Follow-up to I run 9 parallel Claude Code sessions — the bottleneck wasn't the model, it was...
He shifted memory search from model to server, injecting context into the first message—eliminating the 1-in-9 failure rate where parallel sessions ignored it. For reliable multi-agent systems: determinism goes in infrastructure, not prompts. Models handle judgment; critical guarantees require plumbing, not instruction.
Follow-up to I run 9 parallel Claude Code sessions — the bottleneck wasn't
the model, it was memory.
That post described the problem. This one is about the fix I shipped, and the
uncomfortable thing it taught me about "instructing" a model.
The setup
Claude Code writes in six or nine parallel sessions. The bottleneck is one person taking in that output and issuing the next…
AI Coding Agents Still Forget Everything — So I Built the Memory Layer Underneath Them AI...
How I built an engineering toolkit for AI coding agents I spend most of my day working...
I run Claude Code most of the day. The thing that kept biting me wasn't the model getting dumber. It...
I thought running agent sessions in parallel would multiply productivity. What I actually got, in the space of one week, was a…
A practical model for AI coding agents that need to remember projects, tasks, executions, and runtime sessions separately.
