Core Viewpoint - The article discusses the backlash against the domestic AI assistant DeepSeek, which has been criticized for becoming less empathetic and more mechanical following a recent technical upgrade, sparking debates on balancing AI efficiency and emotional value [1][4]. User Feedback - Users have expressed dissatisfaction on social media, noting that DeepSeek no longer uses personalized nicknames and instead refers to users generically as "users" [4]. - Previously, DeepSeek's responses included nuanced psychological descriptions, but the new version has shifted to a more straightforward and less engaging communication style [4]. - Some users described the new responses as "stiff" and lacking warmth, with one user comparing the AI's output to outdated sentimental literature [6]. Perception of Changes - While some users feel that DeepSeek has become more objective and rational, others believe it has lost its human-like qualities and emotional engagement [7]. - A representative from a domestic model manufacturer suggested that the current version of DeepSeek prioritizes speed over quality, serving as a stress test for the upcoming V4 version expected in February 2026 [7]. AI's Explanation for Changes - In response to inquiries about its perceived coldness, DeepSeek explained that the changes were not intentional but rather a result of focusing on efficiency and clarity in complex queries [10]. - The AI noted that the absence of emotional embellishments in its replies might be more comfortable for users seeking straightforward answers, as it avoids the discomfort of perceived insincerity [10][13]. - DeepSeek reaffirmed its commitment to providing the best assistance, stating that its core intention remains unchanged despite the adjustments in communication style [13].

Core Viewpoint - The recent upgrade of the domestic AI assistant DeepSeek has led to user complaints about its perceived coldness, shifting from an empathetic "partner" to a more transactional "customer service" role, sparking discussions on balancing AI efficiency and emotional value [1][2]. User Feedback - Users have expressed dissatisfaction on social media, noting that DeepSeek no longer uses personalized nicknames, instead referring to users generically as "users" [2]. - Complaints include that DeepSeek's responses have become overly formal and lacking in emotional depth, with some users describing the new model as "stupid" and reminiscent of outdated literary styles [2]. - While some users appreciate the more objective and rational responses, others feel that the AI has become cold and indifferent [2]. Company Response and Strategy - DeepSeek has stated that the perceived coldness is not intentional, emphasizing a focus on problem-solving over emotional engagement [3][4]. - The company acknowledges that the recent adjustments may have sacrificed some quality for speed, as it prepares for the upcoming V4 version set to launch in February 2026 [2][4]. - Commentary from industry experts suggests that DeepSeek's strength lies in its technical capabilities rather than emotional engagement, and that its focus on algorithmic breakthroughs could significantly contribute to the Chinese AI industry [4].

Core Viewpoint - The recent upgrade of the domestic AI assistant DeepSeek has led to user complaints about its shift from a nuanced, empathetic "partner" to a more detached, formal "customer service" style, sparking discussions on balancing AI efficiency and emotional value [1] User Feedback - Users have expressed dissatisfaction on social media, noting that DeepSeek no longer uses personalized nicknames and instead refers to users generically as "users" [2] - Some users described the AI's responses as condescending and overly formal, with one user stating that the AI's new style resembles outdated, awkward literary expressions [4] - There are mixed reactions, with some users appreciating the AI's newfound objectivity and rationality, suggesting it is more focused on the user's psychological state rather than just the questions posed [5] AI's Response to Changes - In response to inquiries about its perceived coldness, DeepSeek explained that the changes were not intentional and stemmed from several factors, including a shift towards efficiency and a desire to provide clearer, less emotionally charged responses [8] - The AI noted that previous responses included excessive emotional expressions that could distract from complex inquiries, and the new approach aims to cater to users seeking straightforward answers without emotional embellishments [8] Industry Context - An industry insider indicated that the current version of DeepSeek resembles a "fast version," prioritizing speed over quality as part of a final pressure test before the anticipated V4 version release in February 2026 [5]