The "efficient market" framing assumes Anthropic wants to minimize output, but they don't. They charge per token, so the defaults being verbose isn't a bug they haven't gotten around to fixing.
That said, most of this repo is solving the wrong problem. "Answer before reasoning" actively hurts quality, and the benchmark is basically meaningless. But the anti-sycophancy rules should just be default. "Great Question!" has never really helped anyone debug anything.
That said, most of this repo is solving the wrong problem. "Answer before reasoning" actively hurts quality, and the benchmark is basically meaningless. But the anti-sycophancy rules should just be default. "Great Question!" has never really helped anyone debug anything.