Sophia Unger
Staff Writer
Building a personal computer has never been harder and more expensive. OpenAI CEO, Sam Altman, is one of many to thank for that. If you are like me, you are aware of how expensive it is to build your own PC right now. On top of the graphics processing units (GPU) and motherboard, random-access memory (RAM) prices have skyrocketed about 500% within the past year. Yes, you are reading that correctly. What was once $150 is now more than $700, making 32 gigabytes of memory sound unattainable.
RAM acts as your device’s short-term memory and stores data needed to access applications. The faster the RAM, the faster data flows through the central processing unit (CPU). If your computer does not have enough capacity, it will poorly function, if at all. Dynamic random-access memory (DRAM), is a common type of RAM used for everyday devices. This is an essential piece of hardware for any phone, computer, or console. However, this necessity is also shared with AI and data centers, as they need memory for large language models and overall efficiency.
So what do the billionaires think? Altman recently had an interview with John Collison, the president of Stripe. Looking at AI from “a global allocation perspective,” Altman believes that DRAM in your home computers is inefficient and not being used to its maximum potential. He seemingly criticized computer owners, by saying that the world’s limited DRAM is only used 5% of the time. After experiencing public backlash from this interview, Altman posted on X, “I actually do want the world to have tons of memory and people to put as much as they want in their computers.” This was a complete turnaround from what he claimed in his interview, though he also called it “annoying” that people were taking his remarks out of context. No amount of context can save the public from knowing his true intentions.
The inconsistent remarks that Altman makes about RAM lead me to question his goal. In this interview, the argument of local inference was brought up as a supporting claim for increased DRAM for data centers. He mentions that people think the “problem is solved now” and that we do not need everything in the data centers due to open-source models. Regardless of how out of touch people like Altman are to the multiple threats inflicted on our environment and well-being that data centers cause, the last thing that tech executives want is to get rid of them. If AI were processed locally within your device, rather than depending on data centers, this would harm them financially. So does Altman really stand by his claim of wanting people to have tons of memory within their devices?



