Meta is building AI model to rival OpenAI’s most powerful system

September 11, 2023 Coin Telegraph

chatGPT

Meta is reportedly in the process of building a new, more powerful and open-source AI model to rival the most powerful systems of its rival OpenAI.

Meta, the parent company to social media platforms Facebook and Instagram, says it's developing a new artificial intelligence (AI) that will rival the most advanced model from OpenAI, according to a Wall Street Journal exclusive.

WSJ reported that individuals familiar with the matter said Meta aims for the new AI model to be “several times” more powerful than its Llama 2 model, which it released earlier this year.

For the moment, the WSJ sources say Meta’s plans for the new system are for it to be open-source, and therefore allow other companies to build AI tools to produce high-level text, analysis and other types of output.

The company has also been building data centers necessary to create such a high-level system while acquiring more of Nvidia’s H100 semiconductor chips - the most powerful and coveted chips currently available on the market.

Llama was trained on 70 billion parameters, and while OpenAI hasn’t released its parameters for GPT-4; it's estimated around 1.5 trillion.

The sources said Meta anticipates training to begin for the large language model (LLM) in early 2024 and to be ready for release sometime next year. It is likely to be released after Google’s expected forthcoming LLM Gemini.

Microsoft is a primary backer of OpenAI and also collaborated with Meta to help make Llama 2 available on Azure, its cloud-computing platform. However, the sources said Meta plans to train its upcoming model on its own infrastructure.

This development comes as major tech companies and governments are racing to create, deploy and control high-level AI systems.

Recently, the United Kingdom government announced that it plans to spend $130 million on high-powered chips to create AI systems.

Across the globe in China, the country’s new legislation on AI recently went into effect. Since then the CEO of Baidu, a major China-based tech company, said that over 70 AI models have been released in the country.

Magazine: AI Eye: Get better results being nice to ChatGPT, AI fake child porn debate, Amazon’s AI reviews

US regulators deny blocking AI chip exports to Middle East

September 1, 2023 Coin Telegraph

China

The U.S. Department of Commerce denied blocking AI chip sales to the Middle East and remained silent about whether the requirements were only imposed on specific countries.

The United States Department of Commerce said on Aug. 31 that the Biden administration has “not blocked chip sales to the Middle East,” according to a report from Reuters.

This comes after disclosures were revealed in a Nvidia report that the U.S. government had expanded requirements for export licenses for artificial intelligence (AI) chips.

Advanced Micro Devices (AMD), a direct competitor of Nvidia, also received a similar letter from regulators.

The Commerce Department made no comment as to whether the requirements were imposed on specific U.S. companies. However, the new rules would require Nvidia and AMD to obtain licenses prior to selling flagship chips to “some Middle Eastern countries,” according to the filing.

Neither of the two companies has revealed if they have applied for said licenses or if there was any feedback on licensing for that region.

The company warned regulators in the quarterly report that being “effectively excluded from all or part of China” could potentially “harm” long-term results for the company.

In October 2022, the Biden administration issued the initial export controls in an effort to slow China from developing high-level AI systems with powerful semiconductor chips made by U.S. companies.

Officials in Washington said they are considering tightening the aforementioned regulations even more in a statement made on June 29, which would further limit the computing power on chips available in the Chinese market.

The moves made by the U.S. government have been under close watch from other regulators around the world. Shortly after the U.S.’s initial regulations came into effect, an agreement was made with the Netherlands and Japan to restrict exports of semiconductor manufacturing equipment to China.

Officials in the United Kingdom, France and Germany have all openly said they are considering screening Chinese foreign direct investment in crucial sectors such as AI.

China has responded by saying it will control the export of gallium and germanium products, the primary raw materials needed to produce AI chips.

Magazine: AI Eye: Get better results being nice to ChatGPT, AI fake child porn debate, Amazon’s AI reviews

US officials extend export curbs on Nvidia AI chip to ‘some Middle Eastern countries’

August 31, 2023 Coin Telegraph

China

U.S. regulators have told AI chip maker Nvidia and its rival AMD to curb exports of high-level semiconductor chips used to develop AI to certain Middle Eastern countries.

Regulators in the United States have asked Nvidia to curb exports of artificial intelligence (AI) chips to “some Middle East countries,” according to its latest quarterly report.

