Connect with us

Published

on

David Silver, leader of the reinforcement learning research group at DeepMind, being awarded an honorary “ninth dan” professional ranking for AlphaGo.
JUNG YEON-JE | AFP | Getty Images

Computer scientists are questioning whether DeepMind, the Alphabet-owned U.K. firm that’s widely regarded as one of the world’s premier AI labs, will ever be able to make machines with the kind of “general” intelligence seen in humans and animals.

In its quest for artificial general intelligence, which is sometimes called human-level AI, DeepMind is focusing a chunk of its efforts on an approach called “reinforcement learning.”

This involves programming an AI to take certain actions in order to maximize its chance of earning a reward in a certain situation. In other words, the algorithm “learns” to complete a task by seeking out these preprogrammed rewards. The technique has been successfully used to train AI models how to play (and excel at) games like Go and chess. But they remain relatively dumb, or “narrow.” DeepMind’s famous AlphaGo AI can’t draw a stickman or tell the difference between a cat and a rabbit, for example, while a seven-year-old can.

Despite this, DeepMind, which was acquired by Google in 2014 for around $600 million, believes that AI systems underpinned by reinforcement learning could theoretically grow and learn so much that they break the theoretical barrier to AGI without any new technological developments.

Researchers at the company, which has grown to around 1,000 people under Alphabet’s ownership, argued in a paper submitted to the peer-reviewed Artificial Intelligence journal last month that “Reward is enough” to reach general AI. The paper was first reported by VentureBeat last week.

In the paper, the researchers claim that if you keep “rewarding” an algorithm each time it does something you want it to, which is the essence of reinforcement learning, then it will eventually start to show signs of general intelligence.

“Reward is enough to drive behavior that exhibits abilities studied in natural and artificial intelligence, including knowledge, learning, perception, social intelligence, language, generalization and imitation,” the authors write.

“We suggest that agents that learn through trial and error experience to maximize reward could learn behavior that exhibits most if not all of these abilities, and therefore that powerful reinforcement learning agents could constitute a solution to artificial general intelligence.”

Not everyone is convinced, however.

Samim Winiger, an AI researcher in Berlin, told CNBC that DeepMind’s “reward is enough” view is a “somewhat fringe philosophical position, misleadingly presented as hard science.”

He said the path to general AI is complex and that the scientific community is aware that there are countless challenges and known unknowns that “rightfully instill a sense of humility” in most researchers in the field and prevent them from making “grandiose, totalitarian statements” such as “RL is the final answer, all you need is reward.”

DeepMind told CNBC that while reinforcement learning has been behind some of its most well-known research breakthroughs, the AI technique accounts for only a fraction of the overall research it carries out. The company said it thinks it’s important to understand things at a more fundamental level, which is why it pursues other areas such as “symbolic AI” and “population-based training.”

“In somewhat typical DeepMind fashion, they chose to make bold statements that grabs attention at all costs, over a more nuanced approach,” said Winiger. “This is more akin to politics than science.”

Stephen Merity, an independent AI researcher, told CNBC that there’s “a difference between theory and practice.” He also noted that “a stack of dynamite is likely enough to get one to the moon, but it’s not really practical.”

Ultimately, there’s no proof either way to say whether reinforcement learning will ever lead to AGI.

Rodolfo Rosini, a tech investor and entrepreneur with a focus on AI, told CNBC: “The truth is nobody knows and that DeepMind’s main product continues to be PR and not technical innovation or products.”

Entrepreneur William Tunstall-Pedoe, who sold his Siri-like app Evi to Amazon, told CNBC that even if the researchers are correct “that doesn’t mean we will get there soon, nor does it mean that there isn’t a better, faster way to get there.”

DeepMind’s “Reward is enough” paper was co-authored by DeepMind heavyweights Richard Sutton and David Silver, who met DeepMind CEO Demis Hassabis at the University of Cambridge in the 1990s.

“The key problem with the thesis put forth by ‘Reward is enough’ is not that it is wrong, but rather that it cannot be wrong, and thus fails to satisfy Karl Popper’s famous criterion that all scientific hypotheses be falsifiable,” said a senior AI researcher at a large U.S. tech firm, who wished to remain anonymous due to the sensitive nature of the discussion.

“Because Silver et al. are speaking in generalities, and the notion of reward is suitably underspecified, you can always either cherry pick cases where the hypothesis is satisfied, or the notion of reward can be shifted such that it is satisfied,” the source added.

“As such, the unfortunate verdict here is not that these prominent members of our research community have erred in any way, but rather that what is written is trivial. What is learned from this paper, in the end? In the absence of practical, actionable consequences from recognizing the unalienable truth of this hypothesis, was this paper enough?”

What is AGI?

While AGI is often referred to as the holy grail of the AI community, there’s no consensus on what AGI actually is. One definition is it’s the ability of an intelligent agent to understand or learn any intellectual task that a human being can.

But not everyone agrees with that and some question whether AGI will ever exist. Others are terrified about its potential impacts and whether AGI would build its own, even more powerful, forms of AI, or so-called superintelligences.

