<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Mustafa on Medium]]></title>
        <description><![CDATA[Stories by Mustafa on Medium]]></description>
        <link>https://medium.com/@mustafaxyz9?source=rss-5bd8d2734ef5------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Stories by Mustafa on Medium</title>
            <link>https://medium.com/@mustafaxyz9?source=rss-5bd8d2734ef5------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 06:08:34 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@mustafaxyz9/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[EigenLayer: How Restaking Will Transform Security for Ethereum-Based Protocols]]></title>
            <link>https://medium.com/@mustafaxyz9/eigenlayer-how-restaking-will-transform-security-for-ethereum-based-protocols-fd37c01be44e?source=rss-5bd8d2734ef5------2</link>
            <guid isPermaLink="false">https://medium.com/p/fd37c01be44e</guid>
            <category><![CDATA[proof-of-stake]]></category>
            <category><![CDATA[eigenlayer]]></category>
            <category><![CDATA[restaking]]></category>
            <category><![CDATA[ethereum]]></category>
            <category><![CDATA[eth]]></category>
            <dc:creator><![CDATA[Mustafa]]></dc:creator>
            <pubDate>Tue, 02 Jan 2024 16:56:51 GMT</pubDate>
            <atom:updated>2024-02-22T16:02:09.066Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Introducing EigenLayer’s restaking primitive, its effects on the Ethereum blockchain, the risks that emerge as a consequence, and its practical applications.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5yxq4sJfDvjlJDOCPVZlxw.png" /></figure><h3>Context</h3><p>For years, <a href="https://www.forbes.com/advisor/investing/cryptocurrency/proof-of-work/">Proof of Work</a> (PoW) consensus mechanisms have faced criticism for causing significant environmental damage. For instance, <a href="https://www.nature.com/articles/s41598-022-18686-8">estimates</a> indicate that Bitcoin mining consumed about 75 TWh of electricity per year in 2020, exceeding Portugal’s annual electricity usage of 48 TWh in the same year. Moreover, the environmental impact of Bitcoin mining has escalated, with <a href="https://www.nature.com/articles/s41598-022-18686-8">studies</a> showing that each BTC mined in 2021 was responsible for emitting 126 times the amount of Carbon Dioxide as each BTC in 2016.</p><p>Therefore, it was a major moment in the blockchain industry when Ethereum switched from the PoW consensus mechanism to <a href="https://ethereum.org/en/developers/docs/consensus-mechanisms/pos/">Proof of Stake</a> (PoS) in September 2022. PoS nodes, from an environmental perspective, use only a fraction of the energy consumed by their PoW counterparts, which engage in complex computations. In the PoS system, each validator node stakes 32 ETH and is responsible for verifying the validity of new blocks and transactions added to the blockchain. Validators receive additional ETH as compensation and incentive for their honest service. However, they also risk losing their staked ETH through a process known as ‘Slashing’ if they validate illegitimate transactions or fail to perform their duties when called upon.</p><p>Since 32 ETH represents a considerable sum and a significant barrier for many aspiring to become network validators, third parties have <a href="https://ethereum.org/en/staking/pools/">developed</a> staking pools. These pools facilitate a collaborative approach that allows individuals with smaller ETH holdings to combine their resources and collectively stake in shared validators. Staking pools not only make participation more accessible but also provide stakers with <a href="https://www.paradigm.xyz/2021/04/on-staking-pools-and-staking-derivatives">Liquid Staking Derivatives</a> (LSDs). These LSDs, representing staked ETH on a 1:1 basis, offer ETH stakers the benefits of retaining their ETH’s liquidity while staking to engage in other financial activities, such as using it as collateral for borrowing to potentially compound (increase) yield. This approach reduces the opportunity cost of ETH staking, enabling users to support network security while maintaining asset liquidity for other uses.</p><p>Several protocols have emerged to offer decentralized LSDs, like <a href="https://docs.lido.fi/">Lido’s</a> stETH and <a href="https://docs.rocketpool.net/overview/">Rocket Pool’s</a> rETH, while exchanges like <a href="https://www.coinbase.com/cbeth">Coinbase</a> have capitalized on their large user deposits to create centralized LSDs, such as cbETH. The growing popularity of liquid staking has taken the industry by storm, exemplified by Lido <a href="https://defillama.com/protocols">becoming</a> the largest DeFi protocol by Total Value Locked (TVL), exceeding $20 billion in TVL as of December 2023.</p><h3>The Problem</h3><p>Creating LSDs from staking pools is one strategy that enhances the utility of staked ETH. However, <a href="https://messari.io/report/eigenlayer-and-securing-new-middleware">many middleware services</a> on Ethereum, such as <a href="https://www.gemini.com/cryptopedia/crypto-oracle-blockchain-overview">Oracle Networks</a> (which provide external and real-world data to smart contracts), cannot directly benefit from staked ETH for securing their networks and services. This limitation stems from the fact that staking ETH primarily secures Ethereum’s main layer, not the subsidiary networks of these services, due to Ethereum’s technical constraints. Consequently, these services and their associated tokens have been unable to leverage Ethereum’s pooled trust. Instead, they have had to tackle the challenges of network growth independently. The process of securing these independent networks and validating their transactions is referred to as <a href="https://docs.eigenlayer.xyz/overview/key-terms">Actively Validated Services</a> (AVS).</p><p>Securing an AVS instead of Ethereum’s main-chain security presents <a href="https://docs.eigenlayer.xyz/overview/readme/whitepaper">several</a> limitations.</p><ol><li><em>Bootstrapping Trust</em>: Projects on Ethereum that develop their own AVSs face the challenge of building a new network of nodes willing to validate transactions and uphold network security. This is particularly challenging in the early stages when attracting users who expect robust security. AVSs often encounter a chicken-and-egg problem: potential nodes hesitate to join due to the perceived risk of a new network, yet the network’s security strengthens as more nodes join.</li><li><em>Value Leakage</em>: AVSs create new trust pools with separate nodes for transaction validation. As a result, users of these services incur additional fees separate from the transaction fees already paid on the Ethereum network. Consequently, instead of fees circulating within the Ethereum ecosystem, a portion of the value is extracted by the AVSs, leading to a loss in economic flow within Ethereum. Despite providing the infrastructure for AVSs, Ethereum does not fully capitalize on the economic benefits generated by these services.</li><li><em>Opportunity Cost</em>: Validators face opportunity costs when committing capital to secure new and risky AVSs. By choosing to stake in these projects, they forgo potentially safer investment options, such as staking ETH directly on the main chain. This decision carries high opportunity costs due to added risks such as token price volatility, network failure, and the potential for exploitative behavior.</li><li><em>Lower Trust</em>: Generally, AVSs are secured by smaller amounts of staked assets compared to the Ethereum network as a whole, due to their smaller scale. This disparity makes AVSs and middleware services more vulnerable targets than compromising the Ethereum network itself. Consequently, AVS services generally offer lower levels of trust and security compared to the main Ethereum network.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gd2PNBJ_hXcaGh8Wx1TpUw.png" /></figure><h3>Eigenlayer</h3><p><a href="https://docs.eigenlayer.xyz/overview/readme/whitepaper">Eigenlayer</a> was founded in 2021 by <a href="https://www.linkedin.com/in/sreeram-kannan-a397a47/">Sreeram Kannan</a>, a professor at the University of Washington, to address the limitations associated with AVSs. The company’s most <a href="https://www.coindesk.com/business/2023/03/28/staking-protocol-eigenlayer-raises-50m-amid-crypto-winter/">recent</a> $50 million Series A fundraising round in February 2023 was led by <a href="https://www.blockchaincapital.com/">Blockchain Capital</a>, one of the largest venture capital firms in the blockchain industry.</p><p>Eigenlayer identified an opportunity to leverage the substantial amounts of ETH already being staked to secure the Ethereum network and utilize it to secure additional projects, protocols, and AVS built on Ethereum. Validators currently staking ETH can opt into restaking (reusing) their existing ETH to secure these additional protocols. Additionally, if users are staking their ETH via liquidity staking pools like Lido, Eigenlayer offers them the option to stake their LSDs to secure the protocols instead. By participating in restaking, validators have the opportunity to earn additional rewards.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*caSbEuYqWFoLNb4TuEumyA.png" /></figure><h4>Slashing</h4><p>Although validators can freely restake their ETH into additional modules and compound their yield, crypto-economic security is maintained through slashing conditions specific to each module. This ensures that validators operate honestly when validating transactions or risk losing their ETH. While modules can tailor their slashing conditions to their network’s specific needs, the smart contracts on Eigenlayer enforce the slashing. Validators effectively set the withdrawal credentials of their staked ETH to point to Eigenlayer’s smart contracts, meaning that Eigenlayer processes withdrawals of their ETH and rewards. If a restaking validator acts maliciously, Eigenlayer slashes their ETH and prevents them from further participating in validation.</p><h4>Delegation</h4><p>There may be users who wish to restake through Eigenlayer but lack the technical capabilities to run a node, do not want the additional responsibilities, or do not trust themselves to validate without risking slashing. Eigenlayer offers a flexible solution, allowing them to delegate their ETH to independent operators on Eigenlayer, who then validate on their behalf. Slashing conditions still apply to these operators based on their performance, so users are encouraged to research and select trustworthy operators with a proven track record. In return for their services, users agree to forfeit a portion of their rewards to the operators. The agreements regarding compensation fees and the distribution of rewards are all automated through Eigenlayer’s smart contracts.</p><h4>Free Market</h4><p>The Eigenlayer platform fosters a dynamic free market for validation services on Ethereum by allowing validators the freedom to opt into or out of any module. This flexibility encourages modules to develop creative incentives and rewards to attract validators to secure their services and network. Validators can select which modules to secure based on the balance of rewards and risks (such as slashing conditions). Conversely, modules also have the option to select validators based on specific criteria. For instance, a module prioritizing maximum decentralization could stipulate conditions only to accept validation from a decentralized group of validators. This could involve accepting restake only from independent validators/stakers rather than from services like Lido. Over time, as more modules opt for decentralized validators, this trend could positively impact the Ethereum network as a whole by providing stakers with an additional incentive to stake with independent validators, promoting decentralization rather than only pooling their ETH in liquid staking projects. This market-driven approach efficiently allocates security resources, represented by stake, across the network. Even delegated operators can capitalize on this free-market environment by building credibility or developing innovative incentives to encourage users to delegate their ETH for restaking.</p><h4>Liquid Restaking Tokens (LRTs)</h4><p>Unlike staking pools like Rocket Pool and Lido, which issue liquid tokens in return for staking, Eigenlayer does not provide liquid tokens for restaking. Stakers simply reuse their already staked ETH to secure additional modules built on Ethereum. Eigenlayer recognizes that restakers face varying risks depending on the modules they opt into, making it challenging to represent these diverse restaking positions with a standard token. Nevertheless, emerging protocols such as <a href="https://www.rio.network/">Rio Network</a> address this by issuing novel <a href="https://rio.gitbook.io/rio-network-updates/updates/introducing-rio-network">Liquid Restaking Tokens</a> (LRTs) to reflect these varied restaking positions.</p><h3>Eigenlayer Addresses AVS Limitations</h3><p>Returning to the earlier-discussed limitations of AVSs, we can now understand how Eigenlayer addresses each issue.</p><ol><li><em>Pooled Trust</em>: Rather than onboarding validators from external networks and attempting to build new trust networks from the ground up, modules on Eigenlayer can tap into the security provided by the extensive set of validators already present on the Ethereum network. This approach significantly streamlines the trust-building process.</li><li><em>Value Retention</em>: With Eigenlayer, the Ethereum ecosystem doesn’t lose economic value to fees for external validation systems as it did previously. This is a major boost for retaining economic value in Ethereum. Additionally, Eigenlayer amplifies Ethereum’s <a href="https://www.investopedia.com/terms/n/network-effect.asp">network effects</a>, where increased participation improves the network’s overall value. As security is bolstered and services built on Ethereum are enhanced, Ethereum itself becomes more appealing for further development. This interconnectedness allows for more secure activities and deeper integration among various modules.