Of course, it stems and translates into Hakka and other languages, and we’ve chosen this for two reasons. One is that evaluation in alignment by the language community is easier, because the input and the output is of a comparable size. If this is generated from a few questions, it’s hard to evaluate. But for translation—culture translation and multi-modality translation—it’s easy to evaluate, so that’s what we’re going to focus on. For instance, we’ll give equal opportunities to the various different communities in Taiwan in order for them to become cybersecurity analysts because we expect that humans will not write most of Python code by hand, soon, maybe in a few months from now.