skip to content
Topic

system-design

5 essays on this topic.

Papers
  1. A Skill Is Not a Prompt

    The useful unit in agent systems is not a better instruction. It is a tested capability package: judgment, code, checks, routing, and boundaries.

  2. The Test Before the Output

    The line between tool and skill is whether you can write the test before seeing the result.

  3. Judgment Is a Moving Boundary

    The line between tool and skill isn't a property of the task. It's a property of how well you understand the task.

  4. Skills Should Die

    Every AI skill should be trying to make itself unnecessary. The ones that survive are the ones that haven't been understood yet.

  5. The LLM Is the Tool

    When the transformation is predictable, the LLM is just a runtime. A cheaper, more flexible runtime than custom code.