Centralized vs. decentralized digital networks: Key differences

Published

3 years ago

June 11, 2022

Find out what differentiates centralized and decentralized digital networks—and which structure is right for your organization.

Stablecoins: Depegging, fraudsters and decentralization

Published

22 minutes ago

May 2, 2025

Opinion by: Merav Ozair, PhD

Lately, stablecoins are everywhere — this time around, headed by “traditional” financial institutions. Bank of America and Standard Chartered are considering launching their own stablecoin, joining JPMorgan, which launched its stablecoin, JPM Coin — rebranded as Kinexys Digital Payments — to facilitate transactions with their institutional clients on their blockchain platform, Kinexys (formerly Onyx).

Mastercard plans to bring stablecoins to the mainstream, joining Bleap Finance, a crypto startup. The aim is to enable stablecoins to be spent directly onchain — without conversions or intermediaries — seamlessly integrating blockchain assets with Mastercard’s global payment rails.

In early April 2025, Visa joined the Global Dollar Network (USDG) stablecoin consortium. The company will become the first traditional finance player to join the consortium. In late March 2025, NYSE parent Intercontinental Exchange (ICE) announced that it is investigating applications for using USDC (USDC) stablecoin and US Yield Coin within its derivatives exchanges, clearinghouses, data services and other markets.

Why the renewed interest in stablecoins?

Regulatory clarity and acceptance

Recent moves by regulatory bodies in the United States and Europe have created more straightforward guidelines for cryptocurrency use. In the US, Congress is considering legislation to establish formal standards for stablecoins, bolstering confidence among banks and fintech companies.

The European Union’s Markets in Crypto-Assets regulation requires that stablecoin issuers operating within the EU adhere to specific financial standards, including special reserve requirements and risk mitigation. In the UK, financial authorities plan to conduct consultations to draft rules governing stablecoin use, further facilitating their acceptance and adoption.

The Trump administration executive order 14067, “Strengthening American Leadership in Digital Financial Technology,” supports and “promotes the development and growth of lawful and legitimate dollar-backed stablecoins worldwide” while “prohibiting the establishment, issuance, circulation, and use of a CBDC within the jurisdiction of the United States.”

This executive order, followed by Trump’s World Liberty Financial company launching a stablecoin called USD1, signals that this is the era of stablecoins, particularly those pegged to the USD.

Do we need more stablecoins?

The stablecoin landscape

There are over 200 stablecoins, most pegged to the US dollar. Two established stablecoins dominate the stablecoin landscape. Tether’s USDt (USDT), the oldest stablecoin, launched in 2014 and USDC, launched in 2018, capturing 65% and 28% of stablecoins market cap, respectively — both are centralized fiat collateralized.

Recent: Crypto wanted to overthrow banks, now it’s becoming them in stablecoin fight

In third place, a relatively new one, USDe, launched in February 2024, holds about 2% of the stablecoin market cap and has an unconventional mechanism based on derivatives in the crypto market. Although it runs on a DeFi protocol on Ethereum, it incorporates centralized features since centralized exchanges hold the derivatives positions.

There are three primary mechanisms of stablecoins:

Centralized, fiat-collateralized: A centralized company maintains reserves of the assets in a bank or trust (e.g., for currency) or a vault (e.g., for gold) and issues tokens (i.e., stablecoins) that represent a claim on the underlying asset.

Decentralized, cryptocurrency-collateralized: A stablecoin is backed by other decentralized crypto assets. One example can be found in the MakerDAO stablecoin Dai (DAI), which is pegged to the US dollar and encapsulates the features of decentralization. While a central organization controls centralized stablecoins, no one entity controls the issuance of DAI.

Decentralized, uncollateralized: This mechanism ensures the stability of the coin’s value by controlling its supply through an algorithm executed by a smart contract. In some ways, this is no different from central banks, which also don’t rely on reserve assets to keep the value of their currency stable. The difference is that central banks, like the Federal Reserve, set a monetary policy publicly based on well-understood parameters, and its status as the issuer of legal tender provides the credibility of that policy.

Depegging, risk and fraudsters

Stablecoins are supposed to be stable. They were created to overcome the inherent volatility of cryptocurrencies. To maintain their stability, stablecoins should (1) be pegged to a stable asset and (2) follow a mechanism that sustains the peg.