Ian Hogarth, an entrepreneur turned angel investor, told CNBC that he hopes reinforcement learning isn’t enough to reach AGI. “The more that existing techniques can scale up to reach AGI, the less time we have to prepare AI safety efforts and the lower the chance that things go well for our species,” he said.

Winiger argues that we’re no closer to AGI today than we were several decades ago. “The only thing that has fundamentally changed since the 1950/60s, is that science-fiction is now a valid tool for giant corporations to confuse and mislead the public, journalists and shareholders,” he said.

Fueled with hundreds of millions of dollars from Alphabet every year, DeepMind is competing with the likes of Facebook and OpenAI to hire the brightest people in the field as it looks to develop AGI. “This invention could help society find answers to some of the world’s most pressing and fundamental scientific challenges,” DeepMind writes on its website.

DeepMind COO Lila Ibrahim said on Monday that trying to “figure out how to operationalize the vision” has been the biggest challenge since she joined the company in April 2018.

Continue Reading

Technology

Embattled grocery startup Getir exits the U.S. and Europe, will refocus on Turkey

Published

on

By

Embattled grocery startup Getir exits the U.S. and Europe, will refocus on Turkey

Companies such as Getir and Gorillas promise to deliver items to shoppers’ doors in as little as 10 minutes.

Angel Garcia | Bloomberg via Getty Images

Grocery delivery startup Getir announced on Monday that it is quitting international markets including the U.K., Germany, the Netherlands and the U.S., marking a major setback for the once hyped online grocery industry.

The Istanbul, Turkey-based firm said in a statement that it was withdrawing from its U.S. and European markets and would now refocus its financial resources on Turkey.

The company said it raised a new investment round led by Abu Dhabi sovereign wealth fund Mubadala and venture capital firm G Squared “to bolster its competitive position in its core food and grocery delivery businesses in Turkey.”

Getir said it generates 7% of its revenues from the U.K., Germany, the Netherlands and the U.S.

“Getir expresses its sincere appreciation for the dedication and hard work of all its employees in the UK, Germany, the Netherlands, and the U.S.,” the company said.

Pandemic grocery hype fades

Struggling space

Continue Reading

Technology

Tesla jumps 10% in premarket trading after passing key hurdle to roll out full self-driving in China

Published

on

By

Tesla jumps 10% in premarket trading after passing key hurdle to roll out full self-driving in China

SpaceX owner and Tesla CEO Elon Musk arrives on the red carpet for the Axel Springer Award 2020 on Dec. 1, 2020 in Berlin, Germany.

Britta Pedersen | Getty Images

Shares of Tesla rose sharply in U.S. premarket trading on Monday after the electric car maker passes a significant milestone to roll out its full self-driving technology in China.

The company’s share price spiked more than 10% just after 7:30 a.m. ET, as investors reacted to news surrounding Tesla CEO Elon Musk’s visit to China.

Tesla on Sunday said that local Chinese authorities removed restrictions on its cars after passing the country’s data security requirements.

The move raised expectations that Tesla’s driver-assistance software Full Self Driving (FSD) would soon be available in the country, which is the largest market for electric vehicles.

While Tesla’s electric cars are some of the most popular vehicles in China, they have reportedly been banned from some government-related properties due to data security concerns.

Separately, the Biden administration earlier this year announced a probe into whether imported cars from China pose national security risks due to their ability to potentially collect sensitive data.

FSD is an upgrade to Tesla’s Autopilot driver assistant. Tesla has offered its FSD technology in China for years, but with a restricted feature set that limits it to operations, such as automated lane changing.

Data security concerns have been a key obstacle preventing Tesla from achieving a full rollout of the system in China.

Tesla also reportedly scored a deal with Baidu that would give Musk’s firm access to the Chinese internet giant’s mapping and navigation technology for Tesla’s FSD feature.

The agreement would allow Tesla to tap into Baidu’s mapping service license, which is a requirement for intelligence driving systems to operate on public roads in China, Reuters reported, citing two anonymous sources familiar with the matter.

CNBC was unable to independently verify the report. Tesla and Baidu were not immediately available for comment.

With the license, which foreign companies can only clinch in partnership with local Chinese firms, Tesla will be allowed to legally operate FSD on Chinese roads, and its fleets will be able to gather data about traffic, road signs and routes.

The breakthrough for Tesla toward bringing its FSD self-driving technology to China marks a key win for the firm at a time when it is facing hefty competition in the Chinese market. Local rivals such as Warren Buffett-backed electric vehicle maker BYD, Nio, and Xpeng have ramped up their competition with Tesla in recent years.

BYD was temporarily the largest electric vehicle maker globally, producing more than 3 million new energy vehicles in 2023. The firm recently lost its crown as world’s largest EV maker, after a 43% plunge in sales in the first quarter.

