if you're like me, you just cranked gpt-5.2 reasoning effort to xhigh because using anything than SOTA is unenlightened, but friends, i'm pleased to share that i'm now using medium reasoning effort 85% of the time, and my work is fast again