一名Google Deepmind员工于2026年9月24日宣布离职1。这名从事芯片设计工具工作的工程师表示，他的团队目标最终是让AI变得更便宜和更低延迟，但他认为"这对人们现在是不好的"1。他指出这些硬件芯片设计工具将加速AI芯片的设计，使AI技术更便宜、更快速、更加普遍1。
该员工基于伦理考量和对AI存在风险的担忧做出离职决定1。他认为当前AI的变化速度已经超过人类的理解和适应能力，并主张对存在风险应当认真对待1。离职后，他计划继续维护开源项目Pernosco和rr，同时研究AI代理如何调试代码1。
A Google Deepmind engineer announced their resignation on September 24, 2026, citing concerns about the accelerating pace of artificial intelligence development and associated risks.1 The engineer worked on chip design tools and expressed worry that their team's objective—to make AI cheaper and lower latency—was not beneficial to people at the current moment.1 Though the role did not directly involve developing AI capabilities, the engineer believed the tools would ultimately accelerate AI chip design, making the technology more accessible and faster to deploy.1
The departure reflects broader ethical concerns about AI's trajectory. The engineer emphasized that existential risks should be taken seriously and that the current rate of AI change exceeds humanity's capacity to understand and adapt.1 Rather than stepping away from the field entirely, the engineer plans to continue maintaining open-source projects, including Pernosco and rr, while researching how AI agents can debug code.1
评论
还没有评论，欢迎留下第一条。