</li><li><em>No Opportunity Cost</em>: When the same capital is utilized across multiple services, it is said to be <a href="https://www.investopedia.com/ask/answers/06/amortizationvsdepreciation.asp">amortized</a>. Instead of bearing the cost of each service individually, Eigenlayer enables validators to amortize their staked ETH across various services simultaneously, reducing the per-service restaking cost. Validators no longer face the dilemma of choosing between securing Ethereum for lower but safer returns or investing in a riskier AVS token for potentially higher gains. Eigenlayer allows them to pursue both options by amortizing their ETH without incurring additional costs for opting into additional modules on the platform.</li><li><em>Higher Trust</em>: Generally, the more capital staked on a network, the more secure it becomes. By reducing the opportunity cost associated with restaking, Eigenlayer facilitates easier access for modules to the substantial amounts of ETH staked on the main layer. Consequently, any potential attacker must control a significantly larger amount of ETH to launch a <a href="https://www.coindesk.com/learn/what-is-a-51-attack/">51% attack</a> (where a single person or group gains control of over 50% of the network) on a module.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DKGhGJtHXtA6QHt8FVUKMw.png" /></figure><h3>Risks and Considerations</h3><p>While Eigenlayer offers several beneficial features to the Ethereum ecosystem, it is essential to consider the risks and limitations as well.</p><h4>Operator Collusion</h4><p>Although Eigenlayer theoretically enables modules to be secured by a restake amount equal to the total ETH staked on Ethereum’s main layer, this scenario is likely unrealistic. In practice, only specific subgroups of validators might opt to secure a given module, raising the potential for collusion. Consider this scenario, adapted from the Eigenlayer <a href="https://docs.eigenlayer.xyz/overview/readme/whitepaper">whitepaper</a>.</p><p>Take an AVS on Ethereum, named Testcoin, with $1 million in locked value, secured by $6 million worth of restaked ETH. For an attacker to gain control over Testcoin’s network and access the $1 million in locked value, they would need control of over 50% of the network, or $3 million (half of the restaked ETH). At first glance, Testcoin appears to be crypto-economically secure because the cost of a failed attack (getting slashed at $3 million) outweighs the potential $1 million gain from a successful attack, deterring rational economic actors.</p><p>However, the situation changes if the same $6 million in restaked ETH also secures other modules. Suppose the same ETH is securing 10 modules (including Testcoin), each with $1 million in locked value. In this case, the cost to the attacker for corrupting these modules remains at $3 million, but the potential gain increases to $10 million (1$ million ×10 modules). This scenario renders the system crypto-economically insecure, as rational economic actors might find it economically viable to attempt an attack with potential gains ($10 million) exceeding the cost of failure ($3 million).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lsXXUlevxdvHZe0hvcDqeA.png" /></figure><p>Eigenlayer proposes multiple solutions to this dilemma. The first, though quite unrealistic, is the implementation of a <a href="https://www.investopedia.com/terms/h/hard-fork.asp">hard fork</a> in the Ethereum network. A hard fork represents a significant change where the network splits into multiple branches, with members choosing to continue on the version that aligns with their beliefs. In this scenario, nodes in the network would agree to disregard the branch of Ethereum where the attacker compromised Testcoin and instead build on a branch where the attack is negated. However, coordinating a hard fork, which requires most of the Ethereum network’s agreement, is challenging. It is unlikely that the network would initiate a hard fork for a few million dollars in lost value. However, if the attack causes a critical infrastructure to fail, then it is possible.</p><p>A more practical solution involves modules on Eigenlayer setting limits on the restake they accept from validators who are overly committed to restaking in other projects. These modules could specify terms in the Eigenlayer smart contracts to encourage participation from validators involved in a limited number of other modules. This aims to distribute restaking more evenly and reduce the risk of collusion or over-concentration of restaked ETH in a few hands.</p><h4>Unintended Slashing</h4><p>When dealing with large and mature protocols operating under the PoS consensus mechanism, the likelihood of a validator’s assets being mistakenly slashed due to an accident or a smart contract error is very low. However, this risk becomes more pronounced with newer protocols still in the early stages of development and not yet fully battle-tested. Such protocols risk unintentional slashing, where validators might incur losses despite acting honestly. Since Eigenlayer aims to extend the security of the Ethereum blockchain to AVSs and modules built on top of the main layer, the risk of unintentional slashing is heightened, as many of these modules do not possess the same level of maturity and size as Ethereum itself.</p><p>To mitigate this, Eigenlayer performs a protocol audit as a prerequisite for any module that utilizes Eigenlayer for restaking. This audit aims to identify any security vulnerabilities in the module’s smart contracts and operational mechanics.</p><p>A second solution proposed by Eigenlayer is the establishment of a governance layer. This layer allows community members to coordinate and vote on resolving issues, such as unintentional slashings. Equipped with a <a href="https://www.coindesk.com/learn/what-is-a-multisig-wallet/">multi-signature wallet</a> (crypto wallets that require the signatures of multiple users before transferring funds), this governance body has the power to veto slashing decisions. Modules can opt into having this veto committee oversee slashing dilemmas, which might be beneficial in the early stages of a protocol. This arrangement instills confidence in restakers, knowing that a veto committee can overturn mistaken slashes. As the protocol matures, they can withdraw from the veto committee’s oversight on slashing conflicts.</p><p>Eigenlayer primarily employs a reputation-based <a href="https://docs.eigenlayer.xyz/overview/readme/whitepaper">method</a> to build this veto committee, selecting esteemed individuals from the Eigenlayer and broader Ethereum community. It’s important to note that the committee has no authority to initiate slashing on its own. Rather, its role is limited to arbitrating in conflicting slashing incidents.</p><h4>Centralization Risks</h4><p>While a reputation-based veto committee can address issues like unintentional slashing, it also introduces centralization risks. Crypto has taught us that protocols relying on trustless verification rather than trust in good faith operate more securely. Although a committee comprising well-intentioned actors from the Ethereum and Eigenlayer communities sounds nice, it raises questions about defining “good faith,” determining who’s definition to go with, and whether these actors will consistently uphold good faith standards. Eigenlayer’s selection of committee members leads to a concentration of decision-making power. Furthermore, the potential for collusion exists, especially if committee members from similar communities have interacted prior and developed incentives to collude. In the short run, it might be acceptable to allow Eigenlayer to self-select this committee, but the protocol should consider alternative anti-centralization measures for governance in the long run.</p><p>There are additional centralization risks as well. For instance, some validators might opt out of restaking in certain modules if the rewards do not justify the operational costs. Other validators may lack the computational resources needed to participate effectively in specific modules, limiting their involvement in restaking. Consequently, more computationally capable validators might dominate restaking in various modules, while less advanced validators face limitations.</p><p>Eigenlayer <a href="https://docs.eigenlayer.xyz/overview/readme/whitepaper">contends</a> that many validation tasks can be performed using lightweight nodes, which require minimal computing infrastructure and are low in cost. For many modules, facilitating this approach can ensure restaked security while maintaining decentralization due to the low entry barrier for validators.</p><p>Additionally, Eigenlayer is implementing measures to ensure that workload distribution can be achieved using horizontal scalability with <a href="https://docs.eigenlayer.xyz/overview/readme/whitepaperhttps://docs.eigenlayer.xyz/overview/readme/whitepaper">Hyperscale AVSs</a> rather than solely relying on vertical scaling. Horizontal scaling involves distributing computations across various network nodes, each handling a portion of the workload. In contrast, vertical scaling enhances the capacity of a single node to perform multiple complex operations. Horizontal scaling promotes decentralization by not requiring individual nodes to meet high technical standards for complex tasks. Instead, nodes can manage parts of computations based on their capabilities. This prevents powerful centralized validators from amortizing their validation costs across too many modules. By aggregating the performance of many nodes, horizontal scaling maintains high system output.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Gh1toInc9flVHwvBtbdTrA.png" /></figure><h4>Value Accrual to Native Tokens</h4><p>When ETH is restaked to secure AVSs, it enhances the utility of ETH and preserves economic benefits, such as token value accrual. While this reaps benefits from an ETH perspective, it can lead to the native tokens of AVSs not capturing as much value as they would if validators directly staked the native tokens instead. Native AVS tokens might risk becoming underutilized in liquidity pools or DeFi more broadly. Considering these factors, I anticipate newer AVSs are more likely to join Eigenlayer’s platform, benefiting from the added security of restaking ETH. Conversely, more established and larger protocols might operate independently, securing their networks by native token staking, and thus enjoying the liquidity and value accrual benefits.</p><p>Eigenlayer is attempting to offer protocols the option of <a href="https://docs.eigenlayer.xyz/overview/readme/whitepaper">dual staking</a>. This model allows protocols to establish separate quorums consisting of both ETH restakers and stakers of their native tokens. Validators can participate in securing the network in whichever way they prefer, thus supporting both ETH and the native token. AVSs can be designed to consider responses from both quorums independently and require majority approval from each for decision validation. From a security standpoint, this approach enables modules to inject liquidity into their native tokens while hedging against their volatility by leveraging ETH’s stability.</p><h3>Applications using Eigenlayer</h3><p>Eigenlayer’s solution for pooled security on Ethereum introduces innovative methods for developing a wide range of Web3 applications.</p><h4>Event-Driven Activation</h4><p>In blockchain technology, <a href="https://docs.eigenlayer.xyz/overview/readme/whitepaper">event-driven activations</a> are actions automatically initiated by smart contracts in response to specific events. A common example is <a href="https://www.ledger.com/academy/glossary/liquidation">liquidation</a>, where collateral backing a loan is forcibly sold. This process is automatically triggered by events like the collateral’s value falling below a certain threshold. Presently, Ethereum does not natively support event-driven activations and requires third-party DApps or protocols for execution. These projects often rely on <a href="https://www.gemini.com/cryptopedia/keep3r-network-crypto-kp3r-token-ethereum">keeper networks</a>, nodes that monitor the blockchain and execute actions when predefined conditions are met, to implement event-driven activation. However, a drawback of this approach is that keeper network nodes are not the same nodes responsible for proposing or validating blocks. Therefore, there is a delay between when a keeper node is triggered and an action’s inclusion on the blockchain.</p><p>Eigenlayer offers a solution where block validators or proposers can opt into restaking their ETH for specific modules that perform event-driven activation. This allows nodes acting as keeper networks to also serve as validators/proposers, thereby eliminating the time delay in including event-driven activations on the Ethereum blockchain. Such a system could ensure immediate execution and recording of events like liquidations on the blockchain.</p><h4>Improved Finality Speeds</h4><p>Since Ethereum’s transition to PoS, the network now <a href="https://info.etherscan.com/epoch-in-ethereum/#:~:text=In%20Ethereum%2C%20an%20epoch%20is,6.4%20minutes%20in%20total%20time.">organizes time</a> into two units: slots and epochs, akin to units such as seconds and minutes in conventional timekeeping. A slot lasts approximately 12 seconds and is the period during which new blocks can be proposed for addition to the blockchain. An epoch comprises 32 slots, amounting to about 6.4 minutes. The first block in each epoch serves as a Checkpoint, marking the point at which validators agree on the blockchain’s state.