🖥️🎛️💥 Brah, Da New A.I. Supahcomputah Stay Making Noise, Run by Some Big Kine Computah Chips!
⬇️ Pidgin | ⬇️ ⬇️ English
K, so dis past week, inside one big room in one one-story building ova dea in Santa Clara, Calif., had these tall machines like six-and-a-half feet all humming and going off behind some white cabinets. Ho, da machines was all part of dis new supahcomputah dat just wen’ become operational da month befo’. 👨💻🏢⚡
Cerebras, one start-up from da Silicon Valley, wen’ pull back da curtain on da supahcomputah dis past Thursday. Dey wen’ make dis computah with their own spesho kine chips, made jus’ fo’ powering A.I. kine tings. Da chips stay big, yeah? Like one dinner plate, or 56 times bigga than da regula’ kine chips fo’ A.I. Every single Cerebras chip get da powa of hundreds of da old school chips. 🎉🍽️💡
Cerebras say dey wen’ build da supahcomputah fo’ G42, one company all about dat A.I. G42 say dey planning fo’ use da supahcomputah fo’ creating and powering A.I. stuffs fo’ da Middle East. 🌍🧠🚀
Da CEO of Cerebras, dis guy Andrew Feldman, wen’ say dey like show everybody dat can build one big kine A.I. supahcomputah, and can do ’em faster, with less energy, and fo’ less money. 💸⚙️🏋️
Dis year, everybody like have computing powa and A.I. chips, all cause of da A.I. boom going off around da world. Big kine tech companies like Microsoft, Meta and Google, and choke start-ups too, all been trying fo’ drop A.I. products afta da A.I.-powered ChatGPT chatbot wen’ viral cause da way he talk all like one real human. 🌐🤖💬
But, making A.I. kine stuffs usually take choke computing powa and spesho chips, so had one big rush fo’ find mo’ of those technologies. Back in May, Nvidia, da top dog in making chips fo’ A.I., wen’ say everybody stay hungry fo’ their products — known as graphics processing units, or GPUs — dat their quarterly sales going be more than 50 percent above what Wall Street was thinking. Nvidia’s market value wen’ shoot up past $1 trillion cause of dat. 📈💲🔥
Fo’ da first time, we seeing one big jump in da computer requirements cause of A.I., according to Ronen Dar, who wen’ help start Run:AI, one start-up ova in Tel Aviv dat helps companies make A.I. models. Had one huge demand fo’ spesho chips, and companies stay rushing fo’ get their hands on ’em. 🏁📊🧗
Fo’ get enough A.I. chips, some of da biggest tech companies — including Google, Amazon, Advanced Micro Devices and Intel — wen’ make their own kine. Some start-ups like Cerebras, Graphcore, Groq and SambaNova also trying fo’ get in on da action, trying fo’ get one piece of da market Nvidia stay controlling. 🏭🚀🤑
Chips going be one major part in A.I., to da point dey could change who get da powa among tech companies and maybe even countries. Like, da Biden administration stay considering restricting da sale of A.I. chips to China, cause some American officials saying China’s A.I. abilities could be one threat to da U.S. by making Beijing’s military and security stronger. 🇺🇸🇨🇳💣
Had A.I. supahcomputahs built befo’, including by Nvidia, but not too often dat start-ups go make ’em.
Cerebras, from Sunnyvale, Calif., got started in 2016 by Feldman and four other engineers. Dey like make hardware dat speeds up A.I. development. Over da years, da company wen’ raise $740 million, even from Sam Altman, who heads up da A.I. lab OpenAI, and venture capital firms like Benchmark. Right now, Cerebras valued at $4.1 billion. 🧪💰🚁
‘Cause da chips usually used fo’ power A.I. stay small — often da size of one postage stamp — take hundreds or even thousands of ’em fo’ process one complicated A.I. model. In 2019, Cerebras wen’ show off what dey say was da biggest computer chip eva made, and Feldman say their chips can train A.I. systems between 100 and 1,000 times as fast as da old kine hardware. 📬🧮🏎️
G42, da company from Abu Dhabi, wen’ start working with Cerebras in 2021. Dey used one Cerebras system in April fo’ train one Arabic version of ChatGPT. 🌐🤝🗣️
NOW IN ENGLISH
🖥️🎛️💥 Wow, The New A.I. Supercomputer is Making Waves, Run by Huge Computer Chips!
