Song Lin
Analyst · Citigroup. Please go ahead.
Yes, sure. This is Song Lin. So, I'll try to answer it. So, actually, it's like this, right. So, I think it's -- so first of all, just a comment that there's in general two kinds of AI, right? So, one is the discriminative AI, which is actually what Opera has been using, and it's also the one being used in news recommendation, content recommendation, and other pilots, for instance. And the other is the generative AI, which is what's been popular now. So, I would say that moving forward, we'll probably see both becoming more and more relevant, and the [two will have] (ph) a bit more focus, right? So, the first one, discriminative AI, will be able to give you more relevant contents. I think that we already -- actually it's always ongoing. People just perhaps didn't notice it, but then it's becoming much more visible, right? And then, like, I think the generative AI are more like in the ground where people will be able to see content with their own flavor, right, more like one -- you need [indiscernible] to do that. You can have a particular flavor of ways how do you want to see the news or articles, you cannot, of course, always ask AI to give you a summary based on your flavor, right? So, I think I would say that's more like one algorithm is to be able to give you the relevant content, the other is turn the content into what you might like. So, I think both are going to be very relevant. Like as I said, it's still very early stage. But as that's proven to be very clicking with also users' mentality, right. So, I think for now, we're just going to explore other kind of options. And I think what is happening now is there is more user to actually use that because of differentiating features, that's probably the biggest short-term gain, and also, for instance, the longer-time spent. So, I guess that's a high-level capturing of what might come into view.