In the report released on Aug. 28, the company said the new regulations affect its A100 and H100 chips, which help enhance the speed of machine-learning tasks. Nvidia did not specify what countries in the Middle East are impacted.

Advanced Micro Devices (AMD), a direct competitor of Nvidia, was also given a letter by U.S. regulators requesting the same ban on high-level AI chip exports to some Middle Eastern countries, according to a Reuters report.

The quarterly report from Nvidia states that the new regulatory filing would not have an “immediate material impact” on its business, nor does it make up a “meaningful portion” of its revenue.

However, in a separate statement, it said it is working with the U.S. government to address the matter.

The quarterly report did mention the ongoing AI chip export regulations the U.S. government has placed on China.

Nvidia said that past restrictions have still allowed them to sell alternative products in China, including their less powerful A800 or H800 chips.

However, the company warned that long-term results could be “harmed” if it is “effectively excluded from all or part of China.” Of its $13.5 billion in sales from the recent fiscal quarter ending on July 30, the majority came from the U.S., China and Taiwan, and about 13.9% came from other countries combined.

Initial export controls were implemented in October 2022 by the Biden administration in an effort to isolate China from powerful semiconductor chips.

On June 29, officials in Washington said they are considering tightening restrictions on AI chip exports to China even further and limiting the computing power of chips to stunt the flow of chip availability in the Chinese market.

In response to the measures taken by the U.S., the Chinese government said it will control the export of gallium and germanium products, which are primary components to produce AI chips.

AI-related regulations and restrictions coming from the U.S. have caused other countries to consider their own position in the race to develop powerful systems.

Magazine: AI Eye: Get better results being nice to ChatGPT, AI fake child porn debate, Amazon’s AI reviews

Iris Energy buys 248 Nvidia GPUs worth $10M for generative AI & Bitcoin mining

August 29, 2023 Coin Telegraph

Bitcoin mining

Iris Energy has invested $10 million in the latest generation Nvidia GPUs to explore generative AI while it continues to focus on Bitcoin mining.

Nasdaq-listed Iris Energy has bought 248 state of the art Nvidia H100 GPUs for $10 million as it looks to explore opportunities in generative AI in addition to its core business focus on Bitcoin mining.

The firm expects to receive delivery of the 248 GPUs in the coming months and plans to deploy the hardware to serve opportunities in cloud computing.

Iris Energy co-founder and co-CEO Daniel Roberts said the company was looking to leverage its existing data centers into serving generative AI computing requirements.

“We believe demand for sustainable computing is unlikely to go away, and feel we are uniquely positioned to capture ongoing growth in the broader industry; whether that be ASICs for Bitcoin mining, or GPUs for generative AI and beyond.”

Iris Energy operates in regions that have an abundance of renewable energy including wind, solar, hydro and has set up its modular data centers nearby to the source of low-cost excess renewable energy to be monetized for Bitcoin.

Nvidia's H100 Tensor Core GPU. Source: Nvidia.com.

According to the Iris Energy website, it has four major data center mining facilities, including Canal Flats, Mackenzie and Prince George in Canada’s British Columbia as well as its Childress site in Texas.

Renewable-powered Bitcoin mining operations continue to attract investment, with Genesis Digital Assets Limited opening a new data centre in Sweden in August 2023 that will operate off abundant power from the nearby Porjus Hydroelectric Power Station.

Meanwhile Blockstream recently announced its intent to raise up to $50 million in an official investment note to purchase, store and then sell BTC mining hardware ahead of Bitcoin's next halving event in 2024.

GPU hardware manufacturer Nvidia has also seen significant windfall from the rise of AI-powered tools and AI computing, with its total market capitalization eclipsing $1 trillion in May 2023.

Nvidia also recently teased its next-generation GH200 Grace Hopper Superchip which is touted to be able to process complex generative AI workloads, includling large language models, recommender systems and vector databases.

Magazine: Recursive inscriptions: Bitcoin ‘supercomputer’ and BTC DeFi coming soon

UK to spend $130M on AI chips amid scramble to buy up computing power

August 21, 2023 Coin Telegraph

UK to spend 0M on AI chips amid scramble to buy up computing power

The chips will be used to set up an AI resource as a recent report said 20% of firms can’t get enough computing power for AI.

British Prime Minister Rishi Sunak is set to spend $130 million (100 million pounds) to buy thousands of computer chips to power artificial intelligence, amid a global shortage and race for computing power.

