Concurrent Delivery is about exploring how humans and AI agents plan, coordinate and ship software together. Agents can now work in parallel; making that work fit together is the open problem. We map the challenges, the approaches people are trying, and what the evidence says so far.
The central tensionMore agents. More collisions. Time to rethink delivery.
Have you seen this happen? These are the problems practitioners keep running into once more than one agent works on the same system.
Several agents change overlapping code, files or environments, and each one's work undoes another's.
“Every agent would hit the same bug, fix that bug, and then overwrite each other's changes. Having 16 agents running didn't help because each was stuck solving the same task.”
Generating changes in parallel is easy. Combining them, verifying them and shipping them is not.
“Green is not coherent. Ten agent PRs can each pass CI alone, pass again as a batch, and still be a mess together: three flavors of the same helper, two designs pulling in opposite directions, a thousand lines no human read. A merge queue keeps main buildable. It does not keep main good.”
Developers spend their time managing dependencies, decisions and agent output rather than writing code.
“When 10 of them finished simultaneously, I had 10 results to evaluate without enough context to know which ones I could trust. AI agents don’t have a speed limit, but the person managing them still does.”
Parallel agents put pressure on every step from planning to production. These are the six challenges we track, each with the approaches people are trying.
The research moves as we find things. Here is what came in most recently.
Know a tool, paper, post, talk, or tiny experiment we should be tracking? Send the link and a sentence about why it's interesting. Contributions are welcome—even if the idea is unfinished.