</p><p>For a block to be permanently established on Ethereum, it must be Justified and Finalized. Justification occurs when at least two-thirds of the validators vote to include a specific block at a checkpoint. <a href="https://era.zksync.io/docs/reference/concepts/finality.html">Finality</a>, the stage where a block becomes irreversible, is achieved when the subsequent checkpoint in the next epoch is also justified. Therefore, it takes approximately 12.8 minutes (two epochs) for a block containing transactions to be finalized and permanently added to the blockchain. While this 12.8-minute delay enhances security by providing a buffer before permanently storing blocks, it can restrict the performance of certain Ethereum DApps, such as high-frequency trading platforms, which would benefit from immediate on-chain settlements and finalizations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*orhT9spaj-RglxzvjM5-0Q.png" /></figure><p>Leveraging the robust security of ETH staking, Eigenlayer enables protocols to use crypto-economic incentives to expedite finality on the Ethereum blockchain. For instance, <a href="https://ethereum.org/fil/roadmap/single-slot-finality/">Single Slot Finality</a> (SSF) aims to achieve finality for a transaction within the same slot in which the block is added to the blockchain. Nodes on Eigenlayer could achieve this by opting into signing off on a block’s finality after its addition and committing not to build on any chains that exclude that block.</p><p>SSF is riskier than standard finality, allowing validators less time to verify a block’s legitimacy. Therefore, only those willing to accept this higher risk can opt to restake ETH for SSF without imposing this standard on the entire Ethereum ecosystem. Implementing SSF across the Ethereum blockchain would require fundamental changes to its PoS consensus mechanism. Thus, Eigenlayer provides a platform to experiment with SSF, analyzing its impacts on Ethereum, while minimizing potential negative effects on the broader blockchain.</p><h4>Data Availability Layer</h4><p>In the blockchain world, a <a href="https://www.alchemy.com/overviews/data-availability-layer">Data Availability Layer</a> is a system responsible for ensuring that data (such as transaction information) can be readily accessible to all network participants. This availability is crucial for maintaining transparency, allowing any participant to verify the trustworthiness of information posted on the blockchain. However, as a blockchain expands, this becomes increasingly challenging due to the increasing volume of data and nodes that require access to this data.</p><p>Sharding has emerged as a valuable scalability technique to combat the growing difficulty of storing and accessing data on blockchains. Sharding involves dividing a blockchain into smaller, more manageable pieces (or shards) to distribute the computational and storage load. This method assigns different nodes to specific shards, enabling them to handle a portion of the blockchain (while maintaining communication with nodes in other shards), rather than requiring every node to maintain the entire history of the blockchain. On Ethereum, a particular form of sharding, known as <a href="https://ethereum.org/en/roadmap/danksharding/">Danksharding</a>, has been proposed as a network upgrade to enhance data availability. However, significant changes like this to the Ethereum blockchain often unfold slowly due to the decentralized nature of its governance, which comes at the expense of swift progress.</p><p>Eigenlayer offers the Ethereum community an opportunity to move beyond theoretical discussions and test concepts like Danksharding before full implementation on the main chain. This could provide practical insights into how Danksharding might function on-chain, potentially accelerating its deployment. In fact, the Eigenlayer team has already developed <a href="https://twitter.com/eigen_da">EigenDA</a>, the first implementation of a data availability layer on their platform, which we will explore further below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Z-13Apon-gHlFomIzU6kFw.png" /></figure><h3>EigenDA</h3><h4>Understanding Rollups</h4><p>Before delving deeper, let’s discuss Rollups, which are vital for understanding EigenDA. Rollups are Layer-2 solutions built on Ethereum, designed to address the network’s performance issues. They do this by offloading transaction processing from the main chain, thereby reducing the workload on Ethereum’s nodes. Though transactions are processed off-chain, Rollups still post their transaction data back to the main chain, allowing for finalization by the main Ethereum network. There are two primary types of Rollups on Ethereum: ZK Rollups and Optimistic Rollups.</p><p><a href="https://ethereum.org/en/zero-knowledge-proofs/">Zero-knowledge proofs</a> (ZKPs) are a cryptographic method of allowing one party (prover) to prove to another party (verifier) that a particular statement is true without revealing any information beyond the validity of the statement. <a href="https://chain.link/education-hub/zero-knowledge-rollup">ZK Rollups</a> use these ZKP to validate transactions processed off-chain. They batch transactions together, process them, and then generate a cryptographic proof verifying that the transactions were correctly processed. By batching transactions and processing them off the main congested chain, ZK Rollups enhance <a href="https://shardeum.org/blog/latency-throughput-blockchain/#:~:text=Throughput%20in%20blockchains%20typically%20means,transactions%20processed%20by%20the%20network">throughput</a>, the number of transactions a network can process within a given timeframe.</p><p><a href="https://ethereum.org/en/developers/docs/scaling/optimistic-rollups/">Optimistic Rollups</a> also process transactions off-chain and post their results back to the main Ethereum chain. They assume that transactions posted on-chain are valid unless proven otherwise. Unlike ZK Rollups, Optimistic Rollups do not require complex computations to generate ZKPs. If fraudulent behavior is suspected, users can submit a Fraud Proof, which challenges the validity of a particular transaction. This can only occur during a Challenge Period, an open time window after a roll up batch has been submitted. Once a fraud proof is filed, the Ethereum mainnet verifies it, and if foul play is confirmed, the transaction is revoked from the chain, with penalties applied to the fraudulent party. Due to this challenge window, Optimistic Rollups have a delay in transaction finality.</p><h4>What is EigenDA?</h4><p><a href="https://docs.eigenlayer.xyz/eigenda-guides/eigenda-overview">EigenDA</a> is a decentralized data availability layer built on Ethereum and is the first AVS to launch using Eigenlayer’s restaking mechanics. EigenDA capitalizes on the pooled network security form validators on Ethereum’s main layer. By integrating with EigenDA, rollups can achieve more efficient data handling, thereby reducing transaction costs and increasing throughput. Moreover, this scalability is enhanced as more users and operators opt into restaking to secure the EigenDA system. For ZK Rollups, lacking access to transaction data can impede nodes from reconstructing and verifying the rollup’s state. In the case of Optimistic Rollups, inaccessible transaction data prevents challengers from verifying commitments and generating fraud proofs, potentially allowing invalid transactions to go unchallenged. EigenDA ensures the timely availability of transaction data for nodes interacting with ZK Rollups and Optimistic Rollups.</p><h4>EigenDA’s Contributions</h4><p>EigenDA will provide several important <a href="https://www.blog.eigenlayer.xyz/intro-to-eigenda-hyperscale-data-availability-for-rollups/">contributions</a> to the Ethereum ecosystem. As it is built using core ideas and libraries intended for Danksharding, EigenDA enables Ethereum developers to experiment using validators participating in restaking for EigenDA. This will better prepare the Ethereum community for the integration of Danksharding into the main layer once the upgrade is underway. Additionally, EigenDA’s horizontal scaling architecture facilitates the participation of lightweight nodes by distributing the workload among them. This is a vital countermeasure against centralization, preventing powerful nodes from dominating network participation. As the first AVS on the platform, EigenDA is set to demonstrate the effectiveness of Eigenlayer’s restaking primitive and illustrate how AVS solutions can leverage the trust inherent in Ethereum’s main layer for their subnetworks.</p><p>Several protocols, including <a href="https://www.blog.eigenlayer.xyz/eigenda-layer-n/">Layer N</a> and <a href="https://www.blog.eigenlayer.xyz/eigenda-versatus/">Versatus</a>, have already started integrating successfully with EigenDA, and <a href="https://www.blog.eigenlayer.xyz/intro-to-eigenda-hyperscale-data-availability-for-rollups/">others</a> like <a href="https://celo.org/">Celo</a>, <a href="https://www.mantle.xyz/">Mantle</a>, and <a href="https://fluentlabs.xyz/">Fluent</a> are rumored to be considering similar integration.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AMBFdrdnup4tpVmyiwAchA.png" /></figure><h3>Looking Forward</h3><p>The rise and recent popularity of Eigenlayer signifies a shift in the future of network security for Ethereum-based protocols. Instead of laboriously bootstrapping network trust, many projects may integrate with Eigenlayer, thereby leveraging the established trust of the Ethereum network through restaking. However, numerous risks and questions remain to be addressed, such as mitigating centralization and ensuring protocols can still accrue value to their native tokens.</p><p>Another area worth exploring is how new protocols will develop Liquid Restaking Tokens (LRTs) to represent and quantify the diverse restaking positions users adopt on platforms like Eigenlayer. The decision whether to create their own restaking platforms or to integrate with Eigenlayer presents an interesting development trajectory. Protocols like <a href="https://www.rio.network/">Rio Network</a> and <a href="https://www.renzoprotocol.com/">Renzo Protocol</a> are already pioneering projects centered around LRTs. These projects and others to follow should be kept on short notice as the landscape for liquid restaking continues to evolve on the Ethereum blockchain.</p><h3>Citations</h3><p>Adnan, Faiz. “Epoch in Ethereum.” <em>Etherscan</em>, 2 May 2023, info.etherscan.com/epoch-in-ethereum/.</p><p>Banton, Caroline. “Network Effect: What It Is, How It Works, Pros and Cons.” <em>Investopedia</em>, 2 Aug. 2023, <a href="http://www.investopedia.com/terms/n/network-effect.asp.">www.investopedia.com/terms/n/network-effect.asp.</a></p><p>Betz, Brandy. “Staking Protocol Eigenlayer Raises $50m amid Crypto Winter.” <em>CoinDesk</em>, 28 Mar. 2023, <a href="http://www.coindesk.com/business/2023/03/28/staking-protocol-eigenlayer-raises-50m-amid-crypto-winter/">www.coindesk.com/business/2023/03/28/staking-protocol-eigenlayer-raises-50m-amid-crypto-winter/.</a></p><p>Brickwood, Dustin, et al. “Finality.” <em>Zksync</em>, 22 Aug. 2023, era.zksync.io/docs/reference/concepts/finality.html.</p><p>“Coinbase cbETH Whitepaper.” <em>Coinbase</em>, <a href="http://www.coinbase.com/cbeth/whitepaper.">www.coinbase.com/cbeth/whitepaper.</a> Accessed 27 Dec. 2023.</p><p>Conlon, Jemima. “What Is Sharding?” <em>Ledger</em>, 20 Nov. 2023, <a href="http://www.ledger.com/academy/what-is-sharding#:~:text=Put%20simply%2C%20sharding%20allows%20a,transactions%20from%20the%20original%20network.">www.ledger.com/academy/what-is-sharding#:~:text=Put%20simply%2C%20sharding%20allows%20a,transactions%20from%20the%20original%20network.</a></p><p>Cryptopedia Staff. “Keep3r Network’s KP3R Coin and Jobs Marketplace.” <em>Gemini</em>, 15 Oct. 2023, <a href="http://www.gemini.com/cryptopedia/keep3r-network-crypto-kp3r-token-ethereum.">www.gemini.com/cryptopedia/keep3r-network-crypto-kp3r-token-ethereum.</a></p><p>Cryptopedia Staff. “Oracles and Blockchain: Defi Oracles Examined.” <em>Gemini</em>, 5 Feb. 2022, <a href="http://www.gemini.com/cryptopedia/crypto-oracle-blockchain-overview.">www.gemini.com/cryptopedia/crypto-oracle-blockchain-overview.</a></p><p>“Danksharding.” <em>Ethereum</em>, 30 Nov. 2023, ethereum.org/en/roadmap/danksharding/.</p><p>Dusart, Victor. “Proof-of-Stake (POS).” <em>Ethereum</em>, 26 Sept. 2023, ethereum.org/en/developers/docs/consensus-mechanisms/pos/.</p><p>“Eigenda Overview.” <em>EigenLayer</em>, docs.eigenlayer.xyz/eigenda-guides/eigenda-overview. Accessed 2 Jan. 2024.</p><p>Eigenlayer Team. “Whitepaper.” <em>EigenLayer</em>, Apr. 2023, docs.eigenlayer.xyz/overview/readme/whitepaper.</p><p>Frankenfield, Jake. “Hard Fork: What It Is in Blockchain, How It Works, Why It Happens.” <em>Investopedia</em>, 25 May 2022, <a href="http://www.investopedia.com/terms/h/hard-fork.asp.">www.investopedia.com/terms/h/hard-fork.asp.</a></p><p>​​“Intro to Eigenda: Hyperscale Data Availability for Rollups.” <em>EigenLayer</em>, 19 Oct. 2023, <a href="http://www.blog.eigenlayer.xyz/intro-to-eigenda-hyperscale-data-availability-for-rollups/">www.blog.eigenlayer.xyz/intro-to-eigenda-hyperscale-data-availability-for-rollups/.</a></p><p>“Introducing Rio Network.” <em>Rio Network</em>, rio.gitbook.io/rio-network-updates/updates/introducing-rio-network. Accessed 29 Dec. 2023.</p><p>Jones, B.A., Goodkind, A.L. &amp; Berrens, R.P. Economic estimation of Bitcoin mining’s climate damages demonstrates closer resemblance to digital crude than digital gold. <em>Sci Rep</em> 12, 14512 (2022). <a href="https://doi.org/10.1038/s41598-022-18686-8">https://doi.org/10.1038/s41598-022-18686-8</a></p><p>“Key Terms: Eigenlayer Docs.” <em>EigenLayer</em>, docs.eigenlayer.xyz/overview/key-terms. Accessed 28 Dec. 2023.</p><p>Konstantopoulos, Georgios, and Hasu. “On Staking Pools and Staking Derivatives.” <em>Paradigm</em>, 23 Apr. 2021, <a href="http://www.paradigm.xyz/2021/04/on-staking-pools-and-staking-derivatives.">www.paradigm.xyz/2021/04/on-staking-pools-and-staking-derivatives.</a></p><p>“Layer N X Eigenda: A Case Study in Hyperscale DA for Finance.” <em>EigenLayer</em>, 4 Dec. 2023, <a href="http://www.blog.eigenlayer.xyz/eigenda-layer-n/">www.blog.eigenlayer.xyz/eigenda-layer-n/.</a></p><p>“Lido Whitepaper.” <em>Lido</em>, docs.lido.fi/. Accessed 27 Dec. 2023.</p><p>“Liquidation Meaning.” <em>Ledger</em>, 26 July 2023, <a href="http://www.ledger.com/academy/glossary/liquidation.">www.ledger.com/academy/glossary/liquidation.