The Telegraph reported on Aug. 20 that the United Kingdom aims to build an "AI Research Resource" by mid-2024 as part of Sunak’s plan to make the country an AI tech hub.

The government is reportedly sourcing chips from makers NVIDIA, Intel and AMD — and it is understood that the science funding body UK Research and Innovation — which is leading the effort — is in the late stages of ordering 5,000 NVIDIA graphic processing units (GPUs).

However, while $130 million has been allocated to the project, the funds are reportedly seen as insufficient to match Sunak’s AI hub ambition, meaning government officials could pressure for more funding in an upcoming November AI safety summit.

It follows a recent report that said many companies are struggling to deploy AI due to available resources and technical obstacles.

In March, an independent review of the country’s AI computing capabilities said investment in the space is “seriously lagging” behind international counterparts in the United States and European Union.

At the time, less than 1,000 NVIDIA chips were available for researchers to train AI models — a panel recommended the U.K. make available at least 3,000 top-quality chips to meet immediate needs.

On Aug. 16, S&P Global’s global AI trend report found that many firms reported they’re not ready to support AI, due to not having enough computing power, along with challenges managing data and security concerns.

We teamed with S&P Global to launch a global survey of #AI practitioners & decision-makers to see what’s driving or hindering their AI projects.

Today, we published the survey’s findings in this new 2023 Global Trends In AI report. Get the skinny here https://t.co/k9W5AQ9hwg
— WEKA (@WekaIO) August 16, 2023

While it's still early days for AI — S&P senior research analyst Nick Patience said a deciding factor for who will lead in the space will be decided by who can support AI workloads.

AI Eye: Apple developing pocket AI, deep fake music deal, hypnotizing GPT-4

Nvidia drops new AI chip expected to cut development costs

August 9, 2023 Coin Telegraph

Coin Telegraph

Nvidia, the current market leader for semiconductor chips, revealed its latest chip designed to run large-scale AI systems.

Nvidia, one of the world’s leading developers of semiconductor chips, revealed its latest chip on Aug. 7, designed to power high-level artificial intelligence (AI) systems.

The company said its next-generation GH200 Grace Hopper Superchip is one of the first to be equipped with an HBM3e processor, and is designed to process “ the world’s most complex generative AI workloads, spanning large language models, recommender systems and vector databases.”

Jensen Huang, the CEO of Nvidia, commented in a keynote that it is giving the processor a “boost” and that:

“This processor is designed for the scale-out of the world’s data centers.”

While the GH200 has the same general processing unit as the H100 — the company’s most high-end chip and one of the top in the world — it comes with 141 gigabytes of advanced memory and a 72-core ARM central processor, which is at least three times more powerful than the previous chip.

The latest chip from Nvidia is designed for inference, one of the two primary components of working with AI models after training them. Inference is when the model is used to generate content, make predictions and is constantly running.

Huang said “pretty much any” large language model (LLM) can be run through this chip, and it will “inference like crazy.”

“The inference cost of large language models will drop significantly.”

The GH200 becomes available in the second quarter of 2024, according to Huang, and should be available for sampling by the end of 2023.

This development comes as Nvidia’s market dominance is currently being challenged by the emergence of new semiconductor chips from rival companies racing to create the most powerful products.

Presently, it has over an 80% market share for AI chips and briefly topped $1 trillion in market value.

On May 28, Nvidia introduced a new AI supercomputer for developers to create successors in the style of ChatGPT, with Big Tech companies like Microsoft, Meta and Google’s Alphabet expected to be among the first users.

However, on June 14, Advanced Micro Devices (AMD) released information on its forthcoming AI chip with the capabilities and capacity to challenge Nvidia’s dominance. The AMD chip is set to be made available in the third quarter of 2023.

Most recently, on Aug. 3, the chip developer Tenstorrent received $100 million in a funding round led by Samsung and Hyundai in an effort to diversify the chip market.

Magazine: Experts want to give AI human ‘souls’ so they don’t kill us all

AI chip developer gets $100 million from Samsung and Hyundai

August 3, 2023 Coin Telegraph

AI chip developer gets 0 million from Samsung and Hyundai

The AI chip developing company Tenstorrent just closed a funding round worth $100 million, which was led by manufacturing giants Samsung and Hyundai.

