Connect with us

Published

on

Google CEO Sundar Pichai speaks in conversation with Emily Chang during the APEC CEO Summit at Moscone West on November 16, 2023 in San Francisco, California. The APEC summit is being held in San Francisco and runs through November 17. 

Justin Sullivan | Getty Images News | Getty Images

Google is launching what it considers its largest and most capable artificial intelligence model Wednesday as pressure mounts on the company to answer how it’ll monetize AI.

The large language model Gemini will include a suite of three different sizes: Gemini Ultra, its largest, most capable category; Gemini Pro, which scales across a wide range of tasks; and Gemini Nano, which it will use for specific tasks and mobile devices.

For now, the company is planning to license Gemini to customers through Google Cloud for them to use in their own applications. Starting Dec. 13, developers and enterprise customers can access Gemini Pro via the Gemini API in Google AI Studio or Google Cloud Vertex AI. Android developers will also be able to build with Gemini Nano. Gemini will also be used to power Google products like its Bard chatbot and Search Generative Experience, which tries to answer search queries with conversational-style text (SGE is not widely available yet).

Gemini Ultra is the first model to outperform human experts on MMLU (massive multitask language understanding), which uses a combination of 57 subjects such as math, physics, history, law, medicine and ethics for testing both world knowledge and problem-solving abilities, the company said in a blog post Wednesday. It can supposedly understand nuance and reasoning in complex subjects.

Sundar Pichai, chief executive officer of Alphabet Inc., during the Google I/O Developers Conference in Mountain View, California, US, on Wednesday, May 10, 2023. 

David Paul Morris | Bloomberg | Getty Images

“Gemini is the result of large-scale collaborative efforts by teams across Google, including our colleagues at Google Research,” wrote CEO Sundar Pichai in a blog post Wednesday. “It was built from the ground up to be multimodal, which means it can generalize and seamlessly understand, operate across and combine different types of information including text, code, audio, image and video.”

Starting today, Google’s chatbot Bard will use Gemini Pro to help with advanced reasoning, planning, understanding and other capabilities. Early next year, it will launch “Bard Advanced,” which will use Gemini Ultra, executives said on a call with reporters Tuesday. It represents the biggest update to Bard, its ChatGPT-like chatbot.

The update comes eight months after the search giant first launched Bard and one year after OpenAI launched ChatGPT on GPT-3.5. In March of this year, the Sam Altman-led startup launched GPT-4. Executives said Tuesday that Gemini Pro outperformed GPT-3.5 but dodged questions about how it stacked up against GPT-4.

When asked if Google has plans to charge for access to “Bard Advanced,” Google’s general manager for Bard, Sissie Hsiao, said it is focused on creating a good experience and doesn’t have any monetization details yet. 

When asked on a press briefing if Gemini has any novel capabilities compared with current generation LLMs, Eli Collins, vice president of product at Google DeepMind, answered, “I suspect it does” but that it’s still working to understand Gemini Ultra’s novel capabilities.

Google reportedly postponed the launch of Gemini because it wasn’t ready, bringing back memories of the company’s rocky rollout of its AI tools at the beginning of the year.

Multiple reporters asked about the delay, to which Collins answered that testing the more advanced models take longer. Collins said Gemini is the most highly tested AI model that the company’s built and that it has “the most comprehensive safety evaluations” of any Google model.

Collins said that despite being its largest model, Gemini Ultra is significantly cheaper to serve. “It’s not just more capable, it’s more efficient,” he said. “We still require significant compute to train Gemini but we’re getting much more efficient in terms of our ability to train these models.”

Collins said the company will release a technical white paper with more details of the model on Wednesday but said it won’t be releasing the perimeter count. Earlier this year, CNBC found Google’s PaLM 2 large language model, its latest AI model at the time, used nearly five times the amount of text data for training as its predecessor LLM.

Also on Wednesday, Google introduced its next-generation tensor processing unit for training AI models. The TPU v5p chip, which Salesforce and startup Lightricks have begun using, offers better performance for the price than the TPU v4 announced in 2021, Google said. But the company didn’t provide information on performance compared with market leader Nvidia.

The chip announcement comes weeks after cloud rivals Amazon and Microsoft showed off custom silicon targeting AI.

During Google’s third-quarter earnings conference call in October, investors asked executives more questions about how it’s going to turn AI into actual profit.  

In August, Google launched an “early experiment” called Search Generative Experience, or SGE, which lets users see what a generative AI experience would look like when using the search engine — search is still a major profit center for the company. The result is more conversational, reflecting the age of chatbots. However, it is still considered an experiment and has yet to launch to the general public.

Investors have been asking for a timeline for SGE since May, when the company first announced the experiment at its annual developer conference Google I/O. The Gemini announcement Wednesday hardly mentioned SGE and executives were vague about its plans to launch to the general public, saying that Gemini would be incorporated into it “in the next year.”

