Why I Stopped Letting One Turn Give an Agent Two Orders

Description and use

<h1> Why I Stopped Letting One Turn Give an Agent Two Orders </h1> <p>I used to think duplicate delegation was a small prompting problem. If an agent received two instructions in one turn, surely it could infer which one mattered most. That is a comforting assumption. It is also a poor runtime rule.</p> <p>While building APX, I found a more honest default: <strong>one turn gets one instruction per target agent.</strong> A second instruction to that same agent is refused unless the caller marks it as an explicit follow-up and explains whether it extends or replaces the first one.</p> <p>This is

Problem

<h1> Why I Stopped Letting One Turn Give an Agent Two Orders </h1> <p>I used to think duplicate delegation was a small prompting problem. If an agent received two instructions in one turn, surely it could infer which one mattered most. That is a comforting assumption. It is also a poor runtime rule.</p> <p>While building APX, I found a more honest default: <strong>one turn gets one instruction per target agent.</strong> A second instruction to that same agent is refused unless the caller marks it as an explicit follow-up and explains whether it extends or replaces the first one.</p> <p>This is

Outcome

<h1> Why I Stopped Letting One Turn Give an Agent Two Orders </h1> <p>I used to think duplicate delegation was a small prompting problem. If an agent received two instructions in one turn, surely it could infer which one mattered most. That is a comforting assumption. It is also a poor runtime rule.</p> <p>While building APX, I found a more honest default: <strong>one turn gets one instruction per target agent.</strong> A second instruction to that same agent is refused unless the caller marks it as an explicit follow-up and explains whether it extends or replaces the first one.</p> <p>This is

Evidence

Why I Stopped Letting One Turn Give an Agent Two Orders

Source