a
A funny thing happened on the way to the future. It recently took place at a giant sports arena in San Jose, California, and was described by some geeks as “AI Woodstock.” But while the original music festival attendees were primarily stoned by traditional drugs, San Jose’s roughly 11,000 people were hooked on the Kool-Aid lavished by the tech industry.
they gathered to hear a story
keynote speech
at a technology conference hosted by Jensen Huang, founder of computer chip maker Nvidia and now the Taylor Swift of Silicon Valley. Wearing his usual leather jacket and white-soled trainers, he put on a brave 50-minute performance reminiscent of Steve Jobs in his prime, but with a little less dexterity. Similarly, the audience was reminded of the fanboys who lined up for hours to be allowed into Mr. Jobs’ reality-warping field, with the exception of Mr. Hwang’s fans who, when he signaled applause, I wasn’t paying that much attention.
Still, it made for interesting viewing. Huang is an engaging speaker who has built a remarkable company in the years since first sketching out his Nvidia idea at his Silicon Valley diner in 1993. And the audience was in awe of him because they saw him as a man who foresaw the future long before they did, hoping to get a glimpse of what would happen next.
And in this respect they did not disappoint. Next up is a family of monster machines that includes Nvidia’s 208 billion transistor Blackwell B200 chip and a formidable supercomputer that fits in a rack and has about two miles of copper cable neatly intertwined inside. . There will be a huge round of applause.
The thought that came to my mind when I saw this scene was: How did a small company specializing in graphics cards for gamers become the third most valuable company on the planet? And how did it end so quickly? ?After all, NVIDIA
just worth it
$278 billion in October 2022, currently equivalent to $2.3 billion
Trillion
only chasing Apple and Microsoft.
It’s a great story and I’m sure someone has already written the script. However, even a cursory explanation suggests that from the outset, they were good at anticipating the needs of a particularly demanding user group, namely gamers, and ultimately inspired new types of processors by developing processors that could address their needs. An image emerges of a company that has realized that it has created something. Computer: A graphics processing unit (GPU) that can perform many calculations in parallel, as opposed to a traditional CPU that performs everything serially.
A pivotal moment occurred in 2013.
Mr. Huang decided
GPUs could be useful for an emerging technology called machine learning, which the company will focus on going forward, he said. It was a bold bet at the time, and Wall Street initially thought it was foolish. But when machine learning really took off and parallel processing machines were desperately needed to handle the large amount of computation involved, Nvidia hit the jackpot. If you want to run this kind of AI, you’ll need a lot of Nvidia GPUs. And more importantly, they needed a way to make them work together seamlessly, a kind of operating system. A Stanford software genius named Ian Buck wrote the software for Huang. They called it CUDA (short for Computing Unified Device Architecture) and since then, purchasing his Nvidia kit has become easy for anyone looking to get into the AI business.
Thus, during the greatest gold rush in the history of technology, Mr. Huang found himself virtually the only trader capable of supplying ready-made shovels. During February
his company reported
Quarterly sales were $22.1 billion, an increase of 22% sequentially and 265% year over year.
The Blackwell B200 tip is his latest super shovel. And in his keynote address, Hwang revealed what can be done with it. He announced his DGX GB200 NVL72 (Nvidia does not have human-friendly labeling).
powerful supercomputer
72 Blackwell processors in a single water-cooled rack. I couldn’t find any information on the price, but if I had to ask, I don’t think I could afford it.
However, Google can. So are Microsoft, Meta, Oracle, Tesla, Amazon, and Dell.from what
their boss says, they have already joined the line of supplies. In this case, Mr. Huang certainly saw the future. And it works for him. But whether that will work for the rest of us remains to be seen.
what i was reading
don’t forget me
A thoughtful essay by Rene Walter published in Substack magazine, Good Internet.
Intrinsic hostility to cultural memory
In the digital world.
politics of technology
a
great review
Bill Janeway contributes to the online opinion site Daron Acemoglu and Simon Johnson’s excellent book ”Project Syndicate.”
power and progress.
doubt thomas
Gia Tolentino
new yorker (paywall)
I don’t like fat controllers
in the story of the railway
thomas and friends. Neither do I.
Source: www.theguardian.com
