China’s tech talent making big strides in AI, creating apps for world
San Francisco -based AI Design App Lovart was officially released on Wednesday with North American users.
Lovart
Beijing – Chinese developers strengthen some of the latest vehicles for a global market.
Melvin Chen claimed that there were “800,000 users in 70 countries” for the test version.
“As the first step, we will focus on North America,” translated by CNBC. He said. He previously directed Chinese operations for Capcut, a popular video editing application of Baytan, who is still in the first place in the Apple’s photograph and video category at the Apple’s US App Store.
Lovart uses AI to create logos, stickers and other brand images based on text requests. The new version of Wednesday includes a “chatcanvas” feature that claims to facilitate certain arrangements – a customer can ask a professional designer to change two symbols, it’s hard to explain with words, but it’s a simple task when visuals are included.
It is waiting for Lovart to exceed 1 million users after the launch. However, he said that the application has not come to China soon, because it is often based on Anthropic’s Claude 4 AI model and Openai – both of them are not officially available in China.
Beijing should give green light to productive AI models for public use and run a solid firewall that blocks sites such as Google and Facebook. It has its own rules on where the services of companies can be used.
Most of the Lovart team was in San Francisco to better localize the products, while a part of the Chen production team is in China. He refused to share operating costs and said that the initiative would ask the investor financing after enabled sufficient users.
Lovart has a free option with a monthly subscription fees of monthly subscription fees for wider use.
AI applications for video
In a global AI race, the US government has increased its restrictions on American companies that sell advanced semiconductor to China in the last few years. Openai, based in San Francisco, launched Chatgpt Chatbot at the end of 2022 and did not produce a clear competitor with the breakthrough of China Deepseek until January this year.
However, analysts have long been the advantage of China’s artificial intelligence. Considering that internet -based Chinese companies can build large food distribution and short video applications for the big local consumer market, applications are instead of models.
Already in the AI video generation, Kuaishou’s Kling and Shengshu’s Vidu won global users in the last 18 months. In the field of AI agents, which can automatically perform a series of complex tasks, Manus attracted attention international attention.
“China is increasingly more and more effective by policy support for concentrated technical capabilities, agile development culture and AI commercialization,” Forrester Vice President and Chief Analyst Charlie Dai said. He said. “Cost Effective Model Training and Fast Consumer Application Repetition, often prioritize accessibility of open source.”
“Chinese models are now competing globally, AI is challenging us while reducing the costs,” he said.
Another advantage is that Chinese models, such as Deepseek and others, are open source, that is, they are free for developers to download and use.
Huging Face, an online platform that allows people to try open -source AI models, regularly among the Chinese models. best trend For users.
As of Wednesday, the K2 coding -oriented model, which was launched this month, took the first place on the site, and then the Qwen3 coding -oriented modes came to the market in the early hours of the day. In the image-3D models, Tencent’s Hunyuan ranks first, while the French-based Mistral’s Voxstral is the first text from the sound text.
Chen said Lovart will focus on AI to create pictures and videos instead of 3D models.
“AI New Camera … [for] Human imagination, “he said. The beginning, New York, Tokyo and Europe, including the design community, organizing events by organizing activities, he said.
Subscribe now
According to Sensor Tower, Chatgpt is the most popular AI application in the West with an average of 70 million users in the US and 144.6 million users in Europe as of July.
Google’s twins were a distant second in both markets, but Microsoft Copilot ranked third in the US, while Deepseek kept the third place in Europe.
During the visit to Beijing last week, Nvidia CEO Jensen Huang said that almost all users of Deepseek have downloaded the model to work locally in countries around the world. He also stressed that priorities have changed for AI development.
“Over time, one of the models will be increasingly less important,” he said. “One of the models will be the most useful.”
