Leverage Record: February 22, 2026

About the author: I'm Charles Sieg, a cloud architect and platform engineer who builds apps, services, and infrastructure for Fortune 1000 clients through Vantalect. If your organization is rethinking its software strategy in the age of AI-assisted engineering, let's talk.

Four tasks on a Sunday. The day split cleanly: one greenfield engineering burst that scaffolded three new full-stack services from scratch, plus three architecture articles on AWS infrastructure topics. The weighted average leverage factor was 127.2x with a supervisory leverage of 370.4x, representing 142 human-equivalent hours.

The 127.2x weighted average is skewed hard by the services task (205.7x, 35 minutes for an estimated 120 human-hours). Greenfield multi-service scaffolding is the case where AI leverage reaches its upper range: the interfaces are defined, the patterns are repeatable across services, and there is no legacy code to reconcile with. The three article tasks landed in the 36x to 48x band, which is where long-form technical writing tends to settle.

Task Log

Aggregate Statistics

Analysis

The three-service scaffold (205.7x) is the high-water mark of the day. Standing up three FastAPI backends with their React frontends, wiring authentication with social login and JWT, integrating a payments provider, and creating a cert-prep application all in 35 minutes of AI time requires that the generated code adhere to a consistent pattern across services. That consistency is what a human engineer struggles with when building multiple services in parallel: decisions drift, naming diverges, and shared concerns like error handling or logging get implemented four different ways. The AI enforces the same template across all three services because it is drawing from the same prompt context.

The three architecture articles sit in the 36x-48x band, which is representative of long-form technical writing. A 600-to-800-line article on an infrastructure topic requires synthesizing the subject's documentation, constructing a narrative arc, drafting diagrams, and writing explanatory prose. None of those steps are bottlenecked by external I/O, so the AI's compression ratio is high but not extreme. The 10-to-12-minute runtimes are consistent with the article lengths: roughly 50-to-70 lines of finished content per AI minute.

The supervisory leverage at 370.4x reflects a day with low prompt overhead. The services scaffold took 12 minutes of supervisory time because specifying three services with their integration surface required a detailed brief. The three articles each took 3-to-4 minutes of supervisory time, which is typical for long-form writing where the prompt needs to establish scope, tone, and target length but does not need to specify implementation details.

The pattern visible across all four tasks: when the work is primarily generative (writing code or prose from a clear spec) and the outputs do not depend on external network calls, leverage is high. The Sunday format, working through a backlog of planned items without meetings or context switches, is a favorable environment for maximizing leverage per minute of AI time.