AMD's AI director is not happy with Anthropic's Claude code; issue raised on Github reads: Claude cannot be trusted to perform complex engineering tasks, every senior engineer in my team …
AMD's AI chief, Stella Laurenzo, expressed concerns about Claude's performance, citing a regression in its ability to handle complex engineering tasks. She reported an increase in "stop-hook violations" and a shift to an "edit-first" approach, leading to lower code quality. Anthropic responded, stating that a "redact-thinking" setting only hides reasoning and that users can adjust effort levels for longer thinking.
AMD's AI chief, Stella Laurenzo, expressed concerns about Claude's performance, citing a regression in its ability to handle complex engineering tasks. She reported an increase in "stop-hook violations" and a shift to an "edit-first" approach, leading to lower code quality. Anthropic responded, stating that a "redact-thinking" setting only hides reasoning and that users can adjust effort levels for longer thinking.
