Okay. All right. So the question about the forecast of the future Q3 advertising situation. Yes, kind of there's a weakness because of the macroeconomic situation is not that good, and a lot of uncertainties and also the downgrade of consumptions because of people who just don't have -- less disposable income to buy things and also the competition of companies, the profit margin is lower, and so they are more cautious in investing in marketing spending. And as to sectors, still our largest sector is auto sector and followed by FMCG and IT services. For auto sector, well, it's now kind of stabilized with the price war getting steady and it's the new energy, I mean, electric cars market share is expanding while the luxury cars and also joint ventures are kind of shrinking. So with this kind of market situation, we have this unique events and IPs and create or launch the kind of unique events and combine with our online properties to have some kind of unique marketing opportunities, so that they do spend still on Sohu. Yes. That's about that we're trying to do. Yes, so the whole macroeconomic situation is really not that good. But since our advertising market share is very small, so if we are successful in our social media success platform, we will just get a larger share. So that we hope. But that will not happen in the next few quarters. We'll continue to build our user base. As to your second question about the AI application, I think, yes, it's improved. Especially in the online games part, right, it does improve the efficiency, productivity of online gaming. But for the media platform and social networks, it helps our users to get answers or get results with the search capability on each app. So now people are actually doing information search on each separate apps, not -- so we're using the AI to improve the AI answering capability of the Sohu Video and Sohu News app. But that apps, mostly social networks, now we turn it into a social network platform. So the AI help is only marginal, right? So we are using the open source language model and then based on that, build on that to develop on that. So we're using various models, large language models. About share buyback, yes, we are halfway into it, $83 million already consumed, and we don't have any plan to have other investments or capital investments, because we are fighting on a major user base building and social network wall. So we need a war chest to do the battle. Yes. I hope I answered your question.