</a></p><p>Mcshane, Griffin. “What Is a 51% Attack?” <em>CoinDesk</em>, 11 May 2023, <a href="http://www.coindesk.com/learn/what-is-a-51-attack/">www.coindesk.com/learn/what-is-a-51-attack/.</a></p><p>Mcshane, Griffin. “What Is a Multisig Wallet?” <em>CoinDesk</em>, 14 Dec. 2022, <a href="http://www.coindesk.com/learn/what-is-a-multisig-wallet/">www.coindesk.com/learn/what-is-a-multisig-wallet/.</a></p><p>Nandini, Eshita. “EigenLayer and Securing New Middleware.” <em>Messari</em>, 28 Feb. 2023, messari.io/report/eigenlayer-and-securing-new-middleware.</p><p>Napoletano, Erika. “Proof of Work Explained.” <em>Forbes</em>, Forbes Magazine, 30 Oct. 2023, <a href="http://www.forbes.com/advisor/investing/cryptocurrency/proof-of-work/">www.forbes.com/advisor/investing/cryptocurrency/proof-of-work/.</a></p><p>“Optimistic Rollups.” <em>Ethereum</em>, 15 Oct. 2023, ethereum.org/en/developers/docs/scaling/optimistic-rollups/.</p><p>“Pooled Staking.” <em>Ethereum</em>, ethereum.org/en/staking/pools/. Accessed 27 Dec. 2023.</p><p>“Rocket Pool Whitepaper.” <em>Whitepaper</em>, whitepaper.io/document/551/rocket-pool-whitepaper. Accessed 27 Dec. 2023.</p><p>Ross, Sean. “Amortization vs. Depreciation: What’s the Difference?” <em>Investopedia</em>, 18 Dec. 2023, <a href="http://www.investopedia.com/ask/answers/06/amortizationvsdepreciation.asp.">www.investopedia.com/ask/answers/06/amortizationvsdepreciation.asp.</a></p><p>“Single Slot Finality.” <em>Ethereum</em>, 30 Nov. 2023, ethereum.org/fil/roadmap/single-slot-finality/.</p><p>“TVL Rankings.” <em>DefiLlama</em>, defillama.com/protocols. Accessed 27 Dec. 2023.</p><p>“Understanding Latency and Throughput in Embedded, Computer, and Blockchain Networks.” <em>Shardeum</em>, 5 May. 2023, shardeum.org/blog/latency-throughput-blockchain/#:~:text=Throughput%20in%20blockchains%20typically%20means,transactions%20processed%20by%20the%20network.</p><p>“Versatus X Eigenda: The First Stateless Rollup.” <em>EigenLayer</em>, 11 Dec. 2023, <a href="http://www.blog.eigenlayer.xyz/eigenda-versatus/">www.blog.eigenlayer.xyz/eigenda-versatus/.</a></p><p>Werkheiser, Brady. “What Is the Data Availability Layer?” <em>Alchemy</em>, 4 Aug. 2022, <a href="http://www.alchemy.com/overviews/data-availability-layer.">www.alchemy.com/overviews/data-availability-layer.</a></p><p>“What Is Layer 2?” <em>Chainlink</em>, 24 May 2023, chain.link/education-hub/what-is-layer-2.</p><p>“Zero-Knowledge Proofs.” <em>Ethereum</em>, 7 Sept. 2023, ethereum.org/en/zero-knowledge-proofs/.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fd37c01be44e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Zama: A New Standard for Confidentiality on the Blockchain]]></title>
            <link>https://medium.com/@mustafaxyz9/zama-a-new-standard-for-confidentiality-on-the-blockchain-c0a201b3e777?source=rss-5bd8d2734ef5------2</link>
            <guid isPermaLink="false">https://medium.com/p/c0a201b3e777</guid>
            <category><![CDATA[fhe]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[homomorphic-encryption]]></category>
            <category><![CDATA[zama]]></category>
            <category><![CDATA[blockchain]]></category>
            <dc:creator><![CDATA[Mustafa]]></dc:creator>
            <pubDate>Mon, 30 Oct 2023 01:57:14 GMT</pubDate>
            <atom:updated>2023-10-30T13:29:34.540Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Examining potential use cases for Fully Homomorphic Encryption in Web3 and a deeper dive into Zama’s four main open-source products: TFHE-rs, Concrete, Concrete ML, and fhEVM.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Bk0RX3-8Y41vH5kz-8E6bw.png" /></figure><p>Saying that we need better systems to protect our data privacy is nothing new; we’ve heard this story hundreds of times. It’s no secret that in today’s age, the issue of data privacy is as prevalent as it has ever been since the internet’s inception. The data breaches at Yahoo!, Equifax, and Marriott alone led to <a href="https://dataprot.net/articles/biggest-data-breaches/">nearly</a> $1 billion spent in legal settlement fees. According to a recent industry <a href="https://www.ibm.com/reports/data-breach">report</a> by IBM, the global average cost of a data breach in 2023 was USD 4.45 million, marking a 15% increase over three years.</p><p>Specifically, in Web3, technologies such as Zero-Knowledge Proofs (ZKPs) and Trusted Execution Environments (TEEs) have been used to protect user data. Recently, however, Fully Homomorphic Encryption (FHE) has emerged as a popular alternative (and complement) to many of these privacy-centric tools. FHE is a form of encryption that allows computations to be performed on data while it remains encrypted. Considering that infrastructure-based projects are <a href="https://messari.io/report/state-of-crypto-fundraising-q3-2023?referrer=research-reports">leading average monthly fundraising</a> in the blockchain industry in 2023, with $207 million, I anticipate that new FHE projects will fit well into this category and expand infrastructure fundraising leading into 2024.</p><h3>Zama</h3><p>Zama is a remote-friendly, Paris-based company pioneering open-source FHE tools for Web2 and Web3 applications. They have built several open-source products that make it easier for developers to leverage FHE for various use cases in areas such as blockchain and AI. Zama has completed three successful fundraising rounds, raising $50 million to date. Their most <a href="https://app.dealroom.co/companies/zama">recent</a> round was a $42 million Series A funding in February 2022, led by Protocol Labs.</p><p>To demonstrate the power of FHE firsthand, Zama has created some user-friendly FHE programs available on their Huggingface profile. The first program is a <a href="https://huggingface.co/spaces/zama-fhe/encrypted_sentiment_analysis">sentiment analysis tool</a>. It takes a short text as input, encrypts it using FHE, analyzes the encrypted message, and then informs you of the sentiment of your text — indicating how positive or negative the message is. Another program they’ve developed is an <a href="https://huggingface.co/spaces/zama-fhe/encrypted_image_filtering">image filtering tool</a>. It takes an image as input, encrypts it using FHE, applies a filter to the encrypted image, and then provides the filtered image as output. Their last program is a <a href="https://huggingface.co/spaces/zama-fhe/encrypted_health_prediction">health prediction tool</a>: users enter their symptoms, which are then encrypted using FHE. The program analyzes these encrypted symptoms and informs you of the most likely health issues.</p><p>Zama’s four main product offerings:</p><ol><li>TFHE-rs</li><li>Concrete</li><li>Concrete ML</li><li>fhEVM</li></ol><h3>TFHE-rs</h3><p>Zama’s <a href="https://docs.zama.ai/tfhe-rs">first open-source tool</a> is TFHE-rs, which stands for the jargony phrase “Fast Fully Homomorphic Encryption over the Torus using Rust”. Let’s break it down together.</p><p>Developers have various schemes to implement FHE, which can differ in aspects, such as the mathematical principles they’re based on and their performance speed. One particular version of FHE operates over the “torus” and is “fast”. A torus, in layman’s terms, is a donut. Formally, it’s the mathematical term for the geometric shape resembling a donut. TFHE-rs employs the mathematical structures and formulas associated with tori (plural for torus) for its encryption.</p><p>Interestingly, representing numbers cyclically using tori can enhance the performance speed of FHE schemes. This is why it’s termed “Fast Fully Homomorphic Encryption over the Torus.” TFHE-rs is crafted in the Rust programming language, indicated by the “-rs” suffix in its acronym. This is a typical extension for Rust libraries, akin to how “.jpeg” is used for image files, including your favorite memes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bP_0pDXxX7ujEwoVjS1N2g.png" /><figcaption><em>Figure 1</em></figcaption></figure><p>One of the advantages of Zama’s TFHE-rs library is its range of features. These include:</p><ul><li><em>Rust and C API</em>: The library provides an interface compatible with Rust and C programming languages. This makes FHE both more accessible and more versatile for developers.</li><li><em>Client-side WebAssembly API</em>: This feature allows developers to execute code for languages other than JavaScript directly in the browser, achieving speeds almost as fast as running them in native environments. It is a helpful API because it allows developers to utilize TFHE-rs for web applications.</li></ul><p>An essential feature of TFHE-rs is its resistance to quantum computers, which are orders of magnitude more powerful than conventional computers and can break specific encryption schemes. TFHE-rs can withstand quantum threats because its security is based on a cryptographic problem called Learning with Errors (LWE).”</p><p>The LWE problem refers to the mathematical challenge in which, given a set of linear equations, it is computationally difficult to solve these equations because they are designed to produce results containing errors. Thus, it is challenging to identify the actual solutions and underpins the security of the cryptography. A simple analogy is as follows: Suppose you tasted a spaghetti dish and attempted to deduce the ingredients and their quantities just by its taste. On top of that, imagine that every time you twisted the spaghetti onto your fork, I added a pinch of salt or sugar to your bite, intentionally misleading you. In this analogy, the equations are akin to the ingredients, the error to the unexpected salt or sugar, and solving the LWE problem to guessing the precise recipe.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*llgQ3HPaF5wkdI-5w7H16g.png" /><figcaption>Figure 2</figcaption></figure><p>For security, LWE necessitates the addition of random errors or noise to the message during encryption. However, as more calculations are performed, extraneous noise or errors accumulate with each computation. Some computational steps can produce more noise than others. For instance, the multiplication of ciphertexts (encrypted data) results in almost an exponential increase in noise with each computational step, whereas adding ciphertexts results in roughly a linear increase in noise. This noise level is manageable when only a couple of calculations are needed, but after multiple analyses, the noise can become so prominent that the original data becomes incomprehensible. The data is virtually lost.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*T4CG36qt5_yhZ2V5S60EdA.png" /><figcaption>Figure 3</figcaption></figure><p>To maintain an acceptable noise level for successful FHE implementation, TFHE-rs leverages a technique called ‘Programmable Bootstrapping’. This method mitigates the noise by spending computational power to refresh and reduce it. Picture the process like listening to music on an old radio that accumulates static over time, making the sound less clear. You can eliminate the static by quickly turning the radio off and back on. In this analogy, the encrypted data is likened to the music, errors/noise symbolize the static, and bootstrapping to restarting the radio.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Mq4LHIJopJ8h0tNA4yqy3g.png" /><figcaption>Figure 4</figcaption></figure><p>Programmable Bootstrapping has certain limitations:</p><ol><li>It requires significant computational power.</li><li>It reduces processing speed due to the time required for its implementation.</li><li>Its application must be tailored case-by-case, depending on the computations performed.</li></ol><p>Zama mitigates these limitations by employing ‘Parallelism’ to enhance the efficiency of Programmable Bootstrapping. Parallelism involves executing multiple calculations simultaneously, preventing the distortion of data that can occur with a sequential order of computations. Implementing this requires computers equipped with advanced hardware and software. Thankfully, companies like Microsoft and Intel have dedicated years to developing such capabilities as part of the DPRIVE program.</p><p>In practice, using TFHE-rs involves a user generating two keys: client and server keys. The client key is private and used by the user (you) to encrypt their data. The server key, being public, allows me to work on the encrypted data without revealing its contents. Once I’ve completed operations on your encrypted data, encrypted outcomes are produced, which only you can decrypt with your client key.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dNKalSc0Gna7nB41e2x2MQ.jpeg" /><figcaption>Figure 5</figcaption></figure><h3>Concrete</h3><p>Another <a href="https://docs.zama.ai/concrete">core tool</a> Zama developed is Concrete, an open-source framework that simplifies the challenging process of creating FHE programs for developers. Concrete includes a TFHE (Fast Fully Homomorphic Encryption over the Torus) compiler as part of this framework. This compiler translates regular programming code into actionable instructions that a computer can execute using FHE. It is excellent news for developers, as they can write code as they usually would and make only minor adjustments using Concrete to ensure their programs are fully homomorphically encrypted.</p><p>As previously discussed, performing calculations on FHE-encrypted data can introduce significant noise, leading to errors. Concrete’s default error probability is notably low, with only 1 in every 100,000 executions producing an erroneous output. Developers have the flexibility to modify this error probability’s parameters. However, reducing the error probability below this default setting demands more computational resources and significantly reduces performance speed. Conversely, increasing the error probability can enhance performance.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bJz9-2bPrAkQdKwochyIkw.png" /><figcaption>Figure 6</figcaption></figure><p>Developers building FHE programs have various options at their disposal to optimize calculations for greater efficiency:</p><ul><li><em>Reducing precision</em>: Involves rounding numbers in the program and the data.