Anthropic's Claude Now Runs a Team of AI Assistants for You

Anthropic, the company behind the Claude chatbot, has relaunched a feature called Projects inside Claude Code, its tool aimed at people who write software. The idea is simple: rather than asking one AI to do everything, you now oversee a small crew of AI agents. Each agent is a separate helper that can read, write, and carry out tasks — and they all share the same memory, goals, and folder of files. Think of it less like one assistant and more like a small team that's been briefed on the same project.

Here's how the work gets split up. Inside each project, there are 'threads,' which are individual tasks running at the same time. A 'coordinator' agent keeps an eye on everything and decides who does what. If two threads accidentally change the same piece of code, the clash gets flagged and sorted out the way human programmers already handle it — like two editors marking up the same paragraph. Threads can also break big jobs into smaller pieces, so large assignments finish faster instead of waiting in a single-file line.

The catch is that this is aimed squarely at coders, not the average person. At launch, everything runs in the cloud rather than on your own computer, which means your work and files live on Anthropic's servers — a privacy consideration worth noting. Anthropic says support for local tools is coming 'very soon,' but it isn't here yet. And access is limited: right now it's in beta for 'select' Claude Pro and Max subscribers, with everyone else waiting.

The bigger picture is that AI is shifting from a single chatbot you type at to a managed workforce you supervise. Competitors like Grok are building similar team-of-agents tools. Even if you never touch Claude Code, this is the direction the whole industry is heading — and the skills involved are less about writing code and more about delegating, reviewing, and catching mistakes. The real question isn't whether AI can do the work, but whether you can keep track of it when it does.