Scheduling the work

“The step-based approach also made it simple to run migrations on hundreds of files concurrently”

“Two planners, unaware of each other, implement the same concept in different ways in different parts of the codebase.”

“when we automatically create all of the tickets, we structure it in a manner so that the tickets are made so none of them are blocking each other and if there are ones that are blocking each other, that information is flagged up and highlighted.”

“Two agents working on two unrelated screens will still collide on the last line of that file.”

“This year, an automated dependency upgrade passed our checks, but still failed in production, impacting end users. We are responding by strengthening safeguards, expanding rollback capacity, and scheduling automated changes during owning teams’ working hours.”

“Agents only start working on tasks that aren’t blocked, so execution unfolds naturally and optimally in parallel for this DAG (a sequence of execution steps).”

“In practice, most people could comfortably manage three to five sessions at a time before context switching became painful. Beyond that, productivity dropped.”

“We don’t want thousands of loops being set up across the company without any bounds. You have a managed surface that you go to to set up the loop. So it runs on Sunday when we know we have better uh CI capacity available.”

“now you have your backlog in your pull requests. And that’s the bottleneck now because you can’t review all of that. And then there are going to be agents that review all that code and auto merge and auto deploy it, and what you get is the Homer Simpson car.”

“if you have two different tasks working on the same part of the codebase, you’re going to get merge conflicts. So you need to still like plan out scope and paralyze uh the work. Um and one like one unit of work can always be one agent.”

“Came back to find 6 PRs merged, and only one blocked on a dependency.”

“with each engineer producing three to five PRs per day, even on a team of three, merge conflicts were super miserable, right? Because these PRs tended to be pretty big, we were working on the same parts of the code base.”

“The continuous availability of code reviewers is also not guaranteed all the time, so we need to only migrate parts of the codebase where we have bandwidth to land the changes.”

“We needed to slow down some of the migration work to avoid overwhelming the teams that had to do the reviews”

“now I think the bottleneck is around just creating the work.”

“For instance, most coding tasks involve fewer truly parallelizable tasks than research, and LLM agents are not yet great at coordinating and delegating to other agents in real time.”

“And so, like, at the end of the week, I find out that they've launched like 46 game features.”

“Humans should be able to give many agents logically separate work. Agents should have clean environments and working directories: it unnecessarily wastes tokens on resolution if agents are interfering with one another’s changes.”

“If you were to fan out to a couple of hundred subagents and then tell those subagents to run the build and test of an application, what you’ll get is bad form back pressure. Thus, the instruction above is that only a single subagent should be used for validation”

“We want to get rid of the 20% of our codebase that looks and behaves differently to the rest of our codebase. If we do that, our code becomes extremely predictable. It becomes much easier to write our prompts, and also to review our code.”

“With no hierarchy, agents became risk-averse. They avoided difficult tasks and made small, safe changes instead. No agent took responsibility for hard problems or end-to-end implementation.”

“parallel agents are bottlenecked not by their code quality but by the convergence points in the codebase.”

“agents conflict. They step on each other’s changes. They duplicate work. They make inconsistent architectural decisions. And so the coordination overhead ends up eating up the speed gains all the while you’re burning tokens.”

“every Sunday, when we don’t have that much CI load, it will generate a new PR for this service using that code cleanup.”

“Recurring merge conflicts are the canary telling you that there is a misalignment between a) the work you do, and b) the type of work your architecture supports.”

“So partition by ownership before you start any agents. If you can write down who owns what up front, the branches will probably merge cleanly.”