跳到正文
CNET· Tyler Lacoma·· 16 天前AI 評分56

OpenAI CEO Sam Altman 就個人電腦持有大量 RAM 的言論改口

OpenAI’s Sam Altman Backtracks After Hating on You for Having RAM

AI 導讀

OpenAI CEO Sam Altman 批評個人電腦持有大量 DRAM 後,在 X 澄清自己希望人們可按意願為電腦配置記憶體,並稱此前言論遭斷章取義。文章把爭議置於 AI 數據中心大量需要記憶體、消費電腦及遊戲主機價格上升的背景下,並指出更多 AI 在本機運算可減少對雲端數據中心的依賴。

正文

What does Sam Altman, CEO of ChatGPT creator OpenAI, think about you using RAM (aka memory) in the computer you own? It’s tough to say, because he’s made multiple contradictory statements following an interview last week where he said, “It’s quite terrible to have everybody have a bunch of the world’s limited supply of DRAM sitting in their own computer.”

Altman appeared to put the blame for RAM shortages on computer owners who have the hardware — which is necessary to run graphics and AI, among other things — and only use it “5% of the time,” as he said. Meanwhile, people are blaming data centers and the rapacious RAM acquisition of companies like OpenAI as the chief reason why computers have become astronomically expensive, game consoles are now beyond anyone’s reasonable budget and other device pricing woes. Internet users responded to Altman’s criticism with anger, memes and some sharp criticism of their own.

Read more: What Is RAMageddon? Why AI Is Making Laptops and Phones More Expensive

More from CNET

The saga, which included Altman briefly changing his social media handle to say Ram Altman, led to a flurry of Tuesday tweets on X where Altman backtracked, saying, “I actually do want the world to have tons of memory and people to put as much as they want in their computers, and it is fucking annoying that this was taken out of context.”

A representative from OpenAI did not immediately respond to a request for comment.

What is Altman even trying to say?

Altman’s first comments complaining about RAM sitting in so many personal devices around the world, made on the Cheeky Pint podcast, appear connected to OpenAI’s need for more data centers. Basically, OpenAI’s AI models like ChatGPT, along with other major models from Anthropic, Google and others, need a ton of memory to run the calculations first to train them and then for people to use them.

This need is what’s led to the super-rapid rise of data centers across the world, even if they have to illegally bulldoze forests to do it. In addition to creating concerns about water use and noise pollution, these data centers are sucking up a vast amount of RAM that was once available for other purposes, like consumer PCs and consoles. Data center companies are even stockpiling computer hardware they don’t have the space or energy to operate yet. That’s causing prices to skyrocket, and Altman’s ill-advised comments bemoaned how “from a global allocation perspective,” shortages and high pricing were affecting OpenAI’s own plans.

Huang with a Vera Rubin rack
Nvidia CEO Jensen Huang shows off his company’s Vera Rubin computing platform, the heart of a lot of AI data centers.Faith Chihil/CNET

People pushed back by pointing out that their individual use of RAM is minuscule compared to data centers, and that it sounded like Altman wanted all the world’s memory for his own company, which is already valued in the many billions of dollars (though not as many as it used to be).

In response, Altman clarified with his posts that he has “been arguing for the world to ramp up memory production for a long time,” and he would prefer a world where everyone has as much RAM as they want. Either he didn’t really mean what he said on the podcast, or he does want first dibs on all the RAM but realizes it’s a bad look. It sounds like Altman, much like his chatbot, is trying to say whatever he thinks people want to hear.

Is local AI bad?

Here’s another interesting twist: Most of the AI we use in our computers, phones and even smart speakers already runs partially on “local” hardware built right into our devices. But more complex tasks are sent off to the cloud and data centers to run, then deliver the data back.

The more AI is processed locally, the less we have to rely on the cloud and data centers for the AI tasks we find useful. I’ve called out current examples where once-complicated AI is now processed entirely on local home devices and hubs, such as the free facial-recognition features on the newest Eufy camera. As we make AI increasingly efficient, we may be able to run a growing number of personal features locally or mostly locally. Large-scale commercial uses of AI probably don’t have that luxury, and AI developers need those big data centers and tons of memory to train their new models.

But the more AI functions that are processed locally, the fewer subscriptions we have to pay, and the less money companies like OpenAI will be able to make through their subscription and usage fees. Those companies also won’t be able to collect as much data about us. That’s why I question if Altman really wants everyone in the world to have tons of memory — I suspect his original comments were closer to the truth.

Tyler Lacoma

Tyler Lacoma

Senior Editor / Home Security and Smart Home

Tyler has worked on, lived with and tested all types of smart home and security technology for over a dozen years, explaining the latest features, privacy tricks and top recommendations. With degrees in business management, literature and technical writing, Tyler takes every opportunity to play with the latest AI technology, push smart devices to their limits and occasionally throw cameras off his roof, all to find the best devices to trust in your life. He always checks with the renters (and pets) in his life to see what smart products can work for everyone, in every living situation. Living in beautiful Bend, Oregon, gives Tyler plenty of opportunities to test the latest tech in every kind of weather and temperature. But when not at work, he can be found hiking the trails, trying out a new food recipe for his loved ones, keeping up on his favorite reading or gaming with good friends. See full bio

The Latest from Tyler

來源:CNET · cnet.com