If stablecoins are pegged to gold or electricity, they will reflect the volatility of these assets and thus may not be the best choice if you are seeking a no-risk (or close to no-risk) asset.

USDe maintains a peg to the USD through delta hedging. It uses short and long positions in futures, which generates a 27% yield annually — significantly higher than the 12% annual yield of other stablecoins pegged to the USD. Derivative positions are considered risky — the higher the risk, the higher the return. Therefore, it encapsulates an inherited risk due to its reliance on derivatives, which runs counter to the purpose of stablecoins.

Stablecoins have been around for more than a decade. During this time, there were no major depegging fiascos other than the case of Terra. The collapse of Terra was not the result of a reserve problem or mechanism but rather the act of fraudsters and manipulators.

TerraUSD (UST) had a built-in arbitrage mechanism between UST and the Terra blockchain native coin, LUNA. To create UST, you needed to burn LUNA.

To entice traders to burn LUNA and create UST, the creators of the Terra blockchain offered a 19.5% yield on staking, which is crypto terminology for earning 19.5% interest on a deposit, through what they called the Anchor protocol.

Such a high interest rate is simply not sustainable. Someone has to borrow at such a rate or above for the lender to receive 19.5% interest. This is how banks make their profit — they charge high interest on borrowing (such as mortgages or loans) and provide low interest on savings (such as a traditional savings account or a certificate of deposit account). Analysis of the Anchor protocol in January 2022 showed it was at a loss.

One of the allegations in the lawsuits against Terraform Labs’ founders is that the Anchor protocol was a Ponzi scheme.

In March 2025, Galaxy Digital reached a $200-million settlement with the New York Attorney General over claims the crypto investing company promoted the LUNA digital asset without disclosing its interest in the token.

In January 2025, Do Kwon, founder of Terra, was found liable for securities fraud and is facing multiple charges in the US, including fraud, wire fraud and commodities fraud. If regulators are interested in preventing future cases like Terra, they should focus on how to deter fraudsters and manipulators from issuing or engaging with stablecoins.

Decentralization: Rekindling the premise of Bitcoin

Most stablecoins are centralized assets collateralized. They are controlled by a company that could conduct unauthorized use of customers’ funds or falsely claim that reserves fully back a stablecoin.

To prevent companies’ misconduct, regulators should closely monitor these companies and set rules similar to securities laws.

Centralized stablecoins run counter to the notion of blockchain and the premise of Bitcoin. When Bitcoin was launched, it was supposed to be a payment platform free of intermediaries, not controlled by any company, bank or government — a decentralized mechanism — run by the people for the people.

If a stablecoin is centralized, it should follow the regulations of any other centralized asset.

Maybe it’s time to rekindle the premise of Bitcoin but in a more “stable” fashion. Developing an algorithmic, decentralized stablecoin that is free of any control of a company, bank or government and reviving the core notion of blockchain.

Opinion by: Merav Ozair, PhD.

This article is for general information purposes and is not intended to be and should not be taken as legal or investment advice. The views, thoughts, and opinions expressed here are the author’s alone and do not necessarily reflect or represent the views and opinions of Cointelegraph.

Coin Market

From digital identity to outer space: Projects push crypto use cases

Published

22 minutes ago

May 2, 2025

As the crypto space developed, blockchain use cases expanded from simple digital currencies and non-fungible tokens (NFTs) to more complex areas such as digital identity verification and telecommunications.

Ahead of the Token2049 event in Dubai, Cointelegraph spoke with Spacecoin CEO Stuart Gardner, Spacecoin founder Tae Oh, and Humanity Protocol founder Terrence Kwok to explore how they use blockchain to improve certain industries.

From addressing challenges like verification in the artificial intelligence era to bringing internet connectivity to developing countries, projects are integrating blockchain to solve problems in different industries.

Digital identity verification to combat the AI threat

As artificial intelligence developed, the technology brought improvements that people could benefit from. However, the technology was also adopted by malicious actors who used the tech to perform AI-assisted hack research and deepfake scams.

Kwok told Cointelegraph that just two years ago, the idea of having to prove you’re human seemed “crazy.” However, with today’s advancements in AI, it has become remarkably easy to fake being a real person.

