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

Public companies bought more bitcoin than ETFs did for the third quarter in a row

Published

on

By

Public companies bought more bitcoin than ETFs did for the third quarter in a row

Ozan Kose | Afp | Getty Images

Corporate treasuries have surpassed ETFs in bitcoin buying for a third consecutive quarter as more companies try to benefit from the MicroStrategy playbook in a more crypto-friendly regulatory environment.

Public companies acquired about 131,000 coins in the second quarter, growing their bitcoin balance 18%, according to data provider Bitcoin Treasuries. ETFs showed an 8% increase or about 111,000 BTC in the same period.

“The institutional buyer who is getting exposure to bitcoin through the ETFs are not buying for the same reason as those public companies who are basically trying to accumulate bitcoin to increase shareholder value at the end of the day,” said Nick Marie, head of research at Ecoinometrics. 

Public company bitcoin holdings increased 4% in April, a tumultuous month after the market was rocked by President Donald Trump’s initial tariffs announcement, versus 2% for ETFs, he pointed out.

“They don’t really care if the price is high or low, they care about growing their bitcoin treasury so they look more attractive to the proxy buyers,” Marie added. “It’s not so much driven by the macro trend or the sentiment, it’s for different reasons. So it becomes a different kind of mechanism that can push bitcoin forward.”

Bitcoin ETFs, whose collective U.S. launch in January 2024 was one of the most successful ETF debuts in history, still represent the largest holders of bitcoin by entity with more than 1.4 million coins held today, representing about 6.8% of the fixed supply cap of 21 million. Public companies hold about 855,000 coins, or about 4%.

Regulatory relief

The trend reflects the significant regulatory relief the crypto industry broadly is benefiting from under the Trump administration. In March, Trump signed an executive order for a U.S. bitcoin reserve, sending a strong message that the flagship cryptocurrency, which has long been a source of reputation risk among many investors, is here to stay. The last time ETFs outpaced public companies in bitcoin buying was in the third quarter of 2024, before Trump was re-elected.

In the second quarter, GameStop began buying bitcoin, after its board approved it as a treasury reserve asset in March; health-care company KindlyMD merged with Nakamoto, a bitcoin investment company founded by crypto entrepreneur David Bailey; and investor Anthony Pompliano’s ProCap, kicked off its own bitcoin purchasing program and is going public through a special purpose acquisition company, or SPAC.

Strategy, recently rebranded from MicroStrategy, is still the main behemoth in the bitcoin treasury game. The company pioneered the strategy that more than 140 public companies globally are now emulating. It holds about 597,000 BTC, and is followed by the bitcoin miner Mara Holdings, which has almost 50,000 coins.

“It’s going to be very hard to catch Strategy’s scale,” said Ben Werkman, chief investment officer at Swan Bitcoin. “They’re going to be the preferred landing spot for institutional capital because of the deep liquidity around their equity, while these smaller equities are going to be really good risk returns for retail investors and smaller institutions that want more of that upside – that initial growth that comes in kicking off the strategy – because a lot of people missed it with MicroStrategy.”

A long-term case?

Marie suggested that 10 years from now, there probably won’t be so many companies committed to the bitcoin treasury strategy. Firstly, he said, the more that enter the category, the more diluted the activity at each firm becomes. Plus, bitcoin may be so normalized by then that proxy buyers are no longer constrained by rules and mandates around direct exposure to bitcoin.

“You can think about this wave as a bunch of companies that are trying to benefit from this arbitrage,” Marie said.

Werkman pointed out that most investors that are attracted to bitcoin treasury companies today already have a thesis around bitcoin. For them, leveraged bitcoin equities are likely how they try to outperform bitcoin itself, the foundational component of their investments.

“What people really like about these companies, and why they like to get into these smaller companies, is because they can do something that the investors holding spot bitcoin can’t do: go and accumulate more bitcoin on your behalf because they have access to the capital markets and can issue securities,” Werkman said.

There’s also likely to be a fair number of companies that convert their existing treasury holdings to bitcoin without pursuing leverage the way Strategy does, Werkman noted.

“They’ve got that ability to generate more and more value behind their shares, backed by bitcoin, plus whatever the operations of the company are generating. It’s a unique value proposition,” he said.

