- AMD AI Director Says Claude Code Lost Performance in February 2026 Update
- “Claude Code cannot be trusted” on complex tasks per thousands of coding sessions
- Anthropic says it reduced effort to medium, but Teams and Enterprises can achieve a high level
AMD AI Director Stella Laurenzo claimed that Claude Code has become less effective since February 2026, arguing that it “cannot be trusted to perform complex engineering tasks.”
Laurenzo’s criticism is not unfounded, based on the company’s analysis of more than 6,800 coding sessions, almost 235,000 tool calls, and almost 18,000 reasoning blocks.
“All the senior engineers on my team have reported similar experiences/anecdotes,” Laurenzo wrote, noting that stop-hook violations (in which Claude gave up early, shirked responsibility, or asked for unnecessary leave) increased from zero in early March to about 10 per day later.
Article continues below.
Claude Code is getting worse, warns AMD director
In a GitHub post, user stellaraccident (aka Stellar Laurenzo) identified a strong correlation between the introduction of redact-thinking (redact-thinking-2026-02-12) and a decline in performance on complex tasks. The AMD director maintains that extended reasoning can “support load” for advanced engineering work.
Laurenzo also observed a shift from research-first to editing-first behavior, resulting in lower quality code, poorer compliance with conventions, and overall reduced reliability for long sessions.
Anthropic has already responded to the investigation with a multifaceted explanation. Boris from Claude Code explained that the redact-thinking-2026-02-12 setting only hides the UI reasoning and doesn’t actually reduce it.
The company also introduced adaptive thinking with Opus 4.6, with its model dynamically deciding how long to think to improve performance and efficiency.
“Some people want the model to think longer, even though this requires more time and tokens,” Boris added. “To further improve intelligence, set effort=high via `/effort` or in your configuration.json.”
With medium effort, or effort = 85, now defaulting to users, Anthropic has promised to test higher effort for Teams and Enterprise users so that they “benefit from extended thinking even if it comes at the cost of additional tokens and latency.”
“I appreciate the depth of thought and care that went into this,” Boris also noted, crediting AMD’s Laurenzo for the analysis.
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds. Be sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us on WhatsApp also.
