Claude Code Projects update puts it in the same layer as Grok Bot
Claude Projects now breaks a one-line request into tasks, assigns them to threads, coordinates and reviews results in the cloud, which the author compares to Grok Bot's chief-of-staff model.
Claude Code’s Projects just got an update: describe what needs doing 𝗶𝗻 𝗼𝗻𝗲 𝗰𝗼𝗻𝘃𝗲𝗿𝘀𝗮𝘁𝗶𝗼𝗻, and Claude breaks it into tasks, assigns work to threads, tracks progress, and checks the results. For a release, you could say: investigate this error, fill the gaps in our tests, and update the relevant docs. → Work goes to new or existing threads. Claude coordinates them, reviews the outputs, and brings the results together. → Threads share project memory, including changes like the release moving to Friday and requirements like who to ask before changing the billing service. → Overview lists what needs your attention, and you can follow up from your phone. → Threads run in the cloud. 𝗬𝗼𝘂 𝘀𝗹𝗲𝗲𝗽, 𝗶𝘁 𝘄𝗼𝗿𝗸𝘀. Grok Bot uses a chief of staff to coordinate multiple Bots. Codex can send work from a local task to cloud conversations. Claude has now built that coordination directly into Projects. All three are moving into the same layer: the one that manages the work for you. The agent’s scope just expanded from a conversation to a project.
This post quotes another post: “Claude Code adds parallel project sessions”