OpenAI CEO Sam Altman has responded to significant criticism regarding his recent comments on computer memory, stating that his remarks were taken out of context. The controversy stems from a podcast appearance where Altman discussed the global allocation of dynamic random-access memory, or DRAM, and its use in artificial intelligence workloads.
Altman Clarifies Memory Stance
During the podcast, Altman argued that it is inefficient for individuals to hold large amounts of the world’s limited DRAM supply in their personal computers when that hardware is used only five percent of the time. He was addressing the trend of running large AI models locally on consumer devices rather than relying on centralized data centers. In this supply-constrained environment, he noted that data centers allow for RAM usage to be maximized more effectively.
The statement was interpreted by many in the gaming community as a dismissal of personal computing power in favor of corporate infrastructure. In response to the backlash, Altman engaged in an exchange on the social media platform X, asserting that he actually wants “the world to have tons of memory and people to put as much as they want in their computers.” He added that he has been arguing for a ramp-up in global memory production for a long time.
Hardware Costs and Cloud Concerns
Despite Altman’s clarification, the reaction from tech enthusiasts and gamers remains bitter. The industry has been experiencing what many are calling a “RAMpocalypse,” a period where the demand for memory by AI companies has driven up prices for consumer hardware. This shortage has affected not only high-end personal computers but also gaming consoles. In the lead-up to the launch of Grand Theft Auto 6, the PlayStation Pro is priced at $900, and availability remains scarce.
Beyond the immediate cost of components, there is a deeper concern among gamers that major technology companies are pushing users away from local hardware toward cloud computing services. For an audience already concerned about the erosion of game ownership, the prospect of losing control over their own hardware is unappealing. This sentiment persists even as cloud streaming services like Nvidia’s GeForce Now have improved in performance and accessibility.
These concerns are compounded by the direction of generative AI in the gaming industry. If generative AI features are delivered primarily as cloud services, and games increasingly incorporate these technologies, users may be forced toward cloud computing. This view was echoed by Nvidia DLSS pioneer Bryan Catanzaro at a recent Game Developers Conference panel. Catanzaro suggested that hypothetical games built with AI “world models” would likely not run locally because “AI is fundamentally much more efficient in the cloud.”
Source: PC Gamer