“This new era of models represents one of the biggest science and engineering efforts we’ve undertaken as a company,” Pichai said in Wednesday’s blog post. “I’m genuinely excited for what’s ahead, and for the opportunities Gemini will unlock for people everywhere.”

— CNBC’s Jordan Novet contributed to this report.

Don’t miss these stories from CNBC PRO:

Continue Reading

Technology

Nvidia briefly surpasses $2 trillion in market cap during intraday trading

Published

on

By

Nvidia briefly surpasses  trillion in market cap during intraday trading

Nvidia CEO Jensen Huang speaks onstage during The New York Times Dealbook Summit 2023 at Jazz at Lincoln Center in New York City on Nov. 29, 2023.

Slaven Vlasic | Getty Images

Nvidia briefly surpassed $2 trillion in market cap during intraday trading Friday following the company’s rosy earnings report Wednesday — but it was short-lived.

After rising earlier in the day, shares of Nvidia were down about 1% at 11 a.m. ET. Nvidia stock closed up 16% Thursday.

Nvidia posted $22.10 billion in revenue for its fiscal fourth quarter, a 265% increase from a year ago and above the $20.62 billion expected by analysts polled by LSEG, formerly known as Refinitiv. Nvidia reported $12.29 billion in net income during the quarter, up a staggering 769% from $1.41 billion last year.

The company has benefited from the tech sector’s insatiable demand for artificial intelligence capabilities over the past year. Nvidia makes the pricey graphics processors for the servers that power large AI models.

Nvidia is clearly in a bullish trading cycle, says OptionPlay's Jessica Inskip

Nvidia said it expects $24.0 billion in sales in the current quarter, surpassing the $22.17 billion expected by analysts.

“Fundamentally, the conditions are excellent for continued growth,” Nvidia CEO Jensen Huang said during the company’s quarterly call with investors Wednesday.

— CNBC’s Kif Leswing contributed to this report.

Don’t miss these stories from CNBC PRO:

Continue Reading

Technology

AI can ‘disproportionately’ help defend against cybersecurity threats, Google CEO Sundar Pichai says

Published

on

By

AI can 'disproportionately' help defend against cybersecurity threats, Google CEO Sundar Pichai says

Google CEO Sundar Pichai speaks in conversation with Emily Chang during the APEC CEO Summit at Moscone West on November 16, 2023 in San Francisco, California. The APEC summit is being held in San Francisco and runs through November 17.

Justin Sullivan | Getty Images News | Getty Images

Munich, GERMANY — Rapid developments in artificial intelligence could help strengthen defenses against security threats in cyber space, according to Google CEO Sundar Pichai.

Amid growing concerns about the potentially nefarious uses of AI, Pichai said that the intelligence tools could help governments and companies speed up the detection of — and response to — threats from hostile actors.

“We are right to be worried about the impact on cybersecurity. But AI, I think actually, counterintuitively, strengthens our defense on cybersecurity,” Pichai told delegates at Munich Security Conference at the end of last week.

Cybersecurity attacks have been growing in volume and sophistication as malicious actors increasingly use them as a way to exert power and extort money.

Cyberattacks cost the global economy an estimated $8 trillion in 2023 — a sum that is set to rise to $10.5 trillion by 2025, according to cyber research firm Cybersecurity Ventures.

A January report from Britain’s National Cyber Security Centre — part of GCHQ, the country’s intelligence agency — said that AI would only increase those threats, lowering the barriers to entry for cyber hackers and enabling more malicious cyber activity, including ransomware attacks.

“AI disproportionately helps the people defending because you’re getting a tool which can impact it at scale.

Sundar Pichai

CEO at Google

However, Pichai said that AI was also lowering the time needed for defenders to detect attacks and react against them. He said this would reduce what’s known as a the defenders’ dilemma, whereby cyberhackers have to be successful just once to a system whereas a defender has to be successful every time in order to protect it.

“AI disproportionately helps the people defending because you’re getting a tool which can impact it at scale versus the people who are trying to exploit,” he said.

“So, in some ways, we are winning the race,” he added.

Google last week announced a new initiative offering AI tools and infrastructure investments designed to boost online security. A free, open-source tool dubbed Magika aims to help users detect malware — malicious software — the company said in a statement, while a white paper proposes measures and research and creates guardrails around AI.

Pichai said the tools were already being put to use in the company’s products, such as Google Chrome and Gmail, as well as its internal systems.

U.S. lawmakers reiterate support for Ukraine as President Zelenskyy calls for more aid

“AI is at a definitive crossroads — one where policymakers, security professionals and civil society have the chance to finally tilt the cybersecurity balance from attackers to cyber defenders. 