</li><li><em>Reducing Programmable Bootstrapping</em>: While Programmable Bootstrapping is valuable for enhancing accuracy, optimal efficiency can be achieved by employing it as infrequently as possible.</li><li><em>Increasing Parallelization</em>: This strategy involves breaking the program into smaller segments that can be processed concurrently rather than executing the entire program in one go.</li></ul><h3>Concrete ML</h3><p>Introducing Concrete ML, Zama’s third essential FHE <a href="https://docs.zama.ai/concrete-ml">tool</a>: This open-source machine learning inference framework is constructed on Concrete, merging AI and FHE. Concrete ML’s value lies in its ability to empower conventional data scientists and AI engineers to integrate FHE into their machine-learning models without prior knowledge of cryptography. These developers can continue to use familiar tools they would typically use, like scikit-learn and PyTorch, because they integrate with Zama’s open-source framework.</p><p>Here’s a high-level overview of using Concrete ML to convert a standard machine learning model into one that can make inferences on encrypted data:</p><ol><li><em>Train the Model on Unencrypted Data</em>: Train the model using unencrypted data just as you would with conventional machine learning models.</li><li><em>Quantization of the Model</em>: Through Concrete ML, implement a quantization scheme. This process involves converting decimals and floating-point values into integers, a crucial step given that Concrete ML currently works exclusively with integers.</li><li><em>Compile to FHE</em>: After quantization, Zama’s compiler converts the standard program code to one capable of interpreting encrypted data, ensuring FHE compatibility.</li></ol><p>While Zama has made remarkable strides in creating FHE tools tailored for machine learning, certain limitations remain. One prominent limitation that may already be on your mind is the necessity to train on unencrypted data. Currently, Concrete ML supports FHE solely for inference. Despite the numerous advancements in FHE, we’re still some distance away from being able to train machine-learning models on encrypted data.</p><p>Another limitation is the need for a quantization scheme. Concrete ML operates solely with 16-bit integers and not decimals. This poses a challenge since many models, particularly sophisticated deep learning models, often rely on decimal precision. These precise numbers must be quantized and converted into integers, potentially resulting in a loss of detail and worse model performance. Concrete ML comes equipped with built-in quantization algorithms, sparing developers the task of starting the quantization process from scratch. However, it also allows developers to create their own quantization algorithms, enabling them to tweak features and attempt to minimize detail loss and optimize performance.</p><p>Despite these limitations, it’s not all doom and gloom. Even in its current form, Concrete ML has showcased various exciting machine learning <a href="https://docs.zama.ai/concrete-ml/getting-started/showcase">demos</a>. These include encrypted image filtering, credit scoring, privacy-preserving text generation based on a user prompt (like GPT-2), and a program designed to gauge mood sentiment from short texts.</p><h3>fhEVM</h3><p>There is a prominent dilemma with blockchains. By storing data on-chain, blockchains foster trust through inherent transparency. To mitigate this, many protocols aiming for privacy store data off-chain. However, this approach eliminates contract composability, preventing certain smart contracts from interacting since they aren’t all on-chain. These considerations lead to the question: how can we store data on-chain while preserving confidentiality?</p><p>In response, Zama <a href="https://docs.zama.ai/fhevm">introduced</a> the fhEVM (Fully Homomorphic Encryption on the Ethereum Virtual Machine).</p><p>fhEVM is a protocol enabling developers to execute confidential smart contracts on-chain. This setup is ideal as it ensures that smart contracts remain composable, allowing interactions with one another while preserving the confidentiality of on-chain data. Additionally, fhEVM incorporates Zama’s TFHE-rs library (discussed earlier) and introduces a new TFHE Solidity library. This new library is helpful as it lets developers compute on encrypted data using Solidity, a widely recognized programming language on the Ethereum blockchain.</p><p>The high-level overview of fhEVM is that it uses a global public key, allowing all protocol users to encrypt data on-chain using FHE. The primary reason for this approach is that it enables multiple users, each with their data kept confidential, to interact with one another using smart contracts.</p><p>Since fhEVM uses asymmetric encryption, there are two keys:</p><ul><li><em>Public Key</em>: This key is stored publicly on-chain and is used by all users to encrypt their data and perform calculations on encrypted data.</li><li><em>Private Key</em>: This key possesses the ability to decrypt data. It is not owned by any single user and is instead distributed in ownership among validator nodes in the network. For data to be decrypted, several validators must cooperate and approve this action. This collaborative approval is termed a ‘Threshold Protocol’. This method enhances security through decentralization, preventing rogue misuse of the key.</li></ul><p>Although the key for encrypting data is public and can be used by anyone in the network, precautions are in place to prevent malicious behavior. To add encrypted data to the blockchain and compute on it, a user must submit two things:</p><ol><li>The encrypted data they wish to add to the blockchain.</li><li>A zero-knowledge proof of plaintext knowledge (ZKPoK). Through a ZKPoK, users demonstrate to the network that they know the contents (in unencrypted form) of the encrypted data they are introducing. A zero-knowledge proof is a cryptographic method that verifies knowledge of specific information without disclosing that information.</li></ol><p>Upon submitting these two items, the network verifies the encrypted data and its corresponding zero-knowledge proof. If the verification succeeds, the fhEVM stores the encrypted data on-chain and provides a handle (a reference to that data). If the verification fails, the encrypted data is not added to the blockchain.</p><p>Verification is crucial. It ensures that individuals attempting to add and engage with the encrypted data on the blockchain didn’t merely appropriate someone else’s encrypted data and present it as their own. Recall that the fhEVM stores encrypted data on-chain for public view. Thus, while anyone can attempt to interact with this encrypted data without a proper verification system, they won’t succeed in using it for unintended purposes.</p><p>Additionally, fhEVM enhances its security by using handles that point to the encrypted data. This lets fhEVM track the added encrypted data, especially when operations are executed that might produce new encrypted values. Naturally, when this occurs, the mapping between the handles and the data is updated. The security advantage of this approach is ensuring that only legitimately acquired data is used, as handles are added by the fhEVM only after the data passes the verification stage.</p><p>When it comes to decryption, there are two possible outcomes for the decrypted data:</p><ol><li><em>Publicly Displayed Decryption</em>: The standard approach when validators reach a Threshold Protocol is to decrypt data publicly and save the record on the blockchain. See Figure 7.</li><li><em>Private Decryption</em>: On the other hand, if a user wishes to maintain the confidentiality of their data and not display it publicly on the blockchain, they have that option. The validators in the network reach a Threshold Protocol and allow the user to obtain their data without revealing it on-chain. This is achieved by re-encrypting the original data into a new encrypted version using a different set of encryption keys owned by the user rather than the distributed network. The validators do this using the user’s public key, and then, this newly encrypted data is sent off-chain to the user, where they can decrypt it using their corresponding private key. Through this method, the user can access the raw contents without exposing them on-chain. See Figure 8.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Sryqw-pBf9DeMMQg-GrJ-w.jpeg" /><figcaption>Figure 7</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hNAQKe1CvGzHJssZNNPjPw.jpeg" /><figcaption>Figure 8</figcaption></figure><h3>Web3 Applications Using Zama’s Tools</h3><p>Based on Zama’s unique implementation of an EVM-compatible FHE protocol, they reveal several exciting applications in their <a href="https://github.com/zama-ai/fhevm/blob/main/fhevm-whitepaper.pdf">whitepaper</a>.</p><h4>Confidential Tokens</h4><p>ERC-20 tokens on the Ethereum blockchain are typically stored publicly, allowing anyone to access information such as token balances in wallets and details of previous transactions. From a confidentiality standpoint, this is not ideal, as it will enable people to infer characteristics about one’s real-world identity based on patterns in wallet activity or any KYC protocols to which one might have connected their wallet.</p><p>Using fhEVM, a user could create ERC-20 tokens with the functionality of encrypting balances so they don’t appear in wallets. To someone observing the details of your wallet history, the balances would appear as encrypted gibberish, yet the token would function as it usually would. To achieve this, a developer must substitute typical operations and integers with their fully homomorphically encrypted versions.</p><p>There are several interesting applications for confidential tokens. Companies that pay employees in crypto can use them to prevent coworkers from seeing each other’s salaries. In the DeFi world, new protocols can be designed to implement typical activities such as lending, borrowing, and yield farming but allow these operations to be carried out confidentially.</p><h4>Blind Auctions</h4><p>A blind auction is a type of auction where people submit bids privately, ensuring that neither other bidders nor the auctioneer knows the amount of any bid. Once all bids are collected, they are revealed, and the highest bid wins. Blind auctions are valuable tools for determining the actual value of auction items because people are encouraged to bid based on the value they personally assign to the item rather than being influenced by other bidders.</p><p>Implementing blind auctions on-chain has been challenging for many developers. The two main approaches have been using ZKPs and a two-step process, each with drawbacks.</p><ul><li><em>ZKP Approach</em>: Developers can use ZKP to keep bid data confidential, but this requires storing the bid data off-chain, introducing new confidentiality issues concerning where the bid data is stored.</li><li><em>Two-Step Approach</em>: In this method, participants send a cryptographic hash of their actual bid to the smart contract during the auction. Once bidding concludes, the auction enters a reveal phase where each bidder discloses their exact bid amount. To ensure participants don’t alter their bid amounts, the smart contract computes the hash of the revealed bid, which should match the previously submitted hash. The limitation of this approach is that malicious actors can deliberately send bids without claiming them during the reveal phase. This poses challenges for the auction, as it becomes impossible to ascertain whether the unclaimed bids were legitimate winning bids that were never claimed or if they were submitted by malicious actors intending to disrupt the auction.</li></ul><p>Using fhEVM, blind auctions can be <a href="https://www.zama.ai/post/on-chain-blind-auctions-using-homomorphic-encryption">designed</a> fairly using smart contracts, eliminating the need to rely on a trusted 3rd party to manage the auction. By using confidential tokens, individuals can send encrypted bids to a blind auction smart contract. The contract determines the winner by comparing the encrypted bids and only reveals the winning bid. Since the smart contract compares bids only in their encrypted forms, the privacy of non-winning bids remains intact.</p><h4>Confidential DAOs</h4><p>Decentralized Autonomous Organizations (DAOs) are organizations collectively governed by their constituents. DAOs use smart contracts to enable their members to make collective decisions about spending funds from their treasuries transparently and equitably. To make decisions and transfer funds, members typically submit proposals on which the constituents can vote to approve or disapprove. If a proposal garners sufficient support, it is approved, and the smart contracts automatically transfer the funds from the treasury.</p><p>DAOs primarily consist of two main smart contracts: the Governor contract and the Token contract. The Governor contract handles governance-related activities, such as creating and submitting proposals, voting on proposals, counting votes, and executing actions if a majority vote is achieved. The Token contract oversees token-related activities, like documenting token holders and managing operations such as issuing, burning, or transferring tokens between members.</p><p>Since DAOs primarily rely on token holdings to determine a member’s voting power, <a href="https://medium.com/my-first-dao/minimizing-biases-during-proposal-execution-f1fe76a5c69b">authority bias</a> often ensues, as people with larger amounts of tokens generally have more influence in votes. Sometimes, members follow along and vote in accordance with large holders rather than voting based on their true beliefs. This can lead to biased outcomes in voting rather than achieving true consensus.