The artificial intelligence (AI) chip developer Tenstorrent announced the closing of a $100-million funding round on Aug. 3, which was led by major manufacturing companies Samsung and Hyundai, among others.

The company said it raised $30 million from Hyundai, $20 million from the automobile manufacturer Kia, and $50 million from the Catalyst Fund by Samsung and other participating investors in the round, including Fidelity Ventures, Eclipse Ventures and Maverick Capital.

Tenstorrent is led by its CEO, Jim Keller, who previously developed AI chips for major tech companies, such as Apple, Intel and Tesla.

Heung-soo Kim, executive vice president and head of the global strategy office at Hyundai Motor Group, commented on the group’s expectations from the investment:

“... the Group expects to develop optimized but differentiated semiconductor technology that will aid future mobilities and strengthen internal capabilities in AI technology development.”

Tenstorrent also said the funds will be used to “accelerate the company’s product development, the design and development of AI chiplets,” along with advancements on the roadmap for its machine learning software.

These initiatives come as developers are racing to compete with Nvidia, the current market leader in AI chip development.

The growing demand for high-powered semiconductor chips to develop AI systems briefly shot Nvidia’s market value into the trillion-dollar range.

On June 13, Advanced Micro Devices (AMD) announced details of its forthcoming AI chip that has the potential to also challenge Nvidia’s market dominance.

AMD, Nvidia and Tenstorrent are all United States-based companies that are currently subject to the ongoing chip export restrictions that are primarily targeting China, one of the major markets for the chips.

Initially, United States President Joe Biden issued restrictions on AI chips to China in October 2022 in order to slow down the semiconductor industry. The administration announced that it’s considering tightening the restrictions in July.

Shortly after the U.S. made this statement, China said it plans to tighten controls on exports of the materials needed to make the chips in the first place.

Magazine: Deposit risk: What do crypto exchanges really do with your money?

U.S. considers tightening restrictions on AI chip exports

June 29, 2023 Coin Telegraph

Officials in the U.S. are considering restricting the level of computing power on chips to lessen the flow of AI chips available in the Chinese market.

Officials in the United States are mulling over tighter controls on an export regulation that was created to minimize the availability of artificial intelligence (AI) chips to China.

According to officials close to the source, the stricter regulations would include clamping down on the level of computing power chips can have that are able to be exported. The sources say an update to the rules could come by late July.

Such restrictions on the sale of powerful computing chips have caused alarm to some of the industry’s major players.

Colette Kress, the chief financial officer of Nvidia, one of the world’s leading chip makers, commented on June 28 at an investors conference that:

“..if implemented, [restrictions] would result in a permanent loss of opportunities for U.S. industry to compete and lead in one of the world’s largest markets..."

Kress said an implementation of such regulations would not “immediately impact” the company’s financial results. In late May, the AI chip boom caused Nvidia to momentarily hit $1 trillion in value.

Cointelegraph reached out to the U.S. Department of Commerce for further comment on its potential decision.

Initially, the restrictions against AI chip sales in China were issued by the Biden administration in October 2022 with the intention to slow down the semiconductor industry.

The October ban cut off Chinese developers from access to some of the more advanced chips on the market, including Nvidia’s A100 chips and the latest version, the H100. These two chips are among the most sought-after for high-level AI development.

In May Nvidia reported its second-quarter revenue forecast to be 50% higher than market estimates, along with a 28% increase in company shares.

It was around the same time that the company released additional AI-powered tools including an AI supercomputer that it created to help aid developers to produce ChatGPT successors.

Meanwhile, in China, local developers have been figuring out ways to skirt the impact of U.S. sanctions. Companies have been reported to be studying new methods to develop AI chips through the use of weaker semiconductors and combinations of chips currently available to them.

Magazine: AI Eye: 25K traders bet on ChatGPT’s stock picks, AI sucks at dice throws, and more

Crypto Biz: Six months on from FTX, Tether mines BTC, and Nvidia’s AI superchips

June 2, 2023 Coin Telegraph

Binance

This week’s Crypto Biz also looks at the crypto industry after FTX collapse, Tether’s Bitcoin mining in Latin America, Tabi’s funding round and Nvidia’s artificial intelligence machines.