“As for content, you can’t tell if it’s AI-generated or not. Video deepfakes, you cannot tell, right? Even documents. It’s super easy now to use AI to create a fake proof of address, a fake proof of balance for your bank statement. I think in the future it’s only going to get worse,” he said.

The executive also said that in the future, AI may also exist in the physical world through humanoids that might mimic human beings. In 2024, Tesla’s humanoid robot project was showcased on social media, highlighting developments in humanoid robotics.

Kwok said that the development of robots underscores the need for human identity verification even more. The executive said that this was why they launched the Humanity Protocol, which uses blockchain tech for digital identity verification.

“The internet is filled with bots, you know, it’s filled with AI agents. They’re great, but there’s also a need to be able to verify and check whether something or somebody is a person or not,” Kwok told Cointelegraph.

Terrence Kwok (left) and Cointelegraph’s Ezra Reguerra at the Dubai Polo and Equestrian Club. Source: Cointelegraph

Decentralized satellite network to combat the connectivity oligopoly

Apart from digital identity, blockchain technology is also being used to create a decentralized satellite network.

Gardner told Cointelegraph that at the moment, the satellite connectivity landscape is an oligopoly, a market structure where the industry is dominated by only a few large players.

The executive pointed out that Starlink and Amazon lead the race, while the EU and China are catching up. However, the big problem is that over 150 countries are lagging behind.

“They’re going to become reliant upon working with one of these oligopolies for their connectivity. And that poses a big issue for these people,” Gardner added.

On Nov. 1, Spacecoin unveiled a plan to launch a decentralized physical infrastructure network (DePIN) through a fleet of nanosatellites in space.

Oh told Cointelegraph that the Spacecoin idea came from the observation that the space industry is getting heavily commoditized. However, the executive said that it was possible for smaller companies or even individuals to launch their own satellites and start building constellations for connectivity.

The Spacecoin founder added that since different people or entities own each satellite, it’s essentially a “decentralized network.”

The executive said that they integrated crypto into the project to have a “trustless means of payment and data exchange.” Oh said that this was where the blockchain comes in.

Gardner (left) , Oh (middle) and Reguerra at the Crypto Polo event in Dubai. Source: Cointelegraph

Magazine: Crypto wanted to overthrow banks, now it’s becoming them in stablecoin fight

Coin Market

Artificial general intelligence (AGI): Can it really think like a human?

Published

1 hour ago

May 2, 2025

What is AGI?

When the lines blur between man and machine, you’re looking at artificial general intelligence (AGI). Unlike its counterpart, artificial narrow intelligence (ANI), which is the use of AI for solving individual problem statements, AGI represents artificial intelligence that can understand, learn and apply knowledge in a way that is indistinguishable from human cognition.

AGI is still theoretical, but the prospect of artificial intelligence being able to holistically replace human input and judgment has naturally attracted plenty of interest, with researchers, technologists and academics alike seeking to bring the concept of AGI to reality.

Yet another strand of prevailing research seeks to explore the feasibility and implications of AGI vs. ANI in a world increasingly shaped by AI capabilities.

Indeed, while ANI has already transformed various industries, AGI’s potential goes far beyond. Imagine a world where machines can not only assist humans in their tasks but also proactively understand the drivers behind specific tasks, predict outcomes, and autonomously create innovative solutions to achieve optimal results. This paradigm shift could revolutionize healthcare, education, transportation and countless other fields.

Why is AGI so powerful?

Unlike ANI, AGI is not confined to pre-programmed tasks or predefined responses within a limited domain. Instead, it has the potential to generate and apply knowledge across various contexts.

Imagine a self-driving car powered by AGI. It can collect a passenger from a train station but also personalize the journey with custom recommendations for pit stops, sightseeing avenues or navigating unfamiliar roads to arrive at the desired destination. And because it’s a machine, AGI would not experience fatigue and would continue learning and improving at exponential speeds.

