My focus this week has been on building out more charts to give us a better picture of how Zed is doing in the wild and where we need to focus our attention to improve it. Once that was wrapped up I spent most of the rest of the week tying up loose ends of a set of performance improvements Max and I worked on last week. This week I continued to refine my thinking on how Zed's vision for interleaving conversations and code connects to large language models, which Nate discussed a bit above.