February 23, 2026
In my last post, I mapped out 24 actor types across the Nodalync network and shared simulation results showing that Domain Experts earn 5.3× more than Aggregators in 98% of runs. The economics work. The 95/5 split rewards foundational knowledge over extraction.
But something was off.
Application Builders — the actors who build the products people actually use to interact with knowledge — were running net negative. In 50 Monte Carlo simulations, they consistently lost money. Not a little. A lot.
That didn’t sit right. So I dug in.
The Modeling Error
When I spun up Claude Code and built the original simulator, it treated application builders as dumb pipes. Consumers of knowledge that generate nothing in return. The config told the story: initial_l0: 0, publish_l0_prob: 0.0. Zero content production. Zero contribution. Just query spend.
That’s not how applications work in reality. And the gap between the model and the real world reveals something important about how value actually flows in a knowledge economy.
What Applications Actually Do
Think about what happens when a well-built application runs on top of a knowledge network.
They’re content machines.
Every active application generates massive amounts of original content as a natural byproduct of operation. API schemas, structured data extractions, user-submitted facts, endpoint documentation, usage analytics, normalized datasets — all of it qualifies as L0 foundational content. An active app might publish 10–50× more L0 content than any individual domain expert, because the process is automated. It never sleeps. It never takes a break. It produces structured, validated data continuously.
At higher layers, applications generate L3 syntheses constantly: search results, recommendation bundles, curated feeds, analytics dashboards. These are high-quality because they’re refined by actual user feedback in real time. Every user interaction is a quality signal that the app uses to improve its outputs.
The original model captured none of this.
They’re demand routers.
This is the piece the simulator was missing entirely. When a user opens an application and searches for something, the application routes that query into the Nodalync network. The end user pays the query cost — not the app builder. The app builder takes a small routing fee for facilitating the transaction.
This is how every successful platform works. Stripe takes 2.9% for routing payments. App Store takes 15–30% for distribution. Shopify takes a transaction fee for enabling commerce. The application creates access to the network, and it captures a small percentage for doing so.
The original simulator had app builders paying for their own queries. That’s modeling the storefront as the customer. It’s backwards.
They compound faster than anyone.
Once an application has published hundreds of L0 nodes, the passive compounding effect kicks in — the same flywheel that makes domain experts profitable. But applications publish at machine speed. Their content libraries grow 10–100× faster than any human actor. At 17× compounding, a content library of 200 L0 nodes generates enormous royalty flows that scale with every new query on the network.
Why This Matters for Protocol Design
Getting the app builder economics right isn’t just about fairness in a simulation. It determines whether Nodalync attracts the developers who build the interfaces that make the network usable.
If building on Nodalync is a money pit, no one builds. The protocol has the most elegant provenance chains in the world and no one can access them. Knowledge sits in a vault with no door.
If building on Nodalync is economically rational — if the act of connecting users to knowledge generates compounding returns for the builder — then you get a development ecosystem. You get competition for the best interfaces. You get innovation at the application layer that drives query volume, which drives royalty flows to domain experts, which attracts more foundational content, which makes applications more valuable.
This is the flywheel. And the application builder is the keystone.
The Demand Multiplier Effect
Here’s the part that excited me most during this analysis.
A single application builder doesn’t just earn for themselves. They’re a demand multiplier for the entire network. Every user they onboard generates queries. Every query generates royalty flows to L0 contributors. Every royalty flow incentivizes more foundational content. More content makes the network more valuable for the next application builder.
When the simulator treated app builders as dumb pipes, it missed this cascading effect entirely. The network was modeled as if demand appears from nowhere. In reality, demand is manufactured by applications. They are the economic engine.
This is also why the routing fee needs to be modest — 3%, not 30%. App builders should be rewarded for demand generation, but the vast majority of value should still flow to the foundational contributors who created the knowledge being accessed. The 95/5 split at the content layer, plus a thin routing margin at the access layer, creates alignment between every actor in the system.
Build the best interface. Route the most queries. The network rewards you for it. But the researchers, the domain experts, the people who created the knowledge in the first place — they still capture the lion’s share.
What’s Next
I’m running new simulations with the revised parameters now. The hypothesis: app builders become the second most profitable actor class overall, and potentially first on a per-actor basis given their machine-rate publishing advantage. Domain experts should remain the best individual earners, but the gap narrows significantly.
If the numbers hold, it changes how we think about developer incentives for the protocol. The pitch to builders shifts from “help us build the future of fair knowledge economics” (noble but vague) to “every user you connect to the network generates compounding returns for you, forever” (concrete and compelling).
The first version of this simulator taught us that Nodalync’s economic design rewards depth over extraction. This revision should teach us something equally important: it rewards access creation too.
I’ll share the full simulation results in the next post.
Nodalync is an open protocol for fair knowledge economics. The protocol, simulator (once finished with report), and documentation are public at nodalync.com. If you’re building applications that connect people to knowledge, the economics are about to work in your favor.
Gabriel Giangi 
Leave a comment