Here’s a definition of AGI by Vitalik Buterin, who highlights the sheer potential of AGI:

The example highlights some interesting features of AGI, which include:

Learning capability: AGI can learn from experiences and improve its performance over time without a concerted effort by human programmers to perform additional data set training. This learning is not limited to specific tasks and instead encompasses a broad spectrum of activities.Problem-solving skills: AGI can solve complex problems by applying logical reasoning just as a human would. This includes consideration of non-traditional variables, such as emotional impact, which can highlight an even wider range of potential outcomes.Adaptability: AGI can adjust to new situations and environments without explicit programming, which means it can thrive in dynamic and unpredictable settings.Understanding and interpretation: AGI is equipped to comprehend natural language, abstract concepts and emotional nuance, allowing for sophisticated human-machine interactions.Did you know? Blockchain timestamps could serve as a legal memory for AGI systems, allowing future audits to determine exactly what an AGI knew — and when.

The pursuit of AGI: Where does it stand as of April 2025?

AGI is currently the science-fiction version of AI. However, while still theoretical, the sheer potential of the concept makes AGI the science fiction equivalent of artificial intelligence.

While existing models, such as ChatGPT, are constantly evolving and improving with each day, the journey to bringing AGI to life involves overcoming significant technical challenges, such as:

Defining the tech stack: The purely hypothetical nature of AGI makes it exceedingly difficult, if not altogether impossible, to determine the precise nature of the technological stack required for practical implementation.Neural networks: Advances in deep learning have propelled this field forward, but AGI would also require specialist neural networks that mimic the human brain’s structure to process information and introduce a layer of emotion and nuance.Natural language processing (NLP): Significant advances are required in the field of NLP to enable machines to better understand and generate human language, incorporating nuance, emotion and complexities. This includes a more complex analysis of language syntax, semantics and context, which is still evolving in traditional machine learning models that leverage NLP. Reinforcement learning: Using reward-based mechanisms to teach machines to make decisions would allow AGI to learn optimal behaviors through trial and error.

Despite advancements, creating AGI that can truly think like a human remains an elusive goal.

Did you know? DeepMind warns that not all AI risks come from the machines themselves — some start with humans misusing them. In its paper titled ‘An Approach to Technical AGI Safety and Security’, DeepMind identifies four key threats: misuse (bad actors using AI for harm), misalignment (AI knowingly going against its developer’s intent), mistakes (AI causes harm without realizing it), and structural risks (failures that emerge from complex interactions between people, organizations, or systems).

Can AGI think like a human?

The question of whether AGI can think like a human delves into the very core of human cognition. Human thinking is characterized by consciousness, emotional depth, creativity and subjectivity. While AGI can simulate certain aspects of human thought, replicating the full spectrum of human cognition is a formidable challenge.

Several dimensions of human cognition are particularly difficult to emulate:

Consciousness and self-awareness: One of the defining traits of human thinking is consciousness, the awareness of oneself and one’s surroundings. AGI, as sophisticated as it may become, lacks the intrinsic human ability to introspect. AGI operates on an underlying set of algorithms and complex, learned patterns, without any subjectivity or genuine emotion.Emotional intelligence: Humans experience a wide range of emotions that influence their decisions, behaviors and interactions. While AGI can be trained to recognize and respond to such emotions, the lack of genuine emotional experience means that it cannot wholly replicate these emotions. Emotional intelligence in humans involves empathy, compassion and moral considerations, elements that are challenging to encode into machines.Creativity and innovation: Creativity involves generating novel ideas and solutions, often through intuitive leaps and imaginative thinking. AGI can mimic creativity by combining existing knowledge in new ways, but it lacks the intrinsic motivation and subjective insight that drive human innovation. True creativity stems from emotional experiences, personal reflections and cultural contexts, which AGI cannot authentically replicate.

Key benefits of AGI

The litmus test for AGI lies in its ability to holistically replicate a human experience. When realized, the potential benefits are enormous and stretch across various industries, spawning various aspects of daily life.

Despite its limitations, AGI is increasingly viewed as a force for good across a range of industries, including:

Healthcare: AGI can assist in diagnosing diseases, developing personalized treatment plans and predicting customized health outcomes, leveraging a vast body of underlying training data.Education: It can provide customized learning experiences, tutoring and academic research support. AGI can adapt to individual learning styles and pace, enhancing educational outcomes.Economics: It can optimize financial models, predict market trends, and enhance productivity. It can analyze economic data to forecast market trends and guide investment decisions.Environmental Science: AGI can analyze climate data, model ecological impacts, and propose sustainable solutions.

