Scalable Knowledge Work

Scalable Knowledge Work is a concept articulated by Jan_Wilmake in a 2024 blog post ("Scalable Ideas") describing how large language models shift knowledge work from a one-off activity to a globally replicable pipeline. The claim: for any task that requires expertise but follows a repeatable pattern, an LLM-driven API chain can now execute it millions of times with minimal marginal cost.

The Core Thesis

Jan's formulation: "We're entering an age where we can start copying knowledge work. What I mean by that is that we can find a useful thing to do once and do it on a global scale, potentially millions of times, to create value."

The enabling conditions:

1. LLMs can interpret high-level natural language instructions without rigid input schemas
2. APIs make the outputs of individual steps composable
3. Each step in the pipeline can be a hosted serverless function — independently developed, independently scalable

The GitHub Website Example

Jan used a concrete example to illustrate the concept:

Task: "Find popular GitHub repos that don't have a website yet, and build a compelling landing page for them with a CTA back to the repo. Host this on Vercel and make an issue in the repo named 'I made a website for you!'"

Broken into subtasks:

Each subtask becomes an independently hosted serverless API. Chained together as an ActionSchema, the whole pipeline becomes a single invocable action — deployable at scale across every qualifying repo on GitHub.

Bottleneck: Task Decomposition

Jan identified the key constraint as not execution speed or API cost, but task decomposition knowledge: understanding what the sub-APIs are, what each step requires, and how to chain them. This is where expertise matters — and where tools like openapisearch and ActionSchema's OpenAPI proxy add leverage, by making the space of available APIs discoverable and the wiring between them lower-friction.

Relationship to Adjacent Concepts

Scalable Knowledge Work is the economic thesis that motivates several of Jan's infrastructure projects:

• ActionSchema provides the pipeline execution layer (API chaining, OpenAPI proxy, agent coordination)
• Agent_Friendly_Web provides the data access layer (token-efficient mirrors of major platforms)

Together, they form a coherent picture: the web becomes a substrate for automated knowledge work, and the infrastructure for accessing it must become agent-friendly.

See Also

• ActionSchema
• Agent_Friendly_Web
• Jan_Wilmake