</p><p><a href="https://www.zama.ai/post/confidential-dao-voting-using-homomorphic-encryption">Using fhEVM</a>, DAOs can now set up governance systems with confidential voting, which can help eliminate biases from influencing voting decisions. These ‘Confidential DAOs’ can maintain the secrecy of components like token ownership and votes. The process of transforming a traditional DAO into a confidential one is straightforward. It involves replacing the conventional Governor and Token contracts with confidential counterparts that encrypt values using FHE.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fz3dWzMPy-jDGJGj0Rp08w.jpeg" /><figcaption>Figure 9</figcaption></figure><h4>Decentralized Identifiers (DIDs):</h4><p>Historically, personhood on the web has relied on centralized authorities to build systems that maintain identifiers like our email addresses or website domains. However, DIDs offer an alternative approach where individuals can credibly identify themselves on the web, backed by a decentralized network. Individuals can control their identifiers and data, doing things like updating an identifier without relying on an authority to do so.</p><p>In the past, it has been challenging to create secure DIDs due to the limitations of confidentiality on public blockchains. However, with Zama’s tools, people can store sensitive information in an encrypted and secure manner on the blockchain and take ownership of their online identifiers backed by a decentralized network.</p><p>For example, hospitals can encrypt sensitive data, such as a patient’s COVID-19 history, and store it in a smart contract on the blockchain. In the future, if an authorized party like a travel agency wants to verify information about whether a person had COVID-19 in the past without revealing other sensitive information about the patient, they can send a query to the smart contract. Using FHE, the contract can determine a (Yes/No) response without revealing other identifying information about the patient.</p><h3>Meet the Team</h3><p>The co-founders of Zama are Dr. Rand Hindi (CEO) and Dr. Pascal Paillier (CTO).</p><h4>Dr. Rand Hindi</h4><p><a href="https://www.linkedin.com/in/randhindi/">Rand</a> is an entrepreneur and investor in deep-tech sectors such as AI, Blockchain, and MedTech. At the age of 14, he founded Planet Ultra, a social network startup that connected young people in Paris. He also founded Snips, a privacy-enabled solution for companies to integrate voice assistants into their products. Rand holds a Bachelor of Science in Computer Science and a Ph.D. in Bioinformatics from University College London.</p><h4>Dr. Pascal Paillier</h4><p><a href="https://www.linkedin.com/in/ppaillier/">Pascal</a> is an entrepreneur and researcher with over 25 years of experience in cryptography. He has written dozens of research papers on topics including homomorphic encryption, embedded security, and anonymous credentials. Pascal spent ten years as the CEO of CryptoExperts, a Paris-based startup that advances the development of cryptography. He holds an engineering degree and a Ph.D. in Cryptography from Telecom ParisTech.</p><h3>Looking Forward</h3><p>Zama has done commendable work laying the groundwork and establishing open-source tools for integrating FHE into Web3 applications. We are already seeing other promising Web3 projects use these tools to build their products, with Fhenix <a href="https://medium.com/@Fhenix/fhenix-rising-the-first-fully-homomorphic-encryption-powered-blockchain-to-bring-end-to-end-4c3df5f616b0">recently</a> announcing itself as the first fully homomorphic encryption-powered blockchain. With Zama continuing to refine its tools, more companies will integrate these frameworks to build exciting new products that could significantly enhance privacy in the Web3 world. As these developments unfold, a new standard for online data privacy will be established.</p><h3>Citations</h3><p>“Cost of a Data Breach 2023.” <em>IBM</em>, 2023, <a href="http://www.ibm.com/reports/data-breach.">www.ibm.com/reports/data-breach.</a></p><p>Dahl, Morten, et al. “[Whitepaper] FHEVM: Confidential EVM Smart Contracts Using Fully Homomorphic Encryption.” <em>Github</em>, Zama, 25 Sept. 2023, community.zama.ai/t/whitepaper-fhevm-confidential-evm-smart-contracts-using-fully-homomorphic-encryption/771.</p><p>Danjou, Clément. “Confidential DAO Voting Using Homomorphic Encryption.” <em>Zama.Ai</em>, 12 Oct. 2023, <a href="http://www.zama.ai/post/confidential-dao-voting-using-homomorphic-encryption.">www.zama.ai/post/confidential-dao-voting-using-homomorphic-encryption.</a></p><p>Danjou, Clément. “On-Chain Blind Auctions Using Homomorphic Encryption and the FHEVM.” <em>Zama.Ai</em>, 10 July 2023, <a href="http://www.zama.ai/post/on-chain-blind-auctions-using-homomorphic-encryption.">www.zama.ai/post/on-chain-blind-auctions-using-homomorphic-encryption.</a></p><p>Devens, Chase. “State of Crypto Fundraising — Q3 2023.” <em>Messari</em>, 5 Oct. 2023, messari.io/report/state-of-crypto-fundraising-q3–2023?referrer=research-reports.</p><p>Fhenix. “FHENIX Rising: The First Fully Homomorphic Encryption-Powered Blockchain.” <em>Medium</em>, Medium, 26 Sept. 2023, medium.com/@Fhenix/fhenix-rising-the-first-fully-homomorphic-encryption-powered-blockchain-to-bring-end-to-end-4c3df5f616b0.</p><p>Rojas, Monica. “Minimizing Biases during Proposal Execution.” <em>Medium</em>, Medium, 24 July 2023, medium.com/my-first-dao/minimizing-biases-during-proposal-execution-f1fe76a5c69b.</p><p>Vojinovic, Ivana. “10 of the Biggest Data Breaches in History.” <em>DataProt</em>, 14 July 2023, dataprot.net/articles/biggest-data-breaches/.</p><p>“What Is Concrete ML?” <em>Zama.Ai</em>, docs.zama.ai/concrete-ml. Accessed 30 Oct. 2023.</p><p>“What Is Concrete?” <em>Zama.Ai</em>, docs.zama.ai/concrete. Accessed 30 Oct. 2023.</p><p>“What Is TFHE-Rs?” <em>Zama.Ai</em>, docs.zama.ai/tfhe-rs. Accessed 30 Oct. 2023.</p><p>“What Is Zama’s FHEVM?” <em>Zama.Ai</em>, docs.zama.ai/fhevm. Accessed 30 Oct. 2023.</p><p>“Zama Profile.” <em>App.Dealroom.Co</em>, app.dealroom.co/companies/zama. Accessed 30 Oct. 2023.</p><p>“Zama-FHE (Zama).” <em>Huggingface</em>, huggingface.co/zama-fhe. Accessed 30 Oct. 2023.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c0a201b3e777" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Explaining the Recent Rise of Fully Homomorphic Encryption in the Blockchain Industry]]></title>
            <link>https://medium.com/@mustafaxyz9/explaining-the-recent-rise-of-fully-homomorphic-encryption-in-the-blockchain-industry-c7081fa05458?source=rss-5bd8d2734ef5------2</link>
            <guid isPermaLink="false">https://medium.com/p/c7081fa05458</guid>
            <category><![CDATA[cryptography]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[fhe]]></category>
            <category><![CDATA[big-data]]></category>
            <category><![CDATA[homomorphic-encryption]]></category>
            <dc:creator><![CDATA[Mustafa]]></dc:creator>
            <pubDate>Tue, 10 Oct 2023 10:41:23 GMT</pubDate>
            <atom:updated>2023-10-29T15:27:25.112Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Introducing Fully Homomorphic Encryption (FHE): An exploration of its exciting applications, limitations, and recent developments fueling its popularity.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4K78w0yGSQ4ivemIkBjhRg.jpeg" /></figure><p>When I first heard about ‘Fully Homomorphic Encryption’ (FHE), I wondered about the blockchain space’s tendency to assign long names to trendy concepts. We’ve encountered a fair share of buzzwords that have swept the industry off its feet over the years, the most recent being ‘zero-knowledge proofs’ (ZKPs).</p><p>After doing some digging and exploring the new companies building products with FHE, I noticed a horizon filled with a brilliant new suite of tools. In the upcoming months and years, FHE could be the next big technology to take the industry by storm like ZKPs did. Companies are leveraging recent advancements in various areas of cryptography and cloud computing to pave the way toward a powerful, data-privacy-preserving future. The question is not if we get there but when, and I believe FHE could be a crucial catalyst for advancing data privacy and ownership.</p><p>Over the next few weeks, I will immerse myself in learning more about FHE and researching its limitations, potential, and applications. I’ll share my findings in a series of articles examining different facets of the dialogue surrounding FHE. This week, I’ll introduce this technology and discuss why it has recently garnered much attention. Many people in the industry are talking about it, including <a href="https://multicoin.capital/2023/09/26/the-dawn-of-on-chain-fhe/">Kyle Samani from Multicoin Capital</a>, who had this to say:</p><p>“FHE is the holy grail of cryptography. In time, FHE will reshape the fabric of all computing, both in web2 and web3.”</p><h3>What is Homomorphism?</h3><p>Addressing the elephant in the room, a wise start would be understanding what ‘homomorphic’ means. Tracing its roots, homomorphism originates from mathematics and is <a href="https://datascience.aero/fully-homomorphic-encryption-the-holy-grail-of-cryptography/">defined</a> as a map between two algebraic structures of the same type that preserves a core component between them.</p><p>If you’re like me and prefer a more practical definition, an underlying principle behind the math is that two groups need not be identical to possess the same core properties. For example, imagine two boxes of fruit, each corresponding to a distinct group:</p><p>Box A contains small-sized fruits.<br>Box B contains large-sized fruits.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0JSQstRktFzoua5pp4nrPw.png" /></figure><p>Although the individual fruits differ in size, juicing a small apple and orange together in Box A produces a mixed juice with the same flavor as juicing a large apple and orange together in Box B. Juicing the fruits to yield the same flavor is analogous to preserving a core component between both boxes. Assuming that identical flavor is our primary concern, it doesn’t matter which box we juice the fruits from because a larger/smaller quantity of juice is not our focus. The groups are equivalent where it matters (taste) so that the difference between them (size and quantity) has no impact on their primary function, which we defined as producing a specific fruit juice flavor.</p><p>Drawing a parallel with homomorphism, we have captured its two main two features:</p><ol><li><em>Mapping:</em> We establish a connection between the fruits wherein each small fruit in Box A corresponds to a larger version in Box B. So, the small apple in Box A corresponds to the large apple in Box B, and so forth.</li><li><em>Preservation of Operation:</em> If juicing two small fruits in Box A yields a particular flavor, juicing their corresponding larger versions in Box B should yield the same flavor. The ‘flavor profile’ is preserved despite differences in the size and quantity of the juice obtained.</li></ol><h3>What is Fully Homomorphic Encryption?</h3><p>Tying this back to the central topic of this article, <a href="https://inpher.io/technology/what-is-fully-homomorphic-encryption/">fully homomorphic encryption</a> (FHE) is a specific data encryption method that enables people to perform computations on encrypted data without revealing the raw data. Theoretically, the analysis and calculations performed on the encrypted data should produce results identical to those performed on the raw data. With FHE, we establish a 1:1 connection between data in the encrypted data set corresponding to data in the raw data set. The core component preservation, in this case, is the ability to perform any calculation on the data of either set and yield the same results.</p><p>For context, many companies already take preventative measures to protect user data and maintain differential privacy. Companies rarely store data on the cloud or in their databases in raw, unencrypted form. Therefore, even if attackers gain control of a company’s servers, they still have to bypass the encryption to read and access the data. However, data is not interesting when it’s just sitting there, encrypted and unused. When companies want to perform analyses on data to derive valuable insights, they have no great option but to decrypt the data to do this. When decrypted, the data becomes vulnerable. However, through end-to-end encryption, FHE becomes very useful as we no longer have to decrypt data to analyze it; this is just scratching the surface of what is possible.</p><p>A crucial consideration is whether companies should be allowed to read and store our personal information to begin with. The standard response to this by many has been that companies need to see our data to provide us with better services.</p><p>If YouTube doesn’t store data like my watch and search history, the algorithm can’t operate to its fullest potential and show me the videos I’m interested in. For this reason, many people have considered the tradeoff between data privacy and getting better services worth making. However, with FHE, we no longer need to make this tradeoff. Companies like YouTube can train their algorithms on encrypted data and produce identical results for the end user without infringing on data privacy. Specifically, they can homomorphically encrypt information such as my watch and search history, analyze it without looking at it since it is encrypted, and then show me the videos I’m interested in based on the analysis.</p><p>FHE is a significant step towards building a future where our data is no longer a valuable commodity that we make the compromise on giving away freely to organizations.</p><h3>Applications of Fully Homomorphic Encryption</h3><p>Fully Homomorphic Encryption (FHE) applied correctly is a breakthrough for all sectors that store user data. We are looking at a technology that could transform our entire attitude towards data privacy and the limits of encroachment acceptable by corporations.