I used an LLM to help draft this post, but I’ve edited/rewritten it extensively and endorse it.
Epistemic status: Confident in the existence of an organizational capacity gap in Effective Altruism; seeking feedback on the tractability of proposed civilizational insurance interventions.
When the flowers arrive in early spring, it's already too late to begin beekeeping.
A successful colony requires overwintering — keeping bees alive through months when there's nothing to forage. The problem is that by the time winter arrives, the beekeeper has already extracted the honey the bees stored for survival. The wealth pump is running. Do nothing, and the colony is gone before the first nectar flow.
So you intervene. A feeding jar. A medium super to protect it. An exclusion board to maintain structure. The jar needs a lid with small holes — the feed has to be slow. Too much too fast kills the hive as surely as nothing at all. Every stable system has a narrow adaptive window. Overwhelm it and you get chaos. Ignore it and you get collapse. The beekeeper's job is to find and hold that window.
This is how I think about what's happening inside EA right now.
From Money-Constrained to Organization-Constrained
The movement is in the middle of a phase transition, and I don't think we're talking about it clearly enough.
In January 2026, Fortune reported that all seven Anthropic co-founders — Dario and Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish, and Christopher Olah — pledged to give away 80% of their personal fortunes, a commitment valued at over $21 billion. That's not a rounding error. That's a structural change in what EA is dealing with.
The bottleneck is no longer funding. It's absorptive capacity. The movement lacks the institutional infrastructure to deploy capital at this scale toward systemic stabilization. The jar exists. The holes don't.
Current philanthropic infrastructure is optimized for cause-area depth — technical AI safety, global health, biosecurity. What it isn't designed for is the foundational structural stability that makes those interventions viable in the first place. Nobody is building that layer. That's the gap.
What the Data Actually Shows
Peter Turchin's cliodynamic models — grounded in Structural-Demographic Theory and validated across 800 societies over 10,000 years — identify two primary drivers pushing the Political Stress Indicator toward what Seneca called the cliff: the point where growth was sluggish but ruin is rapid.
The first is the wealth pump. Since 1979, productivity and compensation have decoupled by a factor of 3.5x. The beekeeper has been extracting honey without replacing it. Popular immiseration isn't a metaphor — it's a measurable structural condition with a predictable endpoint.
The second is elite overproduction. When a society generates more credentialed aspirants than it has status positions to absorb, the surplus doesn't quietly accept diminished expectations. Historically, frustrated counter-elites mobilize popular discontent to undermine the social order. The forest fills with deadwood. A single spark becomes an inferno.
Both conditions are present. Both are measurable. The Political Stress Indicator is at levels not seen since the 1850s.
The Ord Hurdle and Why I Think It's Incomplete
In The Precipice, Toby Ord argues that recoverable societal collapse isn't an existential catastrophe — humanity's long-term potential survives intact.
I think this underestimates what happens during the collapse itself.
A Seneca Cliff creates a window of profound institutional fragility. That fragility isn't neutral. During rupture, the probability of genuinely unrecoverable catastrophes — engineered pandemic release, deployment of misaligned AI — increases sharply precisely because the administrative capacity to govern these technologies has broken down. The jar needs holes small enough to prevent overwhelming the hive. Remove the jar entirely and the colony doesn't make it to spring.
Reducing the Political Stress Indicator now is civilizational insurance. It's the precondition for safely navigating what Ord calls the Long Reflection — not a distraction from it.
Recursive Stability: Why Layers Matter
Here's something it took me a while to understand about beehives. They don't have a single point of failure. The queen governs the colony's genetic direction. The worker structure handles day-to-day operations. Individual cells store the resources that buffer against seasonal shocks. Each layer is governed by the one above it. Each provides a floor when the layer above it fails.
Stable civilizations work the same way. So does the architecture I'm proposing.
The root layer is governance — an independent oversight structure, modeled on principles embedded in instruments like Anthropic's Long-Term Benefit Trust, ensuring that the engine of AI wealth pursues the public good rather than narrow extraction. Without this, everything downstream gets captured before it can work.
The institutional layer is The Resilience Commons itself — the plumbing that directs windfalls into vetted, high-impact organizations. Not a grantmaker. Not a program office. A field-building infrastructure that creates the organizational homes capital needs to flow into productively.
The local layer is what I call the Micro-Republic: small, self-reliant networks with high asabiya — the social glue Ibn Khaldun identified as the foundation of stable societies. When superior layers experience turbulence, the foundational units of society need to hold. These aren't fallback positions. They're load-bearing walls.
The principle running through all three layers is simple, even if the execution isn't: sow wisdom before power. No system can absorb unlimited change instantly without catastrophic destabilization. Progress has to be staged, buffered, synchronized with the environment's capacity to integrate it. This isn't caution. It's an engineering constraint — as real in organizations as it is in hives.
What I'm Building
The Resilience Commons directs capital into three neglected areas.
The "Decent Bet" Initiative treats elite overproduction as a status problem. If skilled trades, nursing, and applied engineering carried genuine social dignity — not consolation-prize framing — they'd absorb aspirants who currently have nowhere constructive to go. More chairs. Less musical chairs.
Micro-Republic Resilience builds the local layer — self-reliant networks with shared infrastructure: solar microgrids, water filtration, analog communications. The cellular structure of the hive, built before the colony needs it.
Labor-Augmenting Innovation steers AI toward augmenting human workers rather than simply replacing them. The difference isn't technical — it's a choice. Steering that choice in productive sectors slows the wealth pump at its source.
The build sequence matters as much as the components.
Year one is epistemic grounding — validating the capacity gap thesis, establishing the fellowship pipeline, writing the code for what comes next before the liquidity event arrives.
Year two is operational proof — pilot cohorts, fiscal sponsorship, building the social muscle for large-scale coordination without premature institutional rigidity.
Year three and beyond is when the infrastructure scales to handle capital flows an order of magnitude larger than currently possible. You can't pour ten gallons through a one-gallon jar. The holes have to be ready first.
Why Me
I've spent years working through cliodynamics and structural-demographic theory — the two books linked in my profile represent that framework in full. I've also spent years teaching software to working communities, repairing appliances, keeping bees, and growing food.
That's not a resume flourish. It's the point. I've watched the gap between philanthropic ambition and community reality from both sides of it. The plumbing of daily life is where civilizational safety is actually maintained — not in papers, not in pledges, but in whether the infrastructure holds when the grid goes down.
EA's capacity gap is a plumbing problem. You can't fix plumbing you've never touched.
What I Need From This Community
Three honest questions I don't have good answers to:
Does the "Decent Bet" initiative offer a measurable way to deflate elite aspirations, or am I overstating how much status reframing can move the needle? What would a testable version actually look like?
Is absorptive capacity for large-scale AI windfalls a primary bottleneck the movement is overlooking — or is someone already working on this and I just haven't found them yet?
Is managing a cleansing phase transition a more robust strategy for long-term existential security than attempting to prevent all forms of collapse? Does framing societal rupture as an s-risk amplifier change the calculus for anyone here?
I'm building the infrastructure. Tell me what I'm missing.
James Greathouse is the author of The Science of Social Collapse and The Independent Manifesto. He is the founder of The Resilience Commons.