Just over six months after FTX’s dramatic collapse, the crypto industry can finally begin analyzing the effects of the debacle. The quick ripple effect to other crypto businesses drained liquidity from the industry and prolonged the crypto winter, with Silvergate Bank, BlockFi and Genesis Global Capital among those hit by the exchange’s collapse.

FTX’s bankruptcy has also affected the crypto regulatory landscape, with authorities cracking down on firms — employing controversial methods in some cases — to avoid a deepening blend of traditional finance with cryptocurrencies.

Companies that closed their United States operations citing regulatory pressure in the past months included Bittrex, Nexo and Unbanked, to name a few. Coinbase CEO Brian Armstrong said this week that China stands to benefit most from restrictive crypto policies in the U.S., but only time will tell if this is true.

Companies are also reviewing their business operations due to increased regulatory scrutiny. In response to crypto firms being debanked, Binance has even considered buying a bank in the past months, said its CEO Chanpeng Zhao. Now, the crypto exchange is gearing up for a layoff that will boost its compliance and regulatory capabilities.

While the industry digests the recent events, FTX’s new management claims FTX 2.0 could be launched as soon as next year, hopefully in time to join the club of crypto companies striving to remain in business after November 2022.

This week’s Crypto Biz also looks at Tether’s Bitcoin (BTC) mining operations in Latin America, Tabi’s funding round and Nvidia’s efforts to power the next generation of artificial intelligence (AI) machines.

Buying a bank won’t solve crypto’s debanking issue — Binance CEO

Binance is unlikely to buy up any banking institutions, but it plans to make minority investments in financial institutions that will “hopefully influence them to be more crypto-friendly,” commented Zhao on the growing worry of crypto companies being debanked. The collapse of several U.S. banks in 2023 has prompted concerns that the pool of crypto-friendly banks is shrinking. Former key banking partners, Silvergate, Silicon Valley Bank and Signature Bank, have all capitulated this year. The exchange is also reportedly exploring a solution to reduce counterparty risk by allowing institutional clients to keep their trading collateral at a bank instead of on the crypto platform.

Tether moves into Bitcoin mining in Uruguay

Stablecoin issuer Tether has announced it will launch Bitcoin mining operations in Uruguay in collaboration with a local licensed company. According to Tether, the venture would utilize renewable energy sources aimed at “sustainable” Bitcoin mining and planned to hire additional team members. The mining announcement followed Tether’s plan to “regularly allocate up to 15%” of its profits into BTC purchases. Tether cited Uruguay’s capability of generating 94% of its electricity from renewable sources like wind, solar and hydropower, and its reliable grid. Job listings on its website also suggested expansion into South Africa and Brazil.

Nvidia introduces AI supercomputer to create ChatGPT successors

Nvidia continues to push forward in the race to develop AI tools and applications as the company revealed plans to release more products. Its CEO Jensen Huang recently unveiled a new AI supercomputer platform called DGX GH200 that will aid tech companies in developing successors to the popular AI chatbot ChatGPT. Big Tech firms such as Microsoft, Meta and Alphabet are anticipated to be among some of the pioneering users of the supercomputer equipment. Also developing its own AI chip is Microsoft, which claims it intends to deal with the rising costs of development for in-house and OpenAI projects.

BNB NFT marketplace Tabi raises $10 million in angel funding

Nonfungible token (NFT) marketplace Tabi, previously known as Treasureland, has completed a $10 million angel funding round seeded by venture capital firms Animoca Brands, Draper Dragon, Hashkey Capital, Infinity Crypto Ventures and Youbi Capital. Along with NFT trading and launchpad features, Tabi converts users’ on-chain activities into “experience points,” which can be exchanged for future airdrop rewards and earnings. The protocol also contains a gaming platform aggregating blockchain game transactions and entertainment. Funds will be primarily used to develop Tabi’s gaming ecosystem and construct an on-chain identity protocol.

Crypto Biz is your weekly pulse of the business behind blockchain and crypto, delivered directly to your inbox every Thursday.

Investor Who Called End of Bear Market Doubles Down on Crypto, Says Industry Undervalued and Overlooked

May 31, 2023 The Daily Hodl

Nvidia

Buying a bank won’t solve crypto’s debanking issue — Binance CEO

Tether moves into Bitcoin mining in Uruguay

Nvidia introduces AI supercomputer to create ChatGPT successors

BNB NFT marketplace Tabi raises $10 million in angel funding

Share this video