近日，360集团创始人周鸿祎在一次公开演讲中表示：‘AI圈被图灵奖得主掀桌子了’。这句话迅速引发热议，其背后指的是多位图灵奖得主对当前人工智能发展路径提出尖锐批评。例如，2018年图灵奖得主杰弗里·辛顿（Geoffrey Hinton）近期公开表示对大模型发展方向的担忧，认为当前AI缺乏真正的理解能力，且存在巨大能耗与安全风险；另一位图灵奖得主杨立昆（Yann LeCun）也多次强调，仅靠扩大数据和参数规模无法实现通用人工智能（AGI）。周鸿祎借此比喻意在指出，AI行业正面临一场由顶尖科学家发起的‘范式反思’。过去几年，业界过度追捧大模型、算力竞赛和商业化落地，却忽视了基础理论、可解释性和伦理安全等核心问题。他呼吁中国AI从业者不能盲目跟风，而应重视原创性研究与长期技术积累。这场‘掀桌子’并非否定AI的进步，而是提醒整个行业回归理性，避免陷入技术泡沫。正如周鸿祎所言：‘真正的创新不是堆参数，而是重新思考智能的本质。’
Recently, Zhou Hongyi, founder of 360 Group, remarked in a public speech: ‘The AI community has had its table flipped by Turing Award laureates.’ This vivid metaphor quickly sparked widespread discussion. It refers to sharp criticisms from several Turing Award winners regarding the current trajectory of artificial intelligence development. For instance, Geoffrey Hinton, the 2018 Turing Award recipient, recently expressed deep concerns about the direction of large language models, arguing that today’s AI lacks genuine understanding and poses significant risks related to energy consumption and safety. Similarly, Yann LeCun, another Turing laureate, has repeatedly emphasized that simply scaling up data and model parameters will not lead to Artificial General Intelligence (AGI).Zhou used this analogy to highlight a growing ‘paradigm shift’ initiated by leading scientists. Over recent years, the AI industry has overly focused on large models, computational power races, and rapid commercialization, often neglecting foundational research, interpretability, and ethical safety. He urged Chinese AI practitioners to avoid blind trend-following and instead prioritize original research and long-term technical accumulation.This ‘table-flipping’ is not a rejection of AI progress but a call for the field to return to rationality and avoid falling into a technological bubble. As Zhou put it: ‘True innovation isn’t about stacking parameters—it’s about rethinking the essence of intelligence.’
原创文章，作者：admin，如若转载，请注明出处：https://ktez.cn/7821.html