</p><p>Let’s start by examining how FHE could remodel data practices in the <a href="https://inpher.io/technology/what-is-fully-homomorphic-encryption/">healthcare industry</a>. Many hospitals hold private records of patients stored in their databases, which they must keep confidential for ethical and legal reasons. However, this information is valuable for external medical researchers who can analyze this data to deduce significant insights into illnesses and potential cures. A major obstacle that is slowing down research is maintaining the total confidentiality of patient data when outsourcing it to researchers. There are many ways to anonymize or pseudonymize patient records. Still, they are imperfect and can either reveal too much about someone, making them identifiable, or not reveal enough about their cases, making it difficult to obtain accurate insights about illnesses.</p><p>With FHE, hospitals can encrypt patient data homomorphically, making the protection of patient privacy in the cloud easier. Medical researchers can perform calculations and run analytical functions on the encrypted data without compromising patients’ privacy. Since there is a 1:1 mapping between the encrypted and raw data, the results obtained from the encrypted data set provide real insights that can be applied to actual cases. FHE could rapidly accelerate advancements in the healthcare industry.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Bh7nou2PoRP2BXvFt_ndhg.jpeg" /></figure><p>Another exciting application for FHE is Artificial Intelligence (AI) training. Currently, the AI sector grapples with privacy concerns, which hinder companies from accessing many expansive datasets essential for refining AI algorithms. Companies training AI must choose between using limited public data sets, paying a lot of money to buy private data sets, or creating datasets, which is challenging for smaller companies with fewer users. FHE should address the privacy concerns preventing many data set providers from entering this market. Therefore, improvements in FHE will likely lead to an increase in the number of data sets available to train AI. This will make AI training more financially accessible and refined, given the increased diversification of available data sets.</p><h3>Past Limitations in Fully Homomorphic Encryption</h3><p>If Fully Homomorphic Encryption (FHE) indeed promises to transform modern-day Big Data, why haven’t we seen more of it in action yet?</p><p>While FHE has been a topic people have discussed and researched for years, the reality is that implementing FHE in practice is very difficult. The core challenge lies in the computational power needed to execute FHE. A fully homomorphically secure dataset can produce identical analytical results as its raw data form. This is a challenging feat and demands heavy computation speeds and capabilities, many of which are impractical to implement on existing computers. An operation that typically takes seconds on raw data could take hours or days on homomorphically encrypted data sets. This computational challenge has created a self-perpetuating cycle where many engineers are postponing the undertaking of FHE projects, consequently slowing down its development and constraining the full realization of its benefits.</p><p>A specific example of computational problem engineers face with FHE is addressing the ‘<a href="https://www.darpa.mil/news-events/2020-03-02">noise error’</a>. When performing calculations on homomorphically encrypted data sets, many engineers have faced instances where extraneous noise or errors are produced each time a calculation is made. This is tolerable when only a couple of calculations are needed, but after multiple analyses, the noise can become so prominent that the original data becomes incomprehensible. The data is virtually lost.</p><h3>Why Now?</h3><p>Like <a href="https://medium.com/@matthewdwhite/a-brief-history-of-generative-ai-cb1837e67106">generative AI</a>, once deemed limited and primitive before becoming mainstream, Fully Homomorphic Encryption (FHE) is on a trajectory toward similar progress. Many industry leaders, even those extending beyond the blockchain space, have gathered to organize heavy research and development into FHE. This has led to several recent industry developments, driving a compelling narrative for the advancement of this technology.</p><h4>DPRIVE Program</h4><p>In March 2021, Microsoft, Intel, and the Defense Advanced Research Projects Agency (DARPA) agreed to launch a <a href="https://www.intel.com/content/www/us/en/newsroom/news/intel-collaborate-microsoft-darpa-program.html#gs.640smc">multi-year program</a> to accelerate the development of Fully Homomorphic Encryption (FHE). Named Data Protection in Virtual Environments (DPRIVE), this program signified a substantial progression for FHE. It showcased two industry giants specializing in cloud computing and computer hardware, uniting to address data privacy. They initiated this program to build computers and software capable of managing FHE computing speeds and to establish guidelines for accurate FHE implementation, safeguarding against the data breaches that may arise from incorrect usage.</p><p>As part of the DPRIVE program, engineers have undertaken the task of mitigating the previously mentioned ‘noise error’ by exploring methods to reduce noise to levels that preserve raw data. A promising solution has been designing <a href="https://www.darpa.mil/news-events/2020-03-02">Large Arithmetic Word Size</a> (LAWS) data representations. While conventional computer processors (CPUs) typically use 64-bit words, engineers are developing novel hardware capable of processing words of 1024 bits or more with LAWS. This approach has been effective because research indicates that longer words directly influence the signal-to-noise ratio. Putting it simply, longer words generate less noise with each additional computational step in FHE, allowing for more calculations to be executed before reaching the data loss threshold. By building new hardware to tackle these challenges, engineers involved in the DPRIVE program substantially reduce the computational load required to execute FHE.</p><p>To speed up calculations and get closer to their goal of making FHE 100,000 times faster, the DPRIVE team embarked on an ongoing journey to design new data processing systems that surpass the capabilities of conventional processing and graphing units. They developed a new <a href="https://community.intel.com/t5/Blogs/Products-and-Solutions/HPC/Intel-Completes-DARPA-DPRIVE-Phase-One-Milestone-for-a-Fully/post/1411021">Multiple Instruction Multiple Data</a> (MIMD) system capable of simultaneously managing numerous instructions and data sets. MIMD is analogous to constructing a new superhighway instead of using existing ill-equipped roads to accommodate the traffic required for speedy, real-time calculations of FHE.</p><p>What’s interesting about the DPRIVE program is its extensive use of ‘<a href="https://community.intel.com/t5/Blogs/Products-and-Solutions/HPC/Intel-Completes-DARPA-DPRIVE-Phase-One-Milestone-for-a-Fully/post/1411021">parallelism</a>’ in computer mathematical calculations. This enables developers to perform multiple large-number calculations simultaneously. You can think of parallelism as deploying a bunch of mathematicians to work on different parts of a giant math problem all at once instead of having each of them do their jobs one after the other. Although performing multiple calculations simultaneously facilitates swift problem-solving, the computers must be air-cooled to prevent overheating.</p><p>In September 2022, more than a year and a half after launching the program, Microsoft, Intel, and DARPA <a href="https://community.intel.com/t5/Blogs/Products-and-Solutions/HPC/Intel-Completes-DARPA-DPRIVE-Phase-One-Milestone-for-a-Fully/post/1411021">announced</a> they had successfully completed Phase 1 of the DPRIVE program. They are currently in the process of undertaking Phase 2 of DPRIVE.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_gUmZkht0b0a0TpsWpmetA.jpeg" /></figure><h4>SDKs and Open-Source Libraries</h4><p>With numerous major corporations pioneering advancements in Fully Homomorphic Encryption (FHE), there has been a surge in the availability of Software Development Kits (SDKs) and open-source libraries that allow developers to build upon each others’ work.</p><p>Microsoft <a href="https://www.microsoft.com/en-us/research/project/microsoft-seal/">announced</a> the launch of Microsoft Seal, an open-source library that provides developers with tools to execute homomorphic encryption on data sets. This democratizes access to end-to-end encryption and computation services, enabling a broader range of developers to explore FHE. The library provides examples of homomorphically encrypted programs, complete with detailed commentary, to guide developers toward correct and secure usage.</p><p>Intel also <a href="https://www.intel.com/content/www/us/en/developer/tools/homomorphic-encryption/overview.html#gs.fu55im">launched</a> its own Homomorphic Encryption Toolkit, offering developers tools to facilitate faster homomorphic encryption in the cloud. Intel designed this toolkit flexibly, ensuring compatibility with the latest data processing and computing advancements. It includes specialized functions tailored for lattice cryptography, integrations for seamless operation with Microsoft Seal, samples of homomorphically encrypted schemes, and technical documents to guide users.</p><p>Google’s <a href="https://security.googleblog.com/2019/06/helping-organizations-do-more-without-collecting-more-data.html">Private Join and Compute</a> open-source library provides developers with multi-party computation (MPC) tools. This computational method allows parties to gain shared insights by combining their different data sets without exposing their raw data to each other. Private Join and Compute merges the cryptographic techniques from FHE with Private Set Intersection (PSI) to optimize data confidentiality practices. PSI, another cryptographic method, allows parties with distinct datasets to identify common elements or data points without revealing their data. Google’s approach towards advancing data privacy doesn’t solely concentrate on FHE; it prioritizes the broader concept of MPC by integrating FHE with other impactful data practices.</p><p>The rising availability of reputable open-source libraries for FHE is noteworthy. However, it becomes even more compelling when observing reputable companies experimenting with these libraries in their operations. In April 2021, Nasdaq, a renowned stock exchange and global technology entity for capital markets, <a href="https://www.intel.com/content/www/us/en/newsroom/news/xeon-advances-nasdaqs-homomorphic-encryption-rd.html#gs.64abim">incorporated</a> FHE into its operations. Utilizing Intel’s FHE tools and high-speed processor, Nasdaq tackled financial crime through anti-money laundering efforts and fraud detection. This was achieved by using homomorphic encryption to identify valuable insights and potentially illicit activities in data sets containing sensitive information.</p><h4>Recent Capital Raises</h4><p>In addition to the research and development conducted by previously mentioned firms, several other companies have recently secured substantial funding rounds for initiatives focused on Fully Homomorphic Encryption (FHE)</p><p><a href="https://cornami.com/">Cornami</a>, a large technology company, is celebrated for pioneering the development of scalable cloud computing technology designed specifically for homomorphic encryption. They are engaged in numerous endeavors to create computing systems that support FHE more effectively than conventional CPUs. They also direct initiatives aimed at safeguarding encrypted data against threats posed by quantum computing. In May 2022, Cornami <a href="https://www.securityweek.com/cornami-raises-68-million-quantum-secure-computing-encrypted-data/">announced</a> a successful Series C funding round, securing $68 million led by Softbank and elevating its total capital raised to $150 million.</p><p><a href="https://www.zama.ai/">Zama</a> is another company in the blockchain industry that is building open-source homomorphic encryption tools that developers can leverage to build exciting applications using FHE, blockchain, and AI. Zama has built a Fully Homomorphic Ethereum Virtual Machine (fhEVM) as part of its product offerings. This smart contract protocol enables on-chain transaction data to remain encrypted during processing. Developers exploring various applications with Zama’s library have been impressed with the performance, even in complex use cases. Zama <a href="https://app.dealroom.co/companies/zama">successfully closed</a> its $42 million Series A funding round in February 2022, led by Protocol Labs, elevating its total capital raised to $50 million.</p><p><a href="https://www.fhenix.io/">Fhenix</a> is also an emerging project that is bringing FHE to the blockchain. Their goal is to extend FHE applications beyond confidential payments, opening up doors for exciting <a href="https://multicoin.capital/2023/09/26/the-dawn-of-on-chain-fhe/">use cases</a> of FHE in areas such as decentralized finance (DeFi), bridging, governance voting, and Web3 gaming. In September 2023, Fhenix <a href="https://www.theblock.co/post/252931/fhenix-seed-multicoin-capital">announced</a> the closing of its $7 million Seed funding round, led by Multicoin Capital and Collider Ventures.</p><h3>What’s Next?</h3><p>For years, Fully Homomorphic Encryption (FHE) lingered as an idea that promised robust end-to-end encryption, heralding a future of strong data privacy. Recent developments are beginning to shift FHE from a theoretical dream to a practical reality. Although various companies are competing to pioneer implementing the first robust, fully functional version of FHE, many are collaborating to navigate the complexities of this formidable technology jointly. This collaborative spirit is evident through their implementation of various cross-team programs and the development of open-source libraries that integrate with other libraries.</p><p>Based on my findings, the discussion surrounding FHE seems far-reaching. Over the next few weeks, I am excited to delve deeper, sharing further insights from my research on FHE. Specifically, I am eager to explore more about topics such as:</p><ul><li>Emerging applications of FHE.</li><li>The interplay between Zero-Knowledge Proofs (ZKPs) and FHE.