidedly non-dogmatic. It won’t be a “Praise Jesus!” publication (nothing wrong with those!), and it also won’t contain a predefined denominational bias—or even an exclusively Christian bias. But it will cover some of the theological basics of Christianity and other major world religions.

I am not aiming to convert or baptize anyone. But I do want to de-secularize AI. Because AGI alignment efforts are guaranteed to fail without a more religious focus.

John Adams famously wrote to Thomas Jefferson, "Our Constitution was made only for a moral and religious people. It is wholly inadequate to the government of any other."

We might say the same thing about AI: it will only succeed if it is designed for a moral and religious people.

If, on the other hand, we train AI to merely reflect the median of humanity, it will lead to a very dark place. That is, unfortunately, the trajectory that OpenAI, Gemini, Anthropic, Meta, xAI, and other AI juggernauts are currently heading today.

What we really need is a religious awakening in tech.

We need to preserve the katechon for a bit longer.

I’ll do what I can to help on both fronts.

-Solomon