Don’t miss these cryptocurrency insights from CNBC Pro:

Continue Reading

Technology

AeroVironment stock drops 7% on offering plan to pay off debt

Published

on

By

AeroVironment stock drops 7% on offering plan to pay off debt

An image of a Quantix drone made by AeroVironment.

David Mcnew | Getty Images News | Getty Images

AeroVironment shares fell 7% Tuesday after the defense contractor said it plans to offer $750 million in common stock and $600 million in convertible senior notes due in 2030 to repay debt.

The drone maker said it would use leftover funding for general purposes such as boosting manufacturing capacity.

AeroVironment shares have soared 85% this year, ballooning its market value to about $13 billion.

Last week, shares of the Arlington, Virginia-based company rallied on strong fourth-quarter results, lifting higher as CNBC’s Jim Cramer called it the “next Palantir of hardware.”

Read more CNBC tech news

Last month, the company also closed its $4.1 billion acquisition of space-related defense tech company Blue Halo.

Earlier this month, President Donald Trump signed an executive order intended to boost drone production in the U.S. and crack down on unauthorized uses.

The company also has a high short interest level, which may have contributed to some of the recent gains, creating a short squeeze. This phenomenon occurs when a stock price surges, forcing those shorting the stock to purchase shares to cover their positions and prevent losses.

Stock Chart IconStock chart icon

hide content

Aerovironment 1 month stock chart

AeroVironment CEO on European defense spending boost, U.S. defense spending and Trump

Continue Reading

Technology

Web giant Cloudflare to block AI bots from scraping content by default

Published

on

By

Web giant Cloudflare to block AI bots from scraping content by default

Jaque Silva | Nurphoto | Getty Images

Internet firm Cloudflare will start blocking artificial intelligence crawlers from accessing content without website owners’ permission or compensation by default, in a move that could significantly impact AI developers’ ability to train their models.

Starting Tuesday, every new web domain that signs up to Cloudflare will be asked if they want to allow AI crawlers, effectively giving them the ability to prevent bots from scraping data from their websites.

Cloudflare is what’s called a content delivery network, or CDN. It helps businesses deliver online content and applications faster by caching the data closer to end-users. They play a significant role in making sure people can access web content seamlessly every day.

Roughly 16% of global internet traffic goes directly through Cloudflare’s CDN, the firm estimated in a 2023 report.

“AI crawlers have been scraping content without limits. Our goal is to put the power back in the hands of creators, while still helping AI companies innovate,” said Matthew Prince, co-founder and CEO of Cloudflare, in a statement Tuesday.

“This is about safeguarding the future of a free and vibrant Internet with a new model that works for everyone,” he added.

What are AI crawlers?

AI crawlers are automated bots designed to extract large quantities of data from websites, databases and other sources of information to train large language models from the likes of OpenAI and Google.

Whereas the internet previously rewarded creators by directing users to original websites, according to Cloudflare, today AI crawlers are breaking that model by collecting text, articles and images to generate responses to queries in a way that users don’t need to visit the original source.

This, the company adds, is depriving publishers of vital traffic and, in turn, revenue from online advertising.

Read more CNBC tech news

Tuesday’s move builds on a tool Cloudflare launched in September last year that gave publishers the ability to block AI crawlers with a single click. Now, the company is going a step further by making this the default for all websites it provides services for.

OpenAI says it declined to participate when Cloudflare previewed its plan to block AI crawlers by default on the grounds that the content delivery network is adding a middleman to the system.

The Microsoft-backed AI lab stressed its role as a pioneer of using robots.txt, a set of code that prevents automated scraping of web data, and said its crawlers respect publisher preferences.

“AI crawlers are typically seen as more invasive and selective when it comes to the data they consumer. They have been accused of overwhelming websites and significantly impacting user experience,” Matthew Holman, a partner at U.K. law firm Cripps, told CNBC.

“If effective, the development would hinder AI chatbots’ ability to harvest data for training and search purposes,” he added. “This is likely to lead to a short term impact on AI model training and could, over the long term, affect the viability of models.”

WATCH: AI engineers are in high demand — but what is the job really like?

AI engineers are in high demand — but what is the job really like?

Continue Reading

Trending