Okay. The first question on GP margin, as I explained, the improvement since 2021 was mostly due to the engineering effort to -- also a procurement effort, but mostly engineering effort to optimize the technical design, so that to do the same thing we need less server or bandwidth resources a bit more. We'll continue to optimize the design. But there is this diminishing return of the optimization. So it will get harder. It's not like Internet always gain the same amount of improvement over time, so it will get harder. And that's why we're not guiding kind of a higher GP margin. And in addition, as I mentioned before, our strategy is really to keep a healthy GP margin instead of trying to maximize GP margin whenever we can, so that we can keep enough competitive pressure our competitors as well. So that's on GP margin. In terms of R&D, it's true, we are both investing for today and investing for the future. As we expand globally, we see more use cases, we see more regions, more different macro environment, so we need to invest to really optimize the performance there. But we also made a lot of investments for the future. And we expanded our R&D team so much, almost eight times in the past two years. Obviously, there are a lot of inefficiencies. So looking forward, we will -- as I said, right, we'll be more focused and we'll work hard to remove some of the efficiencies. So in simple terms, we do not expect R&D to continuing to drive as we expand our operation globally. And there should be actually a very strong operating leverage, because a lot of the technology, a lot of the products we build can be leveraged across different use cases and across geographies. When we do the education solution we built in China, or we go to Eastern Europe or we go to South Asia, actually it's essentially the same product with very little modification. So actually we're seeing, in the end, there will be a lot of R&D leverage. So, again, we do not expect R&D Q2 rise [indiscernible] revenue.