Top related persons:
Top related locs:
Top related orgs:

Search resuls for: "SemiAnalysis"


22 mentions found


Analyst firm CCS Insight is predicting a "cold shower" for generative AI in 2024. The firm's chief analyst told CNBC he believes the technology is overhyped and faces immense costs to deploy. AdvertisementAdvertisementAn analyst firm is predicting a "cold shower" for generative AI in 2024. "Just the cost of deploying and sustaining generative AI is immense," said Wood. AdvertisementAdvertisementFor context, AI relies on chips to run, and the firm's prediction comes amid concerns about a global chip shortage.
Persons: , Ben Wood, Wood, Elon Musk, Dylan Patel Organizations: Insight, CNBC, Service, CCS, CCS Insight, Google, Elon, Nvidia, Reuters Locations: London
Google is preparing to launch its answer to rival OpenAI's GPT-4: Gemini. Gemini is a next-gen, multimodal AI model due for release later this year. The tech is a next-gen, multimodal AI model being worked on by a team of researchers pulled from Google's now-merged AI divisions DeepMind and Google Brain. Gemini is multimodalGoogle's Gemini is a multimodal AI, meaning it can process more than one type of data. Researchers behind the SemiAnalysis blog have also predicted that Google's Gemini would likely outperform GPT-4 because of Google's access to top-flight chips.
Persons: OpenAI's GPT, OpenAI's, Sam Altman, AlphaGo Gemini, Google's DeepMind, AlphaGo, Lee Sedol, ChatGPT, Demis Hassabis, DeepMind, Bard Organizations: Google, Service, OpenAI, AlphaGo, Wired Locations: Wall, Silicon, Google's
[1/2] A smartphone with a displayed Arm Ltd logo is placed on a computer motherboard in this illustration taken March 6, 2023. The tech investment behemoth raised nearly $5 billion from Arm's offering while retaining 90.6% of the firm. Known for debt-fuelled acquisition sprees, SoftBank founder and CEO Masayoshi Son flagged in June that the company was shifting back into "offence mode" as he highlighted the potential of artificial intelligence. That's after a year of "defence mode" when tech valuations crashed amid higher interest rates and global banking jitters. Few companies in SoftBank's investment portfolio have demonstrated commercial utility in AI, analysts said.
Persons: Dado Ruvic, behemoth, Masayoshi, Yoshimitsu Goto, SoftBank, SemiAnalysis, Kyle Stanford, There's, Amir Anvarzadeh, PitchBook's Stanford, Anton Bridge, Miyoung Kim, Edwina Gibbs Organizations: REUTERS, Rights, Arm Holdings, SoftBank, HK, Vision, Nvidia, Asymmetric Advisors, Thomson Locations: British
[1/2] A smartphone with a displayed Arm Ltd logo is placed on a computer motherboard in this illustration taken March 6, 2023. The tech investment behemoth raised nearly $5 billion from Arm's offering while retaining 90.6% of the firm. Known for debt-fuelled acquisition sprees, SoftBank founder and CEO Masayoshi Son flagged in June that the company was shifting back into "offence mode" as he highlighted the potential of artificial intelligence. That's after a year of "defence mode" when tech valuations crashed amid higher interest rates and global banking jitters. Few companies in SoftBank's investment portfolio have demonstrated commercial utility in AI, analysts said.
Persons: Dado Ruvic, behemoth, Masayoshi, Yoshimitsu Goto, SoftBank, SemiAnalysis, Kyle Stanford, There's, Amir Anvarzadeh, PitchBook's Stanford, Anton Bridge, Miyoung Kim, Edwina Gibbs Organizations: REUTERS, Rights, Arm Holdings, SoftBank, HK, Vision, Nvidia, Asymmetric Advisors, Thomson Locations: British
The researchers behind the SemiAnalysis blog say Google's upcoming AI Gemini smashes GPT-4. That might explain OpenAI boss Sam Altman's defensive response to a post published over the weekend titled : "Google Gemini Eats The World – Gemini Smashes GPT-4 By 5X, The GPU-Poors." Gemini is a next-gen, multimodal AI model being worked on by researchers at Google's AI arm DeepMind, and is expected to be released later in 2023. In response, SemiAnalysis' Patel posted on X that he got data on Google's GPU stores from a supplier of Google — rather than Google itself. But to say Gemini Smashes GPT-4 by 5x makes it sound like it is 5x better than GPT-4, it's not, its 5x compute.
