OpenAI CEO萨姆·奥尔特曼在社交平台X回应近期的内存言论争议，称其有关个人电脑内存利用率的表态遭到断章取义。他强调，自己希望全球拥有充足的内存，也支持用户根据需要为个人电脑配置内存，并再次呼吁扩大内存产能。
这场争议源于10月29日上线的Cheeky Pint访谈。奥尔特曼在与Stripe联合创始人约翰·科里森讨论本地运行AI模型时表示，从全球资源配置角度看，大量供应有限的DRAM被安装在个人电脑中，却有约95%的时间处于闲置状态；在供应紧张的情况下，集中于云端的数据中心可以提高资源利用率。
相关言论随后引发部分网友批评，有人将其解读为倾向于把内存资源集中至AI数据中心。奥尔特曼在回应中表示，他实际上希望人们能在电脑中安装尽可能多的内存，并称自己长期以来一直主张提升全球内存产量。
在另一条回复中，奥尔特曼写道：“我希望你们拥有很多内存，让我们生产更多。”他还采纳网友建议，将X账号显示名称改为“Ram Altman”。奥尔特曼此次回应并未否认数据中心集中部署内存具有更高利用率的观点，其澄清重点在于通过扩大生产缓解供应压力，避免个人电脑与AI数据中心争夺有限内存资源。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
Open AI CEOSam Altman responded to recent memory arguments on social platform X, claiming that his statements about the usage of personal computer memory were taken out of context. He stressed his desire to have sufficient global memory and supported users to configure personal computers as needed, and reiterated his call for expanded memory capacity.
This controversy originated in Cheeky Pint's live interview on October 29. In a discussion with the co-founder of Stripe, John Coulson, on the local running of the AI model, Altman said that, from a global resource allocation perspective, a large supply of limited DRAMs was installed in personal computers, but about 95 per cent of the time was idle; in the case of supply constraints, concentration on cloud-based data centres could increase resource utilization.
The statements were subsequently criticized by some netizens, who interpreted them as a preference for centralizing memory resources in the AI data centre. In his response, Altman stated that he actually wanted people to install as much memory as possible in computers, and that he had long advocated an increase in global memory production.
In another reply, Altmann wrote, "I hope you have a lot of memory to make more." He also adopted the suggestion of netizens to change the name of account X to “Ram Altman”. In this response, Altmann did not deny the view that the centralized deployment of memory in the data centre had a higher utilization rate, and its clarification focused on reducing supply pressures through increased production and avoiding competition between personal computers and the AID data centre for limited memory resources.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.