Well, basically, it could only be as scalable as Hive is. All the second layer transactions first go through Hive. So, even if you go fast, you can't go faster than 1 block every 3 seconds. I'm not trying to make the fastest or the most scalable second layer blockchain, we just need to be as fast as Hive.
What I'm trying to do is to make the usage of that second layer tool as smooth as using Hive.