Okay, so this past week, inside a large room in a single-story building over there in Santa Clara, Calif., there were these tall, six-and-a-half foot machines buzzing away behind some white cabinets. These machines were all part of a new supercomputer that just became operational the month before. 👨💻🏢⚡
Cerebras, a Silicon Valley startup, unveiled the supercomputer this past Thursday. They created this computer with their own special kind of chips, designed just for powering A.I. products. These chips are big, like the size of a dinner plate, or 56 times larger than the regular kind of chips used for A.I. Every single Cerebras chip contains the power of hundreds of the traditional chips. 🎉🍽️💡
Cerebras said they built the supercomputer for G42, a company focused on A.I. G42 plans to use the supercomputer for creating and powering A.I. products for the Middle East. 🌍🧠🚀
Andrew Feldman, the CEO of Cerebras, said they want to show everyone that it’s possible to build a very large, dedicated A.I. supercomputer, and that it can be done faster, with less energy, and at a lower cost. 💸⚙️🏋️
This year, everyone wants to have computing power and A.I. chips, all due to the A.I. boom happening around the world. Major tech companies like Microsoft, Meta, and Google, and countless startups, have all been trying to release A.I. products after the A.I.-powered ChatGPT chatbot went viral because of its eerily human-like speech. 🌐🤖💬
However, making A.I. products usually requires a significant amount of computing power and specialized chips, leading to a fierce rush to find more of those technologies. Back in May, Nvidia, the leader in manufacturing chips for A.I., announced that the demand for their products — known as graphics processing units, or GPUs — was so strong that their quarterly sales were expected to be more than 50 percent above Wall Street’s predictions. This announcement made Nvidia’s market value soar past $1 trillion. 📈💲🔥
For the first time, we’re seeing a significant increase in computer requirements due to A.I., according to Ronen Dar, a founder of Run:AI, a startup over in Tel Aviv that helps companies develop A.I. models. This has “created a huge demand” for specialized chips, and companies are rushing to get their hands on them. 🏁📊🧗
To secure enough A.I. chips, some of the biggest tech companies — including Google, Amazon, Advanced Micro Devices, and Intel — have developed their own alternatives. Startups like Cerebras, Graphcore, Groq, and SambaNova are also trying to break into the market that Nvidia has dominated. 🏭🚀🤑
Chips are set to play such a key role in A.I. that they could change the balance of power among tech companies and even nations. The Biden administration, for example, is currently considering restricting the sale of A.I. chips to China, with some American officials saying that China’s A.I. capabilities could pose a national security threat to the U.S. by strengthening Beijing’s military and security apparatus. 🇺🇸🇨🇳💣
A.I. supercomputers have been built before, including by Nvidia, but it’s rare for startups to do so.
Cerebras, which is based in Sunnyvale, Calif., was founded in 2016 by Feldman and four other engineers with the goal of building hardware that speeds up A.I. development. Over the years, the company has raised $740 million, including from Sam Altman, who leads the A.I. lab OpenAI, and venture capital firms like Benchmark. Cerebras is currently valued at $4.1 billion. 🧪💰🚁
Because the chips typically used to power A.I. are small — often the size of a postage stamp — it takes hundreds or even thousands of them to process a complex A.I. model. In 2019, Cerebras unveiled what it claimed was the largest computer chip ever built, and Feldman has said that their chips can train A.I. systems between 100 and 1,000 times as fast as the existing hardware. 📬🧮🏎️
G42, the company from Abu Dhabi, started working with Cerebras in 2021. They used a Cerebras system in April to train an Arabic version of ChatGPT. 🌐🤝🗣️