– CNBC’s Evelyn Cheng contributed to this report

Continue Reading

Technology

Oracle boosts its generative AI capabilities as cloud competition heats up

Published

on

By

Oracle boosts its generative AI capabilities as cloud competition heats up

US multinational computer technology company Oracle’s logo is pictured at the Mobile World Congress (MWC), the telecom industry’s biggest annual gathering, in Barcelona on February 27, 2024. The world’s biggest mobile phone fair throws open its doors in Barcelona with the sector looking to artificial intelligence to try and reverse declining sales. (Photo by PAU BARRENA / AFP) (Photo by PAU BARRENA/AFP via Getty Images)

Pau Barrena | Afp | Getty Images

U.S. cloud infrastructure provider Oracle is boosting its generative AI capabilities as cloud competition intensifies and more companies jump into AI.

The AI boom — fueled by the launch of chatbot ChatGPT in November 2022 — is driving an increase in demand for cloud computing services and data centers, as large amounts of data are required in AI model training and the cloud provides access to vast datasets.

Oracle has been introducing generative AI capabilities into its cloud infrastructure and applications to complement the traditional AI already embedded in them.

“The classic AI is very good in terms of detecting patterns or predicting numbers … but you cannot use large language models to predict numbers,” Rondy Ng, executive vice president of applications development at Oracle, told CNBC.

“So we combined the predictive numbering capability with the explained ability in words. So the two together become very powerful and you need both. In the past many years, the number prediction part is already very mature. As part of the product we continue to evolve that and it’s not going to stop. Generative AI is basically the talk of the town right now,” said Ng.

In March, Oracle announced additional generative AI features embedded across applications in finance, supply chain, human resources, sales, marketing, and service. The generative AI capabilities can perform tasks such as generating financial reports and drafting job ads, improving productivity and reducing business costs, Oracle said.

This comes after the firm announced the implementation of generative AI across its technology stack in January.

“We believe Oracle is seeing a renaissance of growth with its AI strategy. [It is] well positioned to be a major beneficiary of the AI revolution,” said Dan Ives, managing director of Wedbush Securities, in emailed comments to CNBC on Wednesday.

“The data Oracle sits on and installed base gives Ellison & co. a major advantage to monetize the software layer of AI,” said Ives, referring to Oracle’s chairman and chief technology officer Larry Ellison.

As firms talked up the generative AI story last year, technology providers have to be one step ahead of the cycle, research firm Gartner said in a report on April 17. “They are bringing GenAI capabilities to existing products and services, as well as to use cases being identified by their enterprise clients.”

JPMorgan has said generative AI and AI could drive incremental IT spending and growth across the software landscape. “Many software vendors, including Oracle, have cited benefits from ongoing investments by businesses into AI technologies,” JPMorgan analysts said in a note on March 12.

Oracle might see an increase in revenue and positive impact on its shares if the company manages to capture a larger-than-expected share of the spending into AI, the U.S. investment bank said. Oracle’s shares have spiked 23.74% in the last 12 months, according to FactSet data.

“Generative AI services [are] basically a huge advantage comparing with our competition. The competition needs to work with different companies and cloud providers for that infrastructure and those kinds of services. We actually take everything into an integrated stack, and we consume that,” Ng told CNBC.

AI growth

Oracle has lagged behind rivals like Amazon, Microsoft and Google in cloud infrastructure service market share, according to Synergy Research Group, which ranked Oracle as the sixth-largest service provider, alongside IBM, globally.

While Oracle was late to cloud infrastructure, the AI boom has increased demand for the company’s AI technology. Ellison had in 2018 dismissed cloud computing as “complete gibberish.”

“Oracle did follow the hyperscalers. [I think] that’s not a competitive concern, say for the rest of 2024 and in the foreseeable future. We’re at the very beginning stage of this whole new generative AI journey,” said Ron Westfall, research director at Futurum Group.

CEO Safra Catz said in March the company added several “large new cloud infrastructure” contracts during the fiscal third quarter. Cloud revenue rose 25% year over year to $5.1 billion, Oracle said.

“Interesting to us is management commentary suggesting its Oracle Cloud Infrastructure backlog is significant and AI isn’t yet really driving revenue, which is expected to be more meaningful in FY25,” said Deutsche Bank analysts on Mar. 12.

Cloud players can monetize AI quicker than other companies, says CFRA's Zino on Microsoft earnings

Ellison said in March that a Salt Lake City data center that Oracle is building can fit eight Boeing 747 airplanes nose-to-tail.

Laying out future market opportunities, Ellison said he sees more national and state government applications being run on platforms like Oracle Cloud Infrastructure, and added that the firm is negotiating sovereign regions with a number of countries.

“Another area [where Oracle] is ahead of the curve, although everybody’s jumping on it, is in terms of offering sovereign AI cloud – a cloud that operates exclusively within a country,” said Westfall.

“More and more countries are going to say when it comes to gen AI, we want all that information, all that data stored within the country.”

In April, Oracle said it would invest more than $8 billion in Japan over the next 10 years to grow cloud computing and AI infrastructure.

Oracle and Nvidia in March announced they will be partnering up to deliver sovereign AI solutions to customers around the world.

Continue Reading

Trending