Scientific Management For AI Workers

We need Scientific Management (Taylorism) for AI coding agents. Specify-then-verify isn’t enough. Agents still decide how to implement the specification, which produces inconsistent results even with the best models. LLM-generated code that demands constant rework is a waste, even when the LLM does the rework. Scientific management can get high-quality output the first time.

The core idea: managers study the work, break jobs into small tasks, identify the best method for each task, then train and supervise workers to follow that method. Sucks for humans. Great for AI. And we already do most of it.

We encode best practices in frameworks. We break features into small tasks. And within a specific programming language, application framework, domain, codebase, and business stage, there is one best way to do something.

Take adding a feature: there’s a specific way to add a DB column, update a service, add a controller, update the UI, and decide what and how to test. Each of those steps breaks into smaller steps, each with a defined method.

I’m still experimenting with how to apply this broadly, but I’ve already experienced that small task with exact instructions produce consistent, reliable output from cheaper open-source models. Over time, the best way to enforce best practices will be to embed them in the codebase itself. LLMs are amazing pattern-matching machines. Give them the best patterns you can.

Giving an agent a specification and a verification method leads to a bunch of wasted time and tokens(money). The agent falls back on heuristics from its training data, and those are almost never the best approach for my specific context. If I break the work into small tasks and tell it exactly how I want it done I get reliable, high-quality output the first time. With rigourously defined tasks and instructtions on how to accomplish them I can find the cheapest model that follow my instructions. Micro managemet sucks for humans but its great for AI.