Persons: Sam Altman, Sam Altman's, Dylan Patel, Daniel Nishball, Google's, Patel, OpenAI's, Altman, SemiAnalysis, Sundar Pichai, Sundar, G86ZRjnNmS, dtS0Bw3I92 — Dylan Patel, dethrones OAI, it's, ChatGPT Organizations: Google, Hacker
SemiAnalysis, a respected research firm, just divided the tech world into 2 groups. You now actually have to have the tech components, other infrastructure, and a smart plan to deploy this incredibly expensive gear. The GPU-PoorPatel and Nishball divide the tech industry into 2 main groups: The "GPU-poor" and the "GPU-rich." Then there are well-known AI firms, such as Hugging Face, Databricks, and Together, that are also GPU-poor. The internet giant is "the Most Compute Rich Firm In The World," with "unbeatably efficient architecture," Patel and Nishall wrote.
Persons: you've, Dylan Patel, Daniel Nishball, Patel, Jules Verne, Rich Patel, Nishball, Elon Musk's, Nishall Organizations: Morning, Nvidia, Google, Meta Locations: Saudi Arabia, UAE
The logo of Nvidia Corporation is seen during the annual Computex computer exhibition in Taipei, Taiwan May 30, 2017. Nvidia's stock price, though, has more than tripled this year and was set to hit an all-time high after Wednesday's results. The company has a near-monopoly on the computing systems used to power services like ChatGPT, OpenAI's blockbuster generative AI chatbot. Huang declined to comment on whether the AI boom will last past next year. The company said the biggest sales driver this quarter was its HGX system, which is an entire computer built around Nvidia's chip.
Persons: Tyrone Siu, Huang, Jensen Huang, quashing, Kinngai Chan, Chan, Dylan Patel, SemiAnalysis, Patel, We're, Stephen Nellis, Max Cherney, Chavi Mehta, Sonali Paul Organizations: Nvidia Corporation, REUTERS, Nvidia, Reuters, Microsoft, Summit, Thomson Locations: Taipei, Taiwan, OpenAI, San Francisco
Arm is hoping for a valuation of up to $70 billion in the IPO, which will launch on the Nasdaq next month. Arm and its owner SoftBank Group have set aside 10% of the shares to be sold in the IPO for its clients, the sources said. Its close relationship with Arm has helped it design chips that curbed its reliance on Intel as a supplier. Many technology companies that seek to make their own chips using Arm’s designs turn to TSMC for its low-cost manufacturing. None of these companies’ investments in Arm’s IPO are certain.
Persons: Dado Ruvic, , Jack Gold, SoftBank, Apple spokespeople, Jay Lee, Masayoshi Son, Son, TSMC, Dylan Patel Organizations: FRANCISCO, Reuters, Arm Holdings, REUTERS, Apple, Intel, Nvidia, Microsoft, Samsung Electronics, Nasdaq, Gold Associates, SoftBank, Devices Inc, APPLE, SAMSUNG, Samsung, MICROSOFT, Devices, Vision Fund Locations: Switzerland, Japan
Arm is hoping for a valuation of up to $70 billion in the IPO, which will launch on the Nasdaq next month. Arm and its owner SoftBank Group (9984.T) have set aside 10% of the shares to be sold in the IPO for its clients, the sources said. Its close relationship with Arm has helped it design chips that curbed its reliance on Intel as a supplier. Many technology companies that seek to make their own chips using Arm's designs turn to TSMC for its low-cost manufacturing. None of these companies' investments in Arm's IPO are certain.
Persons: Dado Ruvic, Jack Gold, SoftBank, Apple spokespeople, Jay Lee, Masayoshi Son, Son, TSMC, Dylan Patel, Milana Vinn, Stephen Nellis, Max Cherney, Greg Roumeliotis, Matthew Lewis Organizations: REUTERS, FRANCISCO, Arm Holdings, Apple, Intel, Nvidia, Microsoft, Samsung Electronics, Reuters, Nasdaq, Gold Associates, SoftBank, Devices Inc, APPLE, SAMSUNG, Samsung, MICROSOFT, Devices, Vision Fund, Thomson Locations: Switzerland, Japan, New York, San Francisco
So, that's the end of OpenAI's ChatGPT moat
  + stars: | 2023-07-20 | by ( Hasan Chowdhury | ) www.businessinsider.com   time to read: +6 min
In the fast-changing world of AI, OpenAI appeared to have created a powerful moat through ChatGPT. But Meta rolling out its own powerful language model, Llama 2, with few restrictions puts ChatGPT at risk. And unlike OpenAI's ChatGPT, Llama 2 doesn't come with an easy-to-use consumer interface. As the first popular entrant into the generative AI space, OpenAI's chatbot had the advantage of capturing the imaginations of millions before Google, Meta and the rest. But open models were always the threat.
