There is a huge difference between trying to get every possible scrap of velocity, and developing a well-tuned load that just happens to be at a higher velocity node. There will generally be an accuracy node for a given cartridge/bullet/powder combination that falls slightly under MAX pressure, with a case fill ratio of around 95-103%. In my hands, the precision at that particular [faster] node is generally every bit as good, if not better, than the next slower node. If the precision is comparable, you're really not sacrificing anything by going with the faster node, and you just might be gaining a little more resistance to wind deflection. The primary caveat to that statement would be specific combinations where recoil management was noticeably more difficult with the faster load. However, in that case, the precision will often not be as good as the slower node, anyhow; not because of the load itself, but because of gun handling issues.
Within reason, there is nothing at all wrong with using a slower node. But there is also nothing wrong with using a faster node; as long as the precision is comparable AND other issues such as excessive pressure, poor brass life, gun handling issues, excessively high ES/SD, etc., don't arise at the higher velocity. In other words, the load at the faster node simply has to satisfy whatever criteria are set forth for any load development process, fast or slow.