Right, I guess I mean why does the poster think more test time scaling means we’ve hit a “wall” versus just a different way to scale?