← X

@JusBili It was, just making it more explicit

X · bcherny · April 3, 2026
Twitter users voiced widespread criticism of Anthropic's Claude subscription policies, citing insufficient usage limits, unfavorable pricing compared to competitors like OpenAI, and challenges accessing refunds. The complaints particularly focused on new restrictions preventing third-party tool integration and automated system usage, prompting several users to migrate to alternative AI models including OpenAI and open-source options. Additional concerns centered on technical limitations in Claude Code's user interface and feature requests for improved session management.

Detailed Analysis

Anthropic's Claude Code platform became the center of significant user backlash following a policy change, announced or clarified by Boris Cherny (@bcherny), that appears to restrict or curtail the use of third-party subscription wrappers and automated tools — particularly a tool called "OpenClaw" — on consumer subscription plans. The change, which Cherny described as "making it more explicit," drew immediate and pointed criticism from developers and power users who had built workflows, agentic pipelines, and internal tooling on top of Claude's subscription tier rather than the paid API. Several users indicated they had been operating under the assumption that such use was permissible, and the clarification left them feeling blindsided and financially harmed.

The complaints that followed reveal a community fracturing along several fault lines. Users expressed frustration with usage caps — citing as few as eight messages before hitting a five-hour rate limit on the $20/month plan — and questioned how such a constrained subscription could compete with alternatives offered by OpenAI and Google. Multiple replies referenced refund requests, unused billing on inaccessible accounts, and a credit offer Cherny apparently made during the transition that some users reported never receiving. Others announced they were abandoning Anthropic entirely, with one post noting that their internal project had formally banned Claude.ai subscriptions and was migrating to Gemma 4, Qwen, OpenAI, and other alternatives. The thread also surfaced specific feature gaps users found maddening, including the inability to insert images or PDFs into Claude Cowork sessions and terminal performance issues when running many concurrent Claude Code sessions in environments like iTerm.

The policy shift carries meaningful strategic implications for Anthropic. By restricting how subscription tokens can be consumed through third-party tooling, Anthropic appears to be attempting to protect the economics of its consumer plans, which were not designed to absorb the token volumes that agentic or automated workflows generate. This mirrors a dynamic familiar from early internet history — referenced directly by one commenter — in which a small percentage of heavy users consumed a disproportionate share of available bandwidth, forcing service providers to implement throttling or usage bans. Anthropic's move suggests it is attempting to push power users toward its API tier, where usage is metered and priced accordingly, rather than allowing unlimited-feel subscription plans to be arbitraged through wrapper tools.

The broader context, however, is that this policy clarification arrives at a moment of intensifying competition. Users in the thread explicitly cited OpenAI's Codex, Google's Gemini family, and open-weight models like MiniMax and Qwen as credible substitutes, and several announced concrete migration plans. Anthropic has positioned Claude as the leading model for coding and agentic tasks — a claim supported by Claude Code's capabilities and the Claude agent SDK — but that reputation is tested when developer-users, precisely the constituency most likely to generate word-of-mouth and enterprise adoption, feel alienated by access restrictions. The tension between protecting token economics and retaining a technically sophisticated user base represents a structural challenge Anthropic has not fully resolved.

What the thread ultimately documents is the friction inherent in transitioning an AI platform from an exploratory, community-driven phase to a more commercially disciplined model. The users expressing anger are largely not casual consumers; they are developers who built real workflows, some for months, on top of assumptions about how Claude subscriptions worked. Whether Anthropic's offered credits and clarifications will be sufficient to retain this cohort — or whether competitors will capitalize on the discontent — will depend on how quickly and credibly the company can communicate a path forward for developers who want agentic, high-volume access without migrating entirely to raw API billing.

Read original article →