Anthropic Faces Developer Backlash Over Claude Code Performance Drop

21 articles · Updated · VentureBeat · Apr 13

Anthropic is facing backlash from developers over a perceived decline in the performance of its Claude Code AI tool.
Users report shallower reasoning, more errors, and increased token consumption, linking these issues to recent changes in effort defaults and interface transparency.
The controversy has sparked debate about transparency and trust, with some fearing it could erode Anthropic’s reputation as it prepares for a potential IPO.

Can Anthropic restore user trust in Claude Code after its performance decline and hidden changes?

How will Anthropic's future compute capacity address current Claude Code performance constraints?

What is Anthropic's long-term strategy to prevent silent performance degradation in its AI models?

Does the "real bug" in adaptive thinking signal deeper challenges for Anthropic's AI architecture?

Will Project Glasswing's success overshadow Claude Code's current reliability and "laziness" concerns?

Are "adaptive thinking" and variable "effort" the inevitable trade-offs for scaling advanced AI agents?