Core Viewpoint - The essence of business in the AI era is to focus on exclusion, prioritizing what not to do rather than spreading resources too thinly across multiple ventures [4][9]. Group 1: Company Insights - The company, founded by Tang Wenbin and his peers from Tsinghua's "Yao Class," has evolved through the AI 1.0 era, learning from both successes and failures [6][7]. - Tang Wenbin emphasizes the importance of not overextending the business and instead concentrating efforts on the most advantageous areas [8]. - The company, Original Intelligence, has focused its first year on model development and AI infrastructure rather than rushing to scale orders or inflate valuations [11][12]. Group 2: Iteration and Adaptation - The key to competitive advantage in the current landscape is iteration efficiency, which allows the company to quickly identify and rectify issues [17][50]. - The company acknowledges that the pace of technological change is faster than anticipated, necessitating rapid adaptation to new challenges [19][20]. Group 3: Data and Model Development - The fundamental challenge in embodied intelligence is the strength of the model, which relies heavily on data to unlock various scenarios [24]. - The company believes that valuable data comes from real-world interactions, where robots can encounter and learn from failures [25][26]. - To overcome the "data deadlock," the company aims to find suitable scenarios for robots that allow for mistakes without severe consequences [27][28]. Group 4: Business Strategy and Market Position - The company is focused on logistics as a primary application area due to its relatively forgiving nature regarding errors, allowing for gradual improvements [28][29]. - The strategy involves a gradual unlocking of scenarios as model capabilities improve, rather than a fixed focus on specific applications from the outset [30][32]. - The company does not view the development of hardware as essential for valuation; instead, it prioritizes solving core problems effectively [36]. Group 5: Training and Benchmarking - The company advocates for a "embodied native model" that engages with physical world data from the beginning to enhance model capabilities [41][42]. - It emphasizes the importance of benchmarking to assess model performance, ensuring that internal evaluations are conducted to maintain quality standards [45][48]. Group 6: Future Outlook - The company anticipates that true commercial applications of embodied intelligence will emerge within the next year, particularly in first-line scenarios [53]. - It recognizes the challenges of entering the consumer market, preferring to establish a strong foothold in the B2B sector first [55].