Anthropic launches Claude Cowork: first impressions of a $100 AI agent

Anthropic Launches Claude Cowork: First Experience Using an AI Agent for $100

The field of artificial intelligence is experiencing rapid development, and companies are actively experimenting with new interaction formats. One such innovation is Claude Cowork — a specialized agent mode presented by Anthropic. This feature is integrated into Claude's desktop application for macOS and offers users a new level of workflow automation. However, as the first experience has shown, the high subscription cost of $100 per month raises questions about real value and the technology's readiness for full-scale deployment.

The emergence of Claude Cowork is linked to the broader trend of creating AI agents capable of performing tasks autonomously. Companies are striving to move beyond simple chatbots and provide tools that can independently analyze information, make decisions, and act in the digital environment. Anthropic, one of the leaders in developing large language models, has decided to join this race by offering its own vision of AI agents. Claude Cowork is positioned as an assistant that can take on routine or complex tasks, freeing up user time for more creative or strategic work. The Claude Desktop application, into which the new mode is integrated, is intended to become a central hub for interacting with advanced AI capabilities.

To assess Claude Cowork's capabilities, a practical test was conducted. The task involved creating a detailed learning plan for studying the Go programming language. This choice was not arbitrary: Go is a popular and in-demand language, and the process of learning it requires a structured approach and understanding of various aspects — from syntax fundamentals to advanced concepts of concurrency and network interaction.

It was expected that the AI agent would be able to independently analyze available resources, determine the optimal sequence of topics, and suggest concrete steps for mastering the material. However, reality turned out to be less rosy. Despite the claimed ability to work autonomously, the agent demonstrated a number of significant limitations.

The process of creating the plan required considerable user intervention, corrections, and clarifications. Autonomy proved to be more illusory than real. The agent had difficulty maintaining context, made logical errors in structuring material, and did not always offer the most relevant or modern approaches to learning.

This raises doubts about its ability to perform truly complex tasks without constant oversight and human guidance.

The implications of Claude Cowork's initial tests are quite significant. First, the high subscription cost of $100 per month seems unjustified given the current level of functionality. Users willing to pay such money expect a high level of autonomy, accuracy, and efficiency from an AI agent, which have not yet been demonstrated.

Second, questions arise about the maturity of AI agent technology in general. The ability of agents to independently perform complex workflows, especially in professional environments, remains questionable. Current implementations appear to require significant refinement and optimization before they can become full-fledged tools, rather than just advanced assistants requiring constant supervision.

This may slow the adoption of such technologies in business processes where stakes are high and errors are not tolerable.

In conclusion, the launch of Claude Cowork is an important step by Anthropic in developing AI agents, but the first experience revealed a number of issues. The technology, despite ambitious claims, has not yet reached a level that justifies the high subscription cost. Real agent autonomy remains questionable, and there is a long way to go before they can become reliable assistants in complex workflows. Users should approach with caution and not place excessive hopes on the current capabilities of AI agents, especially given their high price.