My team rely on a/b testing. It’s the only way we can test whether the tweaks in our ML model produce positive results or not. Playing devils advocate here I t

My team rely on a/b testing. It’s the only way we can test whether the tweaks in our ML model produce positive results or not. Playing devils advocate here I think it’s actually a good idea that YT are testing this. They need to understand how big of a negative impact this would be if released. If enough users hate this change as you and I suspect then hopefully YT see these results in their tests and revisit their approach. A/b testing stops the worst user experiences from getting rolled out. Our company did a high profile a/b test that got a lot of negative feedback from users and the whole redesign it got canned because of it.