Dispatching work to run concurrently is the easy half of parallel execution, and it is the half everyone thinks about. A single control point splits into several branches, each branch runs at the same time as the others, and the total time drops from the sum of the branches to the length of the longest one. That part is close…
A retry is the most ordinary thing an autonomous system does with a failure, and the one a tool is least often built to survive. When a call comes back failed, the model may simply issue it again, on its own judgment, with less context than the engineer who wrote the tool and no way to know whether a second…
Retrying is the most reflexive move in error handling and the least examined. Something failed, so try it again. The reflex is defensible often enough that it survives without scrutiny, and that is exactly the problem. A retry is not a neutral safety net. It is a wager that the thing which just failed will succeed if attempted again, and that wager…
A tool definition is normally understood as a way to make something happen. You declare a function the model can call, it emits the arguments, and something on the other side runs. For pulling structured data out of text, invert that picture entirely. Define the tool, let the model call it, and let nothing on the far side ever run. The call itself is…
Every prompt you send a model is really two documents fused into one. There is the part that is identical on every request in a session, and the part that is different every time. The single most consequential decision in prompt architecture is where you draw the line between them, because that line determines what the model treats as…
Long conversations outgrow the window that holds them. A session that runs for hours, an agent that works a task across dozens of turns, a support thread that keeps accumulating, all of them eventually carry more history than the model can take in at once. The standard answer is to compress the older part of that history into a shorter…
Once a system holds more than one agent, the design question quietly shifts. It stops being about what any individual agent can do and becomes a question about the connective tissue between them: who hands work to whom, where authority sits, and what any single node can observe about the whole. Borrowed from distributed systems, the word for that arrangement is topology, and…
A pipeline that runs its steps in a strict line is the default most systems settle into, and rarely because the work demanded a line. A line is simply the easiest thing to write: code that calls one function, then the next, then the next, reads back as a process and mirrors how anyone would narrate the task aloud, so steps keep…
When a tool succeeds, its return value is data. When a tool fails, its return value is an instruction. An autonomous system reads the failure and immediately decides what to do about it: try the same call again, change the arguments and retry, hand the problem back to a person, or abandon the path and take another. That decision is made by the model, in…
The two directions of travel are not symmetric, and that asymmetry should decide how you start. Grafting autonomy onto a working constrained system is the easy direction: you wait until a real input exposes where the fixed logic gives out, then hand that one decision to a model. Taking autonomy back out of a running system is the hard direction, because…