Anthropic: The Fall from Grace in the AI Development Community
Claude Code and Claude Opus 4.6 initially heralded a golden era for developers, offering a brilliant AI agent and model that quickly overshadowed OpenAI. Anthropic, once the darling of the AI space, has since faced growing criticism for changes perceived as detrimental to developers.
The Triggering Incident
On April 2, 2026, Stella Laurenzo, AMD’s Senior Director in AI, published a pointed critique on Claude Code’s GitHub repository. She titled her analysis “Claude Code is useless for complex engineering tasks with February updates.” This review assessed nearly 6,600 Claude Code sessions and revealed significant performance degradation, leaving users questioning the reliability of the tool.
Deteriorating Performance Metrics
Laurenzo highlighted critical statistics over multiple periods. From January to mid-February, the model adeptly read 6.6 files for every file edited; post-March, this efficiency plummeted to 2.0 files read. Moreover, code changes made blindly surged from 6.2% to an alarming 33.7%. The average visibility of reasoning also underwent a steep decline—from 2,200 characters to just 600.
Anthropic’s Response to Criticism
In response to the backlash, Boris Cherny, responsible for Claude Code, announced changes made to improve “adaptive reasoning” and adjusted effort levels for better performance. While Cherny claimed these changes would not affect the models adversely, skepticism remains high among users who believe the model’s quality has, in fact, diminished.
The Computing Capacity Crisis
Amidst rising demand, Thariq Shihipar from Anthropic acknowledged that session limits were being adjusted to 5 hours during peak usage. This was supposed to affect only 7% of the most intensive users; however, numerous complaints suggest that the broader user base is feeling the impact. too.
Awareness of ‘Nerfing’ Accusations
The term “nerfing”—used to describe deliberate performance degrading—has permeated discussions surrounding Anthropic. Users have decried the perceived downgrades on various social media platforms, prompting Chirny to assert that these characterizations are unfounded.
A Shift in Pricing Models
Anthropic has also shifted its pricing structure for Enterprise plans. Previously, users paid a flat subscription fee, but now they face a base rate plus consumption costs based on API usage. This change could significantly increase expenses for heavy users, rendering the platform less accessible.
Industry-Wide Trends
Anthropic’s distress signals are part of a broader trend in the AI industry. Competitors like OpenAI are also adjusting their pricing models and capacity limits, highlighting a collective struggle in accommodating skyrocketing demand. This trend suggests an overarching issue: the increasing resource constraints across AI platforms are leaving developers and users feeling squeezed.
Conclusion: The Open Bar is Over
As demand for AI services continues to surge, companies are scrambling to recalibrate their offerings. For developers, this translates into either diminished capabilities or increased costs, raising pressing concerns about the future of AI development. The once inviting landscape for AI tools has morphed into a battleground where developers need to tread carefully to avoid falling prey to the tightening grip of major platforms like Anthropic.
With these shifts, the golden era of developers seems to have dimmed, leading to a landscape fraught with uncertainty and frustration.