Persons: OpenAI, chatbot, Mark Zuckerberg, Erin Scott, Warren Buffett, ChatGPT, Kali Hays, Rosalie Chan, Nathan Lambert, Lambert, It's Organizations: Meta, Google, Microsoft, ChatGPT, Stanford University, UC Berkeley
Lately, the giant AI model has become faster, but performance has declined. The world's most-powerful AI model has become, well, less powerful. It's considered the most-powerful AI model available broadly and is multimodal, which means it can understand images as well as text inputs. They think OpenAI is creating several smaller GPT-4 models that act similarly to the large model but are less expensive to run. This week, several AI experts posted what they claimed were details of GPT-4's architecture on Twitter.
Persons: OpenAI's, Peter Yang, I've, Frazier MacLeod, Christi Kennedy, OpenAI, ChatGPT, It's, Sharon Zhou, Theseus, Zhou, Yam, Semianalysis, George Hotz, Soumith Chintala, Oren Etzioni, Greg Brockman, " Brockman, Lilian Weng Organizations: Morning, Twitter, Roblox, Microsoft, Meta, Allen Institute, AI Locations: GPT
Jung Yeon-je | Afp | Getty ImagesSouth Korea's dominance in the memory chip market and a robust artificial intelligence ecosystem gives it an advantage in the global AI chip race, said industry observers. South Korea dominating in the memory market is definitely an advantage," said James Lim, senior research analyst at Dalton Investments. "South Korea seeks to emerge as a prominent player in rapidly growing and promising areas such as AI semiconductors," said Lee. "South Korea has a robust local AI ecosystem, capable of competing with global tech giants," said Sung Nako, executive for large scale AI development at South Korean internet giant Naver. ChatGPT maker OpenAI's CEO Sam Altman had urged South Korea to lead AI chip production during his meeting with South Korean President Yoon Suk-yeol in June.
Persons: Jung Yeon, James Lim, Lee, Dylan Patel, SemiAnalysis, ., TrendForce, Sung Nako, Sam Altman, Yoon Suk, Altman, Dalton's Lim, Geoffrey Cain Organizations: Getty, Dalton Investments, CNBC, Samsung, SK Hynix South, Samsung Electronics, SK Hynix, Science, Micron, South, South Korean, Nvidia, Intel Locations: Seoul, Korea, South Korea, China, U.S
Ted Cannis, a senior executive at Ford, told the Financial Times in December that there is a "large-scale rethinking of logistics operations" across the auto supply chain. "The supply chain is going to be the focus of this decade," Cannis said. Among the companies Apple is relying on to make the Vision Pro is Taiwan's Foxconn — which is the main supplier shifting its supply chain away from China. The move was made after China's COVID-19 lockdowns rocked supply chain and production timelines, but prices are the real driving reason behind the move. "Right now, robustness of our supply chain also needs to be considered to ensure the stable procurement of parts."
Persons: , Donald Trump, lockdowns, COVID, Ashutosh Sharma, Forrester, Ted Cannis, Cannis, China's, Wellsenn, Cowell, Liu Young, TSMC, Morris Chang, Chips, Dylan Patel, SemiAnalysis, Tim Cook, Fang DongxuFeature, Mazda, China's COVID, Masahiro Moro, ", Moro Organizations: Mazda, Service, Privacy, East, Ford, Financial Times, Apple, Vision, Apple's, Cowell e Holdings, Future Publishing, Reuters Locations: China, India, Zhengzhou, Henan, Vietnam, Mexico, Taiwan, Arizona, Nanjing, Fang DongxuFeature China, Japan Japanese, Japan
Google and OpenAI are not positioned to win the AI arms race, a Google engineer said in a leaked doc. A document written by a senior Google engineer suggests the open-source community is rivaling them. Google and OpenAI are not positioned to win a hotly contested AI arms race as the open-source community is "lapping" them, a senior engineer at the search giant has said. Sernau said open-source engineers were doing things with $100 that "we struggle with" at $10 million, "doing so in weeks, not months." He cited challenges around running AI models on a phone as one such major issue, which he said open-source engineers have managed to do on a Google Pixel 6.
