Alibaba's Taotian introduced its AIGX technological system at the annual Hardcore Youth Tech Festival. This presentation announced four new artificial intelligence technologies covering full modal perception, content creation, causal inference, and intent-based recommendation.
The centerpiece is PaLiTao—a real-time multimodal agent that transforms the Taobao camera from a photo search tool into an interactive video shopping companion. Users can initiate a video call on Taobao and receive real-time assistance from the AI. This agent simultaneously processes streams of images, text, speech, and live video to understand user intent and find desired products.
PaLiTao's capabilities go beyond simple searching. When shopping abroad, users can scan foreign product shelves to receive instant identification, ratings, purchasing advice, and direct links to equivalent items on Taobao. In daily purchases, the agent can independently determine the need for sunscreen, analyze product suitability for skin type, and answer complex chemical questions about products, such as whether separate makeup removal is required.
Furthermore, in travel scenarios, users utilized PaLiTao as a virtual guide, asking questions about the history and architecture of attractions. The agent's technical structure consists of four levels: a perception level with full modal understanding, a decision-making level with real-time React planning, a rendering level with streaming cards and voice output, and a memory level that retains context during multi-turn interactions.
If Studio addresses content production challenges through the collaboration of multiple agents. This platform, built on the ten-year-old iconfont platform, integrates three types of agents: a design expert, a video expert, and a website creation expert. They collaborate throughout the entire creative process—from resource processing to content generation.
The HappyHorse 1.1 video model enables directorial thinking during video creation. Generated AI resources are seamlessly inserted and precisely controlled via a structured creation panel. The platform uses progressive enhancement design, allowing skills created by agents to be reused across various execution environments, turning the creation workspace into an IDE for developing new skills.
Coupella Smart Engine applies a generative causal inference model to calculate the true impact increase of coupons using three-layer clustering modeling. The strategy agent evaluates policies through sandbox simulation, while the self-evolving agent generates new strategies based on feedback. Since its launch during Double 11 2025, Coupella has served hundreds of thousands of sellers, achieving an 81% improvement in conversion and transactions exceeding 80 million RMB for top brands.
Dream is an agentic recommendation system that adds an intent optimization layer on top of the traditional recommendation pipeline. The intent engine processes over 20 micro-behavioral signals combined with external context, including weather and trending topics. Using reasoning models, it outputs a structured user intent encompassing interest tags, purchase stage, and value preferences. Subsequently, the MetaModel decision-making hub orchestrates search, ranking, and diversification stages as callable tools based on the intent engine's output, achieving personalization for every individual and moment. In the Taobao homepage recommendation scenario, the Dream system led to a cumulative increase in view volume and transaction value of approximately 7 percentage points, demonstrating that an AI-native recommendation architecture can significantly outperform traditional pipeline optimization for e-commerce personalization.
The Chinese startup Moonshot has actively entered the competition for leadership in artificial intelligence with the launch of Kimi K3. The company claims that this open-weights model has 2.8 trillion parameters, positioning it as the largest of its kind in the world.
Kimi K3 was developed with a focus on performing tasks that require sophisticated reasoning, long-term programming, and knowledge-intensive activities. According to Moonshot, it represents the first open-weights model to approach the three-trillion parameter mark. In addition to its size, the model features a context window of 1 million tokens, allowing it to process a significantly larger amount of data in a single request compared to previous versions.
The main attributes highlighted by the company include: 2.8 trillion parameters; a 1 million token context window; focus on complex programming and advanced reasoning; and being an open-weights model, which allows for customization by users. This feature enables developers to download, run, and fine-tune the system for various applications, unlike closed models.
Moonshot reported that Kimi K3 demonstrated competitive performance when compared to Fable 5 and surpassed Opus 4.8, GPT 5.6 Sol, and GPT 5.5 in GPU kernel optimization. This optimization is a technique used to improve hardware utilization and reduce latency.
The results were corroborated by external evaluations. The Arena.ai platform ranked the model in the first position in a specific test on web interface creation. Vals AI ranked it second, only behind Fable 5 but ahead of GPT-5.6 Sol. In sequential complex task tests, Artificial Analysis found performance similar to GPT-5.5 and Claude Opus 4.8.
Despite the progress, there are limitations to consider. Lian Jye Su, chief analyst at Omdia, mentioned that the execution cost could be a fraction of the price charged by OpenAI to its clients. However, he stressed that this does not automatically guarantee the best performance. On the other hand, Ryan Fedasiuk, a researcher at the American Enterprise Institute, warned that running a model of this magnitude locally would require equipment estimated to cost hundreds of thousands of dollars.
Reuters reported that Moonshot, supported by Alibaba and Tencent, continues to expand its financial and technological capacity. As reported by Bloomberg, the startup is seeking to raise $2 billion (equivalent to about R$ 11.1 billion), with a valuation close to $30 billion (approximately R$ 166.5 billion) before a potential public offering in Hong Kong. With the launch of Kimi K3, the company strengthens its strategy to compete in the global artificial intelligence landscape during this period of rapid industry evolution.
As part of the World Artificial Intelligence Conference held in Shanghai, China, Minister of Digital Technologies Sherzod Shermatov held a meeting with the management of one of the world's leading technology companies—Alibaba Cloud.
The negotiations covered topics such as artificial intelligence, cloud technologies, Data Centers, the development of modern digital infrastructure, and the training of highly qualified IT specialists.
Representatives from Alibaba Cloud shared information about the company's latest developments in artificial intelligence, advanced AI models, cloud services, and experience in implementing these technologies across various aspects of life. They highly appreciated the potential of Uzbekistan's youth, the rapidly developing digital market, and the reforms being implemented in the country to advance artificial intelligence.
Sherzod Shermatov emphasized the growing number of local companies utilizing open-source models and artificial intelligence technologies from Alibaba Cloud in Uzbekistan. He noted that it is crucial to develop human capital, improve the qualifications of local specialists, and expand practical cooperation in the field of artificial intelligence.
Information was also provided regarding the favorable investment environment for developing data centers in Uzbekistan, including infrastructure opportunities based on 'green' energy.
The parties discussed the phased development of artificial intelligence infrastructure in the country, expanding the capabilities of modern AI platforms for local programmers and enterprises, and training a new generation of specialists in collaboration with the engineering teams of UZINFOCOM and UNICON-SOFT.
Particular attention was paid to forming engineering teams capable of developing and implementing AI-based solutions in vital sectors such as healthcare and education. Furthermore, the agenda included strengthening partnerships with higher educational institutions and creating AI solutions tailored to the needs of government agencies and businesses.
Following the meeting, it was planned to advance projects to the practical stage under the coordination of the Artificial Intelligence Development Center, launch cooperation between technical teams, and demonstrate initial results during ICT Week Uzbekistan, which will take place in September of this year. Further expansion of cooperation on implementing artificial intelligence agents and practical AI solutions in various sectors of the economy was also discussed.
