"Melee" large model, Zhihu, Xiaohongshu, and Kuaishou each have their own strengths
It has always been the nature of Internet companies to chase trends. Therefore, following the popularity of the Yuanverse, AIGC and ChatGPT, which suddenly emerged and became popular all over the country, have become the focus of pursuit by enterprises in various industries.
In this context, whether it is a major HE Tuber domestic Internet manufacturer, a technology company, or a content platform, all players with some strength have withdrawn one after another, and a fierce competition has begun around large AI models. The "Thousand Model War" also began.
A wave of entrepreneurship based on large models has begun to spread, and domestic companies that have deployed large models have sprung up like mushrooms after a rain. However, unlike "high-end" players such as Baidu, Alibaba, and Huawei who build general-purpose large models, Zhihu, Xiaohongshu, Kuaishou, etc. have chosen to take a different approach.
As representatives of mainstream content forms such as content Q&A, short pictures and short videos, the exploration of large models by small giants such as Zhihu, Xiaohongshu, and Kuaishou has naturally become the focus of attention from the outside world.
Competition for large models
Large models are undoubtedly the hottest trend this year, and domestic companies have already announced their exploration of large models. Regardless of whether it is investment or self-construction, there are many people who have begun to enter the market to share the pie. Naturally, Zhihu, Xiaohongshu, and Kuaishou will not miss this "train."
As we all know, major Internet companies’ research and development and related layout of large models are highly consistent with their industrial Internet strategies. In contrast, although the entry of the three platforms Zhihu, Xiaohongshu and Kuaishou was unexpected, it is still reasonable.
On the one hand, AI technology is gradually becoming more popular. Players have fully benefited from the wave of artificial intelligence technology and naturally do not want to miss the changes of this era.
The development of AI technology will bring about industrial progress and the birth of new opportunities. As artificial intelligence technology continues to make breakthroughs, large AI models are also booming. The wave of large AI models is sweeping the world, and only by getting on the poker table first can we seize the "trump card" of rules.
Therefore, it is urgent for players such as Zhihu, Xiaohongshu, and Kuaishou to enter the large-scale model industry.
In addition, for these players, once other companies' large models are launched, they may pose a certain threat. In order to prevent risks, it becomes logical to enter the game. Entering the AIGC era, large models based on deep learning will greatly increase the speed and quantity of content production.
In this context, competitors with "technical advantages" are likely to quickly break through the moats of Zhihu, Xiaohongshu, and Kuaishou. Therefore, laying out large models has become a way for these small giants to stay competitive.
On the other hand, in the context of the fading of mobile Internet dividends, choosing to embrace large models can make the content of these players more diversified and is expected to bring new growth points.
The evolution from decision-making AI to generative AI is pushing the application of large models to more diversified scenarios. The launch of large-scale model products by platforms such as Zhihu, Xiaohongshu, and Kuaishou can improve the speed and quality of content production to expand more new content; it can also better meet user needs, thereby attracting more user retention.
In addition, these players all face commercialization difficulties to a greater or lesser extent, and using large models is expected to help them achieve better commercialization.
Xiaohongshu has always been trapped in the problem of "planting grass inside the site and pulling grass outside the site", making it difficult to build a closed loop of commercialization. Zhihu and Kuaishou have also been suffering from losses all year round due to their difficulty in overcoming declining advertising revenue and high operating costs, as well as being deeply involved in live broadcast e-commerce competition. Its training of large AI models can bring a wave of dividends to the content community, thereby feeding back business growth and accelerating commercialization.
From the above perspective, the layout of large models seems to have become a battle that niche players such as Zhihu, Xiaohongshu, and Kuaishou have to fight.
However, due to the different business logic and entry angle of each company, each company's advantages and performance on the large model are also different. Relying on their accumulation of AI technology and content, these platforms have also begun to show their talents in the field of large models.
