Yep. This is exactly the conclusion I reached as an RLHF'er. Reasoning/LRM/SxS/CoT is "just" more context. There never was reasoning. But of course, more context can be good.