OpenAI's losses climbed to $540 million in 2022 as it developed ChatGPT, a new report claims. The Information said that the startup's costs soared in the months before it launched the chatbot. AI tools like ChatGPT are expensive given the significant computing power needed to run them. OpenAI's development of the technology came at a steep price, with its losses roughly doubling to around $540 million last year, a new report claims. OpenAI's revenue is projected to rise significantly this year, with expectations that revenue will hit $200 million this year before climbing to $1 billion in 2024, according to Reuters.
ChatGPT could cost OpenAI up to $700,000 a day to run due to "expensive servers," an analyst told The Information. Microsoft is secretly building an AI chip to reduce the cost, per The Information. "Most of this cost is based around the expensive servers they require," Patel told the tech publication. Microsoft is reportedly working on a secret chipIn an effort to reduce the cost of running generative AI models, Microsoft is developing an AI chip it calls Athena, The Information first reported. Nearly four years later, more than 300 Microsoft employees are now reportedly working on the chip, according to the report.
Generative AI and large language models like OpenAI’s ChatGPT require massive amounts of computing power to run, and typically rely on chips like Nvidia’s graphics-processing units, or GPUs, that are specialized for these types of calculations. Graphcore sells primarily to AI startups looking to build and train models at lower cost, he said, and the company is benefiting from the proliferation of those startups. Shane Rau, who leads International Data Corp.’s semiconductor research, said chip startups are increasingly pivoting to focus their products on supporting large language models. Still, he added, “you’re going to see a combination of real adaptation and marketing.”“There will be the pressure to say: ‘Hey, we’re already relevant, our AI chip technology’s already relevant to generative AI’,” said Mr. Rau. Some chip makers say they expect yet another surge in demand once businesses more widely adopt generative AI.
Executives across the technology sector are talking about how to operate AI like ChatGPT while accounting for the high expense. What makes this form of AI pricier than conventional search is the computing power involved. Still, footing the bill is one of two main reasons why search and social media giants with billions of users have not rolled out an AI chatbot overnight, said Paul Daugherty, Accenture's chief technology officer. Technology experts also said a workaround is applying smaller AI models to simpler tasks, which Alphabet is exploring. The company said this month a "smaller model" version of its massive LaMDA AI technology will power its chatbot Bard, requiring "significantly less computing power, enabling us to scale to more users."
It's in response to the supply chains disruptions companies have faced in recent years. The headaches US corporations have faced are leading more of them to bring their supply chains closer to home and more under their control. Just under half (46%) said they were adjusting supply chains, and around a third (32%) said they were onshoring their workforce. But if more companies shift their supply chains closer to home, this could slowly begin to change. If reshoring companies have to raise wages to attract workers, they may decide to raise prices even further to protect their margins.
Taiwan-based TSMC, the world's biggest chipmaker, announced a $40 billion investment in Arizona last week. That's despite TSMC's founder previously calling US chip production an "expensive exercise in futility." In the event China — which claims the island as its own — invades the island and chip production screeches to a halt, there could be trillions of dollars in economic losses. First, the cost of chip production in the US might not ultimately be "50% more expensive." The factories will be partially subsidized by the US government through the CHIPS and Science Act, a package passed in August that provided $52 billion to boost US semiconductor chip production.
Experts disagree on how much this will protect the US economy in the event of a Chinese invasion of Taiwan. "These investments are helping us build and strengthen the supply chain here in America," Biden said, adding that "American manufacturing is back." "That's the definition of supply chain resilience. "Ultimately, creating a more resilient supply chain for semiconductors involves more than fab plants." To truly achieve supply chain resilience, Rasser says, the US must only boost production further, but needs similar investments in all areas of the chips supply chain, from raw materials to packaging.
The fate of the global economy may rest on the shoulders of one company: TSMC. That's because there could be trillions of dollars' worth of economic activity tied to that one company: Taiwan Semiconductor Manufacturing Company, the world's biggest chipmaker. Ann Wang/ReutersThe semiconductor industry has its roots in the US, as much of the research and development is done on US soil. According to a 2021 report from the Semiconductor Industry Association, in 1990 the US produced 37% of the world's chip supply. These days, the US is responsible for only 12% of global chip production.
Total: 22