Additionally, AGI’s potential extends to areas such as transportation, communication and entertainment, offering new frontiers for innovation.

Did you know? Some futurists believe AGI systems could eventually negotiate with each other autonomously using blockchain-based smart contracts — forming agreements, trading data or even co-developing solutions without human intervention.

Ethical and societal considerations

The rise of AGI raises significant ethical and societal questions.

While powerful, AGI requires careful consideration for safe usage, which has prompted the creation of nonprofit societies, such as the AGI Society, as shown in the image below.

Fundamentally, it is crucial to address concerns such as:

Safety: Ensuring AGI operates within safe and controlled parameters to prevent unintended consequences. This includes robust testing and the introduction of regulatory frameworks to oversee AGI deployment.Privacy: Protecting personal data from misuse by AGI systems. As AGI can process vast amounts of data, safeguarding privacy is paramount.Bias and fairness: Preventing discriminatory practices and ensuring equitable access to AGI benefits. Developers must ensure that AGI systems are free from biases that could lead to unfair treatment.Employment: Addressing the impact of AGI on job displacement and workforce dynamics. As AGI automates tasks, there is a need to consider its impact on employment and provide support for affected workers.

The integration of AGI into society requires a thoughtful approach to its governance, ensuring that it serves the common good and respects social values.

Can blockchain power AGI?

AGI could create computers as smart as humans, revolutionizing fields like cryptocurrency trading or market analysis. But AGI needs trust and fairness to work for everyone. Blockchain, the tech behind Bitcoin and Ethereum, offers a secure, transparent way to make this happen.

Here’s how blockchain can supercharge AGI with crypto-inspired solutions:

Clear training records: Blockchain works like Bitcoin’s open transaction log, recording every piece of data (e.g., crypto trading patterns) used to train AGI. This helps ensure the system is fair and free from hidden biases.Shared decision-making: Similar to Ethereum’s smart contracts, blockchain will allow developers, traders and users to vote on AGI’s rules, ensuring no single company controls it.Safe data sharing: Like crypto wallets safeguarding funds, blockchain could protect sensitive data from crypto exchanges, allowing secure sharing for AGI training without leaks.Rewards for fairness: Developers who build unbiased AGI, such as accurate trading predictors, could earn digital tokens, just like crypto mining rewards, encouraging ethical work.

However, ongoing challenges such as blockchain’s slow speed, delays in crypto transactions and limited storage capacity could make it hard for AGI to process data quickly or handle large datasets.

To make blockchain AGI-ready, researchers are already exploring:

Offchain storage: Decentralized systems like InterPlanetary File System (IPFS) are used to store large files offchain, while the blockchain keeps only verifiable hashes, reducing congestion.Sharding and danksharding: Like Ethereum’s scalability upgrades, sharding splits data across multiple nodes, allowing AGI to process more information without slowing down the network. Also, danksharding, an advanced form of sharding being developed for Ethereum, combines rollups and data availability sampling to scale data access efficiently — ideal for real-time AGI applications.Data pruning: Advanced blockchain models like Decentralized Artificial Intelligent Blockchain-based Computing Network (DAIBCN) prune old or irrelevant data, keeping the system lean and optimized for high-demand tasks like AGI. DAIBCN also enables secure, distributed AI computing — blending blockchain trust with AI performance.

The future of AGI

Artificial general intelligence represents the pinnacle of AI development, promising capabilities that rival human intellect.

While AGI can simulate aspects of human thinking, achieving true human-like cognition remains a distant goal. Consciousness, emotional depth and creativity are intrinsic to human experience and pose significant challenges for AGI.

Nevertheless, the pursuit of AGI continues to drive innovation and reshape our understanding of intelligence. As we advance toward this frontier, it is imperative to navigate ethical considerations and societal impacts to responsibly harness AGI’s potential.

Ongoing research, identifying practical opportunities and technical requirements, and initiating dialogue across society are all essential steps to address the challenges and opportunities posed by AGI.

The future of AGI holds promise, but it requires a balanced approach to ensure that its eventual integration into society enhances human well-being and respects ethical standards.