[1/2] Nvidia’s Explosive Rise from Zero to Trillions (Category Documentary)

I watched this with great interest. documentary Nvidia’s transformation from a startup on the verge of bankruptcy to a company worth $3 The film spans the journey from the bottom to the bottom. 1993 One year before the current dominance of AI. I was interested to see what stages the company went through and what conclusions can be drawn from us as engineers and technical managers.

Foundation and early years Nvidia was founded by three engineers -- Jensen Huang, Chris Malachowski, and Curtis Prim -- who met while working at Sun Microsystems and LSI Logic. "" The legendary meeting took place at Denny's restaurant in San Jose, where over cups of coffee they discussed the future company day after day. "" Their bet was on 3D graphics for PCs and video games, even though they hadn't even seen a personal computer at the time. (jealousy)and the reduction of NV ("next version"), which the founders used to save files . . . with an initial capital of just $40,000 and subsequent funding of $20 Millions from Sequoia Capital, they started their journey.

Failure and miraculous rescue First product NV1 (1995) It turned out to be a disaster: 250,000 Chips sent to partner Diamond Multimedia returned 249,000 . The problem was that the chip used quadratic primitives instead of triangle primitives, which became the Microsoft DirectX standard, plus it was too expensive. "" After the cancellation of Sega's contract for NV2, the company had only one month's salary funding left. 128 (August 1997) It was the product that saved Nvidia -- it sold a million units in four months, as opposed to NV1, which sold a thousand units. "" Since then, the company's unofficial motto has been, "" We're always in the middle of it. 30 days from bankruptcy.” Interestingly, my first computer had a riva tnt2 chip, so I got to know the kids before they even started. 21 century:)

Breakthrough and IPO GeForce 256 (1999) It was the world’s first GPU with a programmable accelerator – a breakthrough in the concept of accelerated computing. In January 1999 Nvidia conducted an IPO, which gave the company more capital and publicity. The next big step was a contract with Microsoft for $200 Millions to develop chips for the original Xbox.

Turning to AI The most interesting decision was made in the middle. 2000-x: CUDA release in 2006 This year, we had a parallel computing platform that allowed GPUs to go beyond graphics. "" Stanford professor Andrew Ng recalls that he was already in the middle of his career. 2008 During the year, his students experimented with CUDA for deep learning and received acceleration in the study. 10-100 once. It was a huge risk: 2000No one knew if AI was going to be profitable. "" Nvidia decided on a massive transformation of the company, adding costs, hiring people and distracting attention from the core business in gaming and computer graphics. "" Huang described it as a "wholesale pivot" - a complete reorientation of the entire company.

Interestingly, I had a colleague at MIPT, whom we took to seep all sorts of websites, and he was parallel CUDA on the scientific trajectory and at some point told us arivederchi and left to do CUDA full time.

Then the AI map went well - in 2012 happened AlexNet and the start of “modern AI”. NVIDIA explicitly notes that AlexNet’s breakthrough was the trigger of the modern AI era, and the GPU was at the center of that wave. And about the insights and conclusions from this documentary we will talk in next.

#Documentary #Infrastructure #AI #ML #Engineering #Software #Leadership #Startup #DistributedSystems