The release coincided with the signing of a pact by major companies at MSC to take “reasonable precautions” to prevent AI tools from being used to disrupt democratic votes in 2024’s bumper election year and beyond.

Adobe, Amazon, Google, IBM, Meta, Microsoft, OpenAI, TikTok and X, formerly Twitter, were among the signatories to the new agreement, which includes a framework for how companies must respond to AI-generated “deepfakes” designed to deceive voters.

It comes as the internet becomes an increasingly important sphere of influence for both individuals and state-backed malicious actors.

Former U.S. Secretary of State Hillary Clinton on Saturday described cyberspace as “a new battlefield.”

“The technology arms race has just gone up another notch with generative AI,” she said in Munich.

“If you can run a little bit faster than your adversary, you’re going to do better. That’s what AI is really giving us defensively.

Mark Hughes

president of security at DXC

A report published last week by Microsoft found that state-backed hackers from Russia, China, and Iran have been using its OpenAI large language model (LLM) to enhance their efforts to trick targets.

Russian military intelligence, Iran’s Revolutionary Guard, and the Chinese and North Korean governments were all said to have relied on the tools.

Mark Hughes, president of security at IT services and consulting firm DXC, told CNBC that bad actors were increasingly relying on a ChatGPT-inspired hacking tool called WormGPT to conduct tasks like reverse engineering code.

However, he said that he was also seeing “significant gains” from similar tools which help engineers to detect and reserve engineer attacks at speed.

“It gives us the ability to speed up,” Hughes said last week. “Most of the time in cyber, what you have is the time that the attackers have in advantage against you. That’s often the case in any conflict situation.

“If you can run a little bit faster than your adversary, you’re going to do better. That’s what AI is really giving us defensively at the moment,” he added.

Germany has been benefitting from a 'peace dividend' for years, defense minister says

Continue Reading

Technology

Ride-hailing giant Grab posts first profitable quarter, announces $500 million share buyback

Published

on

By

Ride-hailing giant Grab posts first profitable quarter, announces 0 million share buyback

A attendee walks past a banner with a Grab logo before a bell-ringing ceremony as Grab begins trading on the Nasdaq, in Singapore, on Thursday, Dec. 2, 2021.

Ore Huiying | Bloomberg | Getty Images

SINGAPORE — Grab posted its first-ever profitable quarter, raking in $11 million in profit, the Southeast Asian ride-hailing giant said in its fourth-quarter earnings report Thursday.

This compares with a $391 million loss recorded in the same period a year ago. The boost was “primarily due to the improvement in Group adjusted EBITDA, fair value changes in investments, and lowered share-based compensation expenses,” the company said.

Revenue for the quarter hit $653 million, exceeding LSEG analysts’ estimates of $634.86 million.

Losses for full year 2023 came to $485 million, down 72% from $1.74 billion a year ago.

In addition to ride-hailing, the company also provides financial services like payments and insurance, as well as deliveries for food, groceries and packages.

“We exited [2023 with] mobility exceeding pre-Covid levels. We are seeing a very strong demand in the mobility space,” Grab CFO Peter Oey told CNBC in an exclusive interview on Friday, adding that tourism is “growing very much.”

“If you look at the deliveries business, we have another record 13% year-over-year growth. We have now more users on our platform also at the same time. So we have really strong momentum,” he said on CNBC’s “Squawk Box Asia.”

Grab CFO says 2024 will be a year of investing

Grab announced Thursday it would be repurchasing up to $500 million worth of class A ordinary shares for the first time.

Grab was largely unprofitable during its years of operation, having amassed billions of dollars in losses since its inception in 2012.

In the initial years of business, tech startups tend to prioritize growth over profitability, which usually means burning a lot of cash. But with global macro uncertainties slowing growth, they have been forced to renew their focus on profitability and be more prudent with costs.

During the fourth quarter, total incentives — which include partner and consumer incentives — were further reduced to 7.3% of total value of goods sold, Grab said in its report. That’s compared to 8.2% in the same period a year ago “as we continued to improve the health of our marketplace.”

Grab had been doling out incentives to attract drivers and passengers to its platform but that’s tapering now as the company moves to drive up profitability.

On whether Grab would reach a time where it wouldn’t need to incentivize people to stay on the platform, Oey said incentives will “always be a lever” for the business.

“I don’t think we’re going to see a world where there’s no incentive whatsoever,” he told CNBC, adding that incentives help “to make sure we have enough supply” of drivers and attract price-sensitive customers.

For 2024, Grab expects revenue to come in between $2.70 billion and $2.75 billion, lower than LSEG analysts’ consensus of $2.8 billion.

Grab’s shares closed 8.41% lower on Thursday. Its share price has plummeted 75.8% from its $13.06 opening price in December 2021, when the firm first listed on the Nasdaq.

Continue Reading

Trending