Allen Bargi · 2026-08-15 · notable
Allen Bargi — working with AI feels more like leadership than coding
Allen Bargi argues that the skill of working with AI is closer to leading people than to programming. The same prompt can give different answers, so he says sharing context and intent beats writing precise instructions.
Bargi's framing: you brief an AI the way you brief a colleague, not the way you write a function.
What is it?
Allen Bargi's note describes a change in what the daily work of building software feels like. Instead of stating exactly what the machine should do, Bargi says the useful move is to explain why it matters and what a good result looks like. He is explicit that this is not a claim that models think — it is a claim about the shape of the interaction.
How does it work?
Bargi builds the argument on one property: the same input can produce different output. A compiler is deterministic, so precise instructions are the right interface for it. A model is not, so Bargi treats reasoning, examples and desired outcomes as the interface instead. Corrections then act like ongoing feedback — saved as reusable instructions, they pull the system toward how a particular person works.
Why does it matter?
This is a practical read on why some developers get far more out of coding agents than others using the same tools. Bargi's answer is that the gap sits in briefing skill, which suggests the thing worth practising is describing intent and standards clearly. He is careful to note the limits of the comparison: a model has no lived experience and carries no accountability.
Who is it for?
developers working with coding agents day to day