On February 5th,a significant development in the world of artificial intelligence emerged when Lenovo announced its personal intelligence agent,"Xiao Tian," would integrate the highly sought-after DeepSeek AI model.Users eagerly accessed the DeepSeek feature through Xiao Tian's AI Space interface.This advancement represents a notable trend in AI technology where platforms are increasingly seeking to leverage powerful AI capabilities.

Simultaneously,Huawei was making similar strides with its system-level intelligence agent,"Xiao Yi," which,upon upgrading to the latest version of HarmonyOS NEXT,introduced the innovative DeepSeek R1 model.This synchrony in development between two major players in the tech sector—Lenovo and Huawei—raises questions about whether this was merely coincidental or indicative of a broader trend within the industry.

The meteoric rise of DeepSeek-R1 in popularity can be distilled into two key concepts: efficiency and open-source accessibility.DeepSeek-R1 presents a compelling alternative to other AI models,delivering impressive reasoning and generalization capabilities at a fraction of the cost when compared to models such as OpenAI's.This efficiency allows more manufacturers to integrate advanced AI functionalities within their products,putting state-of-the-art technology into the hands of everyday users.

The open-source nature of DeepSeek only further accelerates its adoption,enabling companies to utilize and optimize its technology without being beholden to a few large AI corporations.This liberation from reliance on predominant players in the field allows brands like Keep,Huawei,and Lenovo the flexibility to innovate without restrictions,steadily expanding the list of companies benefiting from this powerful AI model.

However,while the integration of DeepSeek into devices by major hardware providers like Huawei and Lenovo is exciting,the journey towards effectively utilizing DeepSeek's capabilities remains incomplete.For example,Xiaomi's integration seems more like an app plugin rather than a deeply embedded system feature.

In the competitive arena of terminal AI,Huawei stands out due to its dedicated focus on developing its own technology.From large-scale cloud models to operating systems and hardware,Huawei is often compared to companies like Google.Yet,with the arrival of DeepSeek-R1,Huawei's approach shifted.By integrating DeepSeek R1 only within HarmonyOS NEXT in a modular fashion rather than a deep integration,Huawei's intentions appear to involve careful observation and strategy development.

Underneath the surface of the Xiao Yi app,users can access DeepSeek R1 capabilities,presenting not just a user interface enhancement but a true evolution of AI experience.This integration allows for sophisticated interactions such as code generation,mathematics calculations,text generation,and complex logical reasoning tasks.Feedback from users suggests that DeepSeek R1 is notably more responsive and informative compared to prior implementations,showcasing an impressive leap in AI sophistication.

As articulated by Huawei Terminal BG CTO Li Xiaolong,the enthusiasm surrounding DeepSeek-R1's launch was profound,prompting rapid expansion efforts within the company.Yet,despite its powerful features,the incorporation of DeepSeek R1 within Xiao Yi still manifests as an external tool rather than a standard AI engine.Users interacting with the Xiao Yi assistant must specifically activate DeepSeek R1 to utilize its advanced capabilities,revealing a transitional state rather than a complete integration into the core functionality of Xiao Yi.

Currently,it appears that Huawei views DeepSeek-R1's implementation more as an enhancement rather than a final evolution of terminal AI.This approach enables quick deployment of the latest AI features but fails to revolutionize the foundational experience of device AI,which still largely operates on established principles.Hence,the way users engage with DeepSeek R1 is reminiscent of a supplemental tool rather than an intrinsic systemic shift.

This method of integration also prompts questions regarding why Huawei opted for this setup amidst the evolution of HarmonyOS NEXT.Notably,the absence of a dedicated DeepSeek app compatible with HarmonyOS NEXT indicates that users must rely on Xiao Yi for accessing DeepSeek capabilities.This makes Xiao Yi the only app within HarmonyOS NEXT that can natively interact with DeepSeek-R1,further signifying its strategic role.

Moreover,Huawei's cautious approach could stem from its commitment to developing its proprietary AI solutions,such as the Pangu model.The integration of DeepSeek R1,an external open-source model,may signify a temporary exploratory phase rather than substantial commitment to external technology.

Additionally,both Huawei and Lenovo have refrained from deploying DeepSeek R1 locally,despite its efficiency.The model necessitates considerable computational power and hardware support,presenting challenges that manufacturers are still navigating as they work towards more streamlined local optimization solutions.With DeepSeek-R1's full model containing 617 billion parameters,this is certainly not well-suited for standard personal devices.While distilled versions are available,they inevitably come with performance compromises.

Despite these constraints,DeepSeek-R1 stands as a harbinger of change within the terminal AI landscape.Qualcomm's CEO,Cristiano Amon,elucidated during a recent financial call that DeepSeek R1 holds promise for Qualcomm,as its chips can efficiently run AI capabilities locally instead of relying on cloud infrastructure.This sentiment captures the shift towards minimizing dependency on external servers while enhancing on-device AI functionality.

The emergence of DeepSeek R1 not only augments the AI experience on personal devices but also signals the evolution of AI in consumer electronics—from smartphones to PCs.As technology continues to iterate,the sophistication of AI assistants is simultaneously improving,evident in models like ChatGPT,Gemini,and Claude,which have not only enhanced conversational capabilities but also opened new avenues for user interaction.

DeepSeek-R1,falling under the MIT open-source license,empowers developers and manufacturers to customize and improve the model to fit their unique needs.This flexibility accelerates the integration of advanced AI capabilities into personal devices,liberating the terminals from reliance on proprietary systems.

The implications of DeepSeek-R1's open-source nature on the terminal market are profound.It streamlines the application and local deployment of AI,enabling manufacturers to create tailored AI solutions without needing to depend on closed-source models from industry giants.In essence,DeepSeek-R1 ushers in a new era of customizable AI experiences,offering enhanced reasoning,adaptability,and real-world applicability across various scenarios.

In summary,while the current integration of DeepSeek-R1 remains somewhat elementary,it lays a foundation upon which the transformation of terminal AI can build.As manufacturers continue to explore and adopt more profound insights from DeepSeek-R1's capabilities,we stand on the cusp of an evolving narrative where the AI experience on personal devices becomes deeper and more intuitive.This shift signifies just the beginning of what a more integrated AI experience might entail,paving the way for a future where AI becomes ubiquitous and essential within our daily technological interactions.