My personal Perp track, except for omni.(variational.io)/?ref=OMNITAOBAO, has all been paused. There are just too many projects in this track, it's way too crowded. Unless there's a very early opportunity to get involved, a few leading projects in Perp have already emerged. It's actually really hard for other projects to break through. After all, the pie is only so big, and the rest can only sip some soup. But who gets to drink that soup is hard to say. By 2026, the all-in prediction track has way better cost-performance than Perp.