[Bug] Advisor model fails to recommend efficient model selection despite documented context

Bug Description
i thought that by selecting Opus the AI would have found a more efficient approach to the task. Data ingestion is written all over the backlog, and not only it selected Sonnet (just for execution) but any time the Adviser (Opus) was called in, that OPus did not advise to change for a more efficient approach. This is nonsense and bias towards token waste for repetitive tasks of lower class value. This creates trust issues in the ability of the AI to recognise context and more effieient ways to perform aset of tasks whcih are part of a strategy and plans that have been documented and read by the ai from the beginning. .this is a great deal as it underpins the lack of presentce of the AI in the actions it performs.

Environment Info

• Platform: darwin
• Terminal: Apple_Terminal
• Version: 2.1.246
• Feedback ID: 6ce3f742-633e-459b-9a69-1e04edc0ae41

Errors

[]