4 ms·
Did you explicitly tell it to use Sonnet or Opus subagents and stick at or below high effort? Asking because such practices make a huge difference in the qualit
by kittoes 3mo ago
Did you explicitly tell it to use Sonnet or Opus subagents and stick at or below high effort? Asking because such practices make a huge difference in the quality of output and the amount of tokens burned. I used one of my accounts to explore ultramax and it was just a token hog that might be worse than Opus.
- giancarlostoro 3mo agoI had it on whatever the recommended settings was, but maybe I should have told it to use Sonnet for most subtasks. Even so, I'm just not that impressed, I felt like I got more done by just using Opus.
- sroussey 3mo agoYou can write a skill (many have) to lead it in how to use different models and efforts for subtasks. For searching the code base, for example, I have it use Haiku, which is fast.
- kittoes 3mo agoYeah, that's what bit me. Even Anthropic's own documentation seems to indicate that Fable is not all that great as your go to model for tasks. What it seems to excel at is a sort of leadership role because it proactively keeps all the subagents in check. If you're not explicit in the prompt or haven't configured your environment then the default behavior is to use subagents that match the host.