I Cancelled Claude: Token Issues, Declining Quality, and Poor Support (nickyreinert.de)

In a recent post, a user shared their disappointing experience with Claude Code after an initially positive start. While the first few weeks showcased impressive speed and adequate token allowances, a spike in token usage and unhelpful automated customer support led to frustration. The user noted a decline in performance quality, evidenced by subpar code suggestions and an inefficient handling of project continuity, which resulted in unnecessary token depletion. The experience was compounded by confusing communications around usage limits, including an unexpected monthly limit that had not been previously documented. This situation highlights significant concerns for the AI/ML community regarding the scalability and customer support capacities of AI platforms. As more users adopt these technologies, the need for robust support systems and transparent communication about usage policies becomes crucial. The user ultimately decided to cancel their subscription, pointing to Anthropic's challenges in managing a growing user base and the importance of maintaining quality and effectiveness in AI services. This case serves as a reminder for developers to address user feedback proactively and ensure that AI products can sustain performance under increased demand.