In the realm of technology, the anticipation surrounding AI assistants has been soaring ever since Steve Jobs laid out his vision for Siri back in 2010. At that time, Jobs recognized the potential of having an assistant that could engage with users in a manner akin to human interaction, offering both companionship and utility. Fast forward to today, and it appears that the goal remains steadfast: people still crave a personal assistant that truly understands them, capable of performing a multitude of tasks and seamlessly executing commands across various domains.The development of large models is significantly shortening the timeline toward achieving this ambitious goal. OpenAI's executives are emphasizing the potential of AI agents to yield groundbreaking advancements, with American Bank's November report noting that agentic AI—an AI with enhanced capabilities for autonomous planning and action—is spearheading a wave of innovation. This means the competition is not just about creating AI that can answer queries, but about building an AI capable of collaboration, planning, and independent functioning in a complex digital ecosystem.As this innovation cycle unfolds, the race is on to develop an AI assistant that is closest to users, connects most broadly, and facilitates collaboration among various agents. The implications are profound: AI's ability to enable natural language interactions could soon replace traditional graphical user interfaces, leading to a reimagining of user interactions across countless scenarios. Yet with this evolution comes the possible disruption of established balances within both software and hardware markets, raising the stakes for existing players and emerging challengers alike.Currently, we find ourselves in the early stages of this competitive landscape. A diverse array of participants is eager to seize the opportunity, ranging from ambitious startups developing large models to established tech giants and mid-sized companies vying for relevance. The landscape encompasses various niches: general AI assistants, specialized assistants, and tool-based AI interfaces, each seeking to carve out its own place in this evolving market.The key to success in this burgeoning realm lies in accumulating advantages in areas such as model technology, agent ecosystems, user acquisition efficiency, and innovations in the commercial chain. Those who excel in these domains stand the best chance of claiming lucrative entry points into the new digital food chain.Recent narratives from American Bank classify AI advancements into three waves: pre-Generative AI, Generative AI, and Agentic AI. The pre-Generative AI phase, extending from the 1940s until ChatGPT's debut in November 2022, saw notable tools like Siri and Alexa emerge but predominantly focused on data-driven insights and decision-making rather than comprehensive interactions. The current Generative AI phase, anticipated to conclude by October 2024, emphasizes two main applications: enhancing natural human-machine interactions and improving productivity through AI capabilities in search, video generation, code creation, and more.As research executive Silvio Savarese from Salesforce AI Research posits, the shift to Agentic AI marks a significant leap forward—where AI begins to autonomously manage entire tasks and take actions on our behalf. Gartner projects that by 2028, Agentic AI will independently handle at least 15% of daily work decisions. While we're not there yet—in 2024, that percentage will remain at zero—numerous companies are laying the groundwork for this transformation.Many AI systems are now prioritizing contextual memory as a bridge to autonomous decision-making. Google's AI assistant, Gemini, exemplifies this trend, maintaining knowledge about users' daily lives, work responsibilities, and preferences. Moreover, technological breakthroughs are ushering in enhanced multimodal capabilities alongside the ability to control applications, thereby expanding the potential of AI assistants. AI assistants can soon perceive through listening, reading, and viewing, leading to more natural interactions that bridge voice and visual inputs.The landscape is witnessing companies launch platforms designed to facilitate third-party AI application ecosystems. Microsoft’s Azure AI Foundry, for example, aims to enable organizations to design, customize, and manage AI applications and assistants. Meanwhile, Baidu has introduced a no-code AI application-building tool, aimed at facilitating millions of "super useful" applications. Observers note that the inability to cultivate a third-party ecosystem has stymied Siri's growth since its inception.The competition for dominance isn't limited to a single group; various players have vested interests as AI assistants emerge. Established platform holders like Alibaba and Baidu are battling to maintain their stakes, while upstart model companies are eager to challenge the status quo. Companies with robust AI models are racing to enhance their applications, leveraging their substantial existing user bases, unique data, and rich resources.Hardware manufacturers, traditionally sidelined during the explosive growth of mobile applications, now find AI assistants offer a ticket to reconstruct their business models. By integrating user habits, data collection, and seamless interactions with AI, they can better leverage their technology in this new landscape. For instance, by adding dedicated buttons for AI assistants on hardware devices, they create immediacy and accessibility in invoking these AI capabilities. Moreover, they hold the unique advantage of implementing secure hybrid AI solutions that can protect user privacy.However, the journey for tool-based products diverges into two potential paths. One path sees products like DingTalk, Quora, and Meitu devising vertical AI assistants aimed at specific functions, while another group looks to embed their content proactively within the larger AI ecosystems. The latter approach allows for adaptability and longevity within this rapidly evolving tech environment.As the conflict for leadership escalates, the stratification of who will grab the proverbial “golden stick” becomes clearer. Ultimately, this struggle revolves around the strength of the infrastructure. Who possesses the most advanced model capabilities, the broadest access to business or consumer touchpoints, and the most effective integration of hardware and software will likely become the leading authority.The evolution of foundational AI models remains paramount. Success will hinge on accurately interpreting user intent and executing tasks efficiently through intricate models. Reports have emerged regarding Google’s Project Jarvis, which intends to enhance its Gemini model by developing a sophisticated new model capable of interpreting clicks, text, and images within a browsing environment.The auxiliary capabilities accessible through various AI assistants will determine their relative positioning. A singular chatbot did not dominate the space; the more diverse their accessible capabilities, the more robust their market presence. This indicates a rich agent ecosystem is vital for providing a broad range of capabilities required in today's market.When mixed with a robust hardware strategy, AI competency will become two sides of the same coin. Manufacturers must navigate this collaborative landscape—where AI OS development, user comprehension, cross-system functionality, and cloud integration converge. This transition represents a major opportunity for hardware makers, such as Honor and Xiaomi, who seek to deepen their influence in this redefined landscape. Companies like ByteDance and Baidu are also rushing to meld hardware with their innovative AI solutions, reinforcing the idea that alignment within both realms is becoming pivotal.In conclusion, while the contest around AI assistants is still emergent, it’s evident that power dynamics are shifting rapidly. As new contenders and legacy players vie for control, the principles governing this innovation will unravel, showcasing a compelling overlap between user needs, market presence, and technology capabilities. The victors in this race will not only embrace innovation but also marry it with effectively harnessing user interaction to orchestrate the future of AI.