According to CoinMeta, Tang Daosheng, Senior Executive Vice President of Tencent Group and Head of the Cloud and Smart Industries Business Unit, responded positively to the concerns that "Tencent is slowing down in AI" in a lengthy statement. He acknowledged that there is indeed anxiety within the company, as competitors have performed more strongly over the past few years. Tang Daosheng pointed out that Tencent's overall computing power is "seriously insufficient," which has slowed down the training of the Hunyuan model and product development, having a significant impact. He believes that what we are experiencing may just be the first kilometer of a AI marathon, and that "enduring longer is more important than winning early." He previewed that the next generation of Hunyuan HY4 will see more breakthroughs and emphasized that Tencent cannot rely solely on basic models. Tang Daosheng also reflected on several mistakes, mentioning that a large amount of promotion was invested in Yuanbao for user growth, but at that time, both the "model and product were not yet ready," resulting in less than ideal outcomes. The team now uses AI extensively for coding, with people mainly responsible for judgment, debugging, and quality control. He predicts that the boundaries between future products, development, and testing will become increasingly blurred, and small teams will be able to accomplish what used to require large teams.