Podcasts
This podcast explores AI Agents on mobile and computers, highlighting Zhipu AI's AutoGLM 2.0. As the first free, general-purpose mobile agent (no invitation code needed), it handles lifestyle and productivity tasks like ordering takeout and buying tickets. The mobile Agent market's potential surpasses the web, making it a key area of competition for tech giants and phone manufacturers. Liu Xiao, AutoGLM 2.0's leader, shares technical insights and the development journey from testing to launch, addressing challenges in cross-application operations, user privacy, and payment upgrades. It emphasizes Reinforcement Learning, especially end-to-end online Reinforcement Learning, for enhancing Agent capabilities and complex tasks, alongside benchmarks like Agent Bench. The discussion covers the pros and cons of Graphical User Interface (GUI) and Application Programming Interface (API) Agent operation modes, the interplay between general and vertical Agents, and phone makers' Agents. The podcast envisions AI Agents connecting more devices and running 24/7, urging industry collaboration to achieve AGI.
This podcast features an in-depth conversation with Michael Guan, the founder of FinalRound AI. He details how he, as a non-ABC (American-Born Chinese) entrepreneur, successfully founded and rapidly grew an AI-driven recruiting tech company from the ground up in Silicon Valley. Michael shares how the company, since its inception in October 2023, has met the job search needs of knowledge workers in North America and Europe through innovative products such as AI resume modification, automated job application, mock interviews, and real-time interview assistants. This has allowed them to quickly accumulate over 7 million users and achieve an annual subscription revenue (ARR) exceeding $10 million. The podcast reveals FinalRound AI's unique growth strategies, including systematic A/B testing, bold pricing strategies, brand restructuring, and the development of internal AI tools to achieve marketing automation (such as AI-driven SEO, content generation, and performance ad delivery). Additionally, Michael introduces the company's entrepreneurial culture in Silicon Valley, such as balancing the 'opportunity anxiety'-driven intense work schedule (996, referring to working from 9 am to 9 pm, 6 days a week) with unlimited vacation policies, and the vision of building a global team across San Francisco, Shanghai, and Bangalore. The episode concludes by looking at how AI is disrupting traditional recruitment processes, innovating corporate talent assessment methods through B2B products like 'AI trial periods,' emphasizing that future talent assessment will focus more on the methodology and tools used to achieve results, rather than just the results themselves.
Driven by his own emotional journey, Ren Yongliang founded Cece, a pan-psychological online platform leveraging big data and AI to address users' emotional needs. He elaborates on Cece's evolution from a tool-based product to a platform integrating community, bilateral services, and AI services, and shares his views on the competition between general-purpose Large Language Models and specialized applications in the AI boom. Facing the impact of AI technology, Ren Yongliang is determined to focus on embodied AI companions, committed to creating a new generation of products with proactive interaction capabilities, capable of integrating into families and providing emotional companionship. The podcast also explores the influence of psychology, Buddhism, MBTI, and other diverse perspectives on Ren Yongliang, as well as his philosophical reflections on human nature, emotions, periodic systems, and the social impact of AI. He is optimistic about the long-term potential of AI but also concerned about its impact on human dignity and value distribution systems, emphasizing the importance of human-machine collaboration and emotional empathy, demonstrating his continuous pursuit of a deep integration of technology and human-centered values.