</li><li>Integrating FHE with Private Set Intersection (PSI) to advance secure Multi-Party Computation (MPC).</li><li>New companies, like Zama and Fhenix, that are pioneering developments in FHE.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jMVheU6PbTvhqWul6QPa-A.png" /></figure><h3>Citations:</h3><p>Arampatzis, Anastasios. “Latest Developments in Homomorphic Encryption.” <em>Venafi</em>, 1 Feb. 2022, venafi.com/blog/what-are-latest-developments-homomorphic-encryption-ask-experts/.</p><p>Arampatzis, Anastasios. “What Is Homomorphic Encryption &amp; How Is It Used.” <em>Venafi</em>, 28 Apr. 2023, venafi.com/blog/homomorphic-encryption-what-it-and-how-it-used/.</p><p>“Building Hardware to Enable Continuous Data Protections.” <em>DARPA</em>, 2 Mar. 2020, <a href="http://www.darpa.mil/news-events/2020-03-02.">www.darpa.mil/news-events/2020-03-02.</a></p><p>Cristobal, Samuel. “Fully Homomorphic Encryption: The Holy Grail of Cryptography.” <em>Datascience.Aero</em>, 7 Jan. 2021, datascience.aero/fully-homomorphic-encryption-the-holy-grail-of-cryptography/.</p><p>“Homomorphic Encryption: What Is It, and Why Does It Matter?” <em>Internet Society</em>, 9 Mar. 2023, <a href="http://www.internetsociety.org/resources/doc/2023/homomorphic-encryption/">www.internetsociety.org/resources/doc/2023/homomorphic-encryption/.</a></p><p>Hunt, James. “FHENIX Raises $7 Million in Seed Round Led by Multicoin Capital.” <em>The Block</em>, The Block, 26 Sept. 2023, <a href="http://www.theblock.co/post/252931/fhenix-seed-multicoin-capital.">www.theblock.co/post/252931/fhenix-seed-multicoin-capital.</a></p><p>“Intel® Homomorphic Encryption Toolkit.” <em>Intel</em>, <a href="http://www.intel.com/content/www/us/en/developer/tools/homomorphic-encryption/overview.html#gs.fu55im.">www.intel.com/content/www/us/en/developer/tools/homomorphic-encryption/overview.html#gs.fu55im.</a> Accessed 8 Oct. 2023.</p><p>“Intel to Collaborate With Microsoft on DARPA PROGRAM.” <em>Intel</em>, 8 Mar. 2021, <a href="http://www.intel.com/content/www/us/en/newsroom/news/intel-collaborate-microsoft-darpa-program.html#gs.ftusxq.">www.intel.com/content/www/us/en/newsroom/news/intel-collaborate-microsoft-darpa-program.html#gs.ftusxq.</a></p><p>“Intel Xeon Advances NASDAQ’s Homomorphic Encryption R&amp;D.” <em>Intel</em>, 6 Apr. 2021, <a href="http://www.intel.com/content/www/us/en/newsroom/news/xeon-advances-nasdaqs-homomorphic-encryption-rd.html#gs.6mpgme.">www.intel.com/content/www/us/en/newsroom/news/xeon-advances-nasdaqs-homomorphic-encryption-rd.html#gs.6mpgme.</a></p><p>Johnson, Rick. “Intel Completes DARPA DPRIVE Phase One Milestone for a Fully Homomorphic Encryption Platform.” <em>Intel</em>, 14 Sept. 2022, community.intel.com/t5/Blogs/Products-and-Solutions/HPC/Intel-Completes-DARPA-DPRIVE-Phase-One-Milestone-for-a-Fully/post/1411021.</p><p>“Microsoft Seal: Fast and Easy-to-Use Homomorphic Encryption Library.” <em>Microsoft Research</em>, 4 Jan. 2023, <a href="http://www.microsoft.com/en-us/research/project/microsoft-seal/">www.microsoft.com/en-us/research/project/microsoft-seal/.</a></p><p>Paillier, Dr. Pascal. “Fully Homomorphic Encryption: The Holy Grail of Cryptography.” <em>Business Age</em>, 9 Mar. 2023, <a href="http://www.businessage.com/post/fully-homomorphic-encryption-the-holy-grail-of-cryptography">www.businessage.com/post/fully-homomorphic-encryption-the-holy-grail-of-cryptography</a></p><p>Samani, Kyle. “The Dawn of On-Chain FHE.” <em>Multicoin Capital</em>, 26 Sept. 2023, multicoin.capital/2023/09/26/the-dawn-of-on-chain-fhe/.</p><p>Walker, Amanda, et al. “Helping Organizations Do More Without Collecting More Data.” <em>Google Online Security Blog</em>, 19 June 2019, security.googleblog.com/2019/06/helping-organizations-do-more-without-collecting-more-data.html.</p><p>“What Is Fully Homomorphic Encryption?” <em>Inpher</em>, 11 Apr. 2021, inpher.io/technology/what-is-fully-homomorphic-encryption/.</p><p>White, Matt. “A Brief History of Generative AI.” <em>Medium</em>, 8 July 2023, matthewdwhite.medium.com/a-brief-history-of-generative-ai-cb1837e67106#:~:text=Although%20most%20people%20will%20admit,of%20Stability%20AI’s%20Stable%20Diffusion.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c7081fa05458" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Interview with Rong Kai Wong, CEO of PARSIQ]]></title>
            <link>https://medium.com/@mustafaxyz9/interview-with-rong-kai-wong-ceo-of-parsiq-39d9c0328b96?source=rss-5bd8d2734ef5------2</link>
            <guid isPermaLink="false">https://medium.com/p/39d9c0328b96</guid>
            <category><![CDATA[ceo]]></category>
            <category><![CDATA[blockchain-data-analytics]]></category>
            <category><![CDATA[rong-kai-wong]]></category>
            <category><![CDATA[parsiq]]></category>
            <category><![CDATA[parsiq-network]]></category>
            <dc:creator><![CDATA[Mustafa]]></dc:creator>
            <pubDate>Wed, 04 Oct 2023 13:57:25 GMT</pubDate>
            <atom:updated>2023-10-04T13:58:27.860Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Learn the story about how a police commander went on to become the CEO of PARSIQ, a leading blockchain data company.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*z7iXFMRtQOZ8kwVSeoxk5w.png" /></figure><p>This week, we are excited to share an exclusive interview with Rong Kai Wong, the CEO of <a href="https://www.parsiq.net/">PARSIQ</a>. <a href="https://www.parsiq.net/">PARSIQ</a> is a blockchain data company that provides Web3 and AI projects with real-time data and infrastructure tools to help speed up their development. <a href="https://www.parsiq.net/">PARSIQ</a> creates custom and robust APIs tailored specifically to the different needs of each project</p><p>Beyond his work as the CEO of <a href="https://www.parsiq.net/">PARSIQ</a>, Rong Kai Wong has obtained a diverse range of experiences across numerous industries. He spent eight years working with the Singapore Police Force as a Commanding Officer and a Head of Operations. His first foray into Web3 was with <a href="https://www.binance.com/en">Binance</a>, where he managed their $100 million BSE accelerator fund, investing in 30+ projects, including <a href="https://www.parsiq.net/">PARSIQ</a>. He is a natural leader who has managed teams as large as 1,000 people.</p><p>In the interview, Rong Kai Wong tells us about his unique story and provides valuable life and professional advice for students and young professionals. His candid and thoughtful insights are timeless.</p><blockquote><strong>Watch the Video Interview</strong></blockquote><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F84qdLiFyxWM%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D84qdLiFyxWM&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F84qdLiFyxWM%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/c29239169b99e920c62cddf51b4b6f29/href">https://medium.com/media/c29239169b99e920c62cddf51b4b6f29/href</a></iframe><blockquote><strong>Read the Transcript</strong></blockquote><h3><strong>Tell us about your story and how you got into the Web3 space.</strong></h3><p>“My story is probably quite hard to replicate in this space. I didn’t start off in the Web3 space after graduation. I graduated from <a href="https://www.imperial.ac.uk/">Imperial College London</a> in 2012 with a Masters of Chemical Engineering. I never did engineering after that because I actually went back to Singapore. I’m a Singaporean. I became a senior police officer with the Singapore police force. I had a scholarship from there and I worked with them for 8 years. I enjoyed my time over there. It gave me a lot of training in terms of critical thinking, stress handling, and man management. I was a superintendent of police, so I managed 1000 officers. And that was a really good experience that I had. I left public service for various reasons. There’s a pool of factors, but at the end of the day, it really boiled down to what I wanted for my life.</p><p>That’s something that I hope the listeners — if you are young, fresh out of university, or still in university — start to think about. And you also need to know that your life priorities change over time. When you’re 23, you’re young, you have a lot of energy, you can go three days without sleeping, you can tackle anything in the world, and you want to go all out for a career — that’s great. Go ahead and do it if that’s what you want. However, you also need to know that’s not going to last. You need to know that down the road, when you’re 30, when you’re 33, or maybe you’ve found someone and you’ve settled down and you have kids, then your priorities change with life and you need to be able to adapt.</p><p>So effectively, what I did was actually live a public service life for my family. I joined <a href="https://www.binance.com/en">Binance</a> after that. That’s how I got into the Web3 space in 2020. I did investments with them for more than a year, as well as ecosystem growth. I was actually managing the 100 million BSE accelerator fund, of course with my team. Out of the 500+ companies and projects that we looked into, we only invested in 31 of them by the time I left. So that’s how strict we were. But of the 31 companies that we invested in, <a href="https://www.parsiq.net/">PARSIQ</a> was one of them. I was very impressed with <a href="https://www.parsiq.net/">PARSIQ</a>. I liked the team, I liked the technology, and I liked how well their growth was. When I left <a href="https://www.binance.com/en">Binance</a>, <a href="https://www.parsiq.net/">PARSIQ</a> actually hired me, initially as COO, and now as CEO.</p><p>My path is not traditional, and it’s not typical. <a href="https://www.parsiq.net/">PARSIQ</a> hired me back then as COO because of my experience with the police force which was very operational. I’ve experienced growing teams from less than 10 all the way to 1000. So anything more than 1000 is new to me. But in Web3, this is usually good enough. They hired me as a COO, and now they made me a CEO. So that’s my story.”</p><h3><strong>What are the similarities and differences between leading a police force division and a Web3 company?</strong></h3><p>“In public service when I was in the Singapore police force, it was very structured in a big company. And actually you tend to realize that whenever you grow a company to a certain size, once you go past 50, you’re approaching the upper limits of how well interpersonal connections can get between humans. There’s studies into it. They say usually that the limit is probably about 200, but the 200 also includes your friends as well. So if you filter it out to just your workspace, then that’s probably about 50 or so. Beyond that, you really do not have the same kind of interpersonal connections. The web of communication gets too complicated. You need to put in systems and you need to put in communication protocols and hierarchies, so to speak. It doesn’t have to be very high up. It could be as flat as you want to make it to be, but there needs to be a system.</p><p>Joining a big organization from the very beginning in my career really helped with that. So I think that for a lot of young students, one of the key challenges that you might be thinking of is this: ‘Should I join a big company and get the money? They will usually pay me higher right at the beginning, I will get a relatively straightforward career trajectory, and I will learn a lot from there’. There’s pros and cons around it. The pros I got from joining the police force was learning about very structured leadership training and management training. Those two things are different, by the way. You might meet very good managers, but they are poor leaders, or vice versa. They are different skill sets.</p><p>I think another very important thing that I brought from my previous experience is, when you’re in the police force, you’re always managing crises and managing the worst things that ever happen to anyone. That gave me a very good and solid operational foundation. In the police force you need to make snap decisions in seconds usually, and if you do the wrong thing, someone may die. After that, and then translating it into the private sector, it’s different. I would say that psychologically it’s a bit easier. Mentally there might be very difficult decisions, but psychologically it’s easier because you do not need to make snap decisions in seconds. And if you make the wrong decision, no one is going to die. So to me, that’s already a win. I had good, solid, and foundational training with the police force and I’m eternally grateful to them for it.</p><p>Bringing that into the private sector, into what I’m doing now, for example as CEO, it’s a lot about identifying where the weaknesses in communications are, and putting people together with different interests. Different teams have different interests. Even if you’re in a relatively smaller company, you have to recognize that. And I think one of the big problems with people, especially in Web3, is that leadership training is not solid. Many people think that if they are in a leadership position, they look at the skill sets needed as the managerial skill sets. These include how to grow the company, how to scale, how to organize the company accordingly, how to hit the KPIs, how to bring in the money, the commercial stuff, etc. Those are great, they are managerial skills.<em> But leadership skills are different. They are from the heart.</em> Leaders come earlier in human evolution. When we talk about leadership skills, it’s always about bringing people together, convincing them to go above and beyond what they’re supposed to do, aligning them with the company’s vision and making it their own. And they say, ‘Yes, this is what I believe in, this is what I’m going to do’. So it’s different. Those skill sets are very hard to train, but they are trainable.”</p><h3><strong>What have been the differences, in terms of your workflow, transitioning from COO to CEO?</strong></h3><p>“Especially in Web3, COOs usually take on everything internal, and they take on everything that doesn’t fall under anyone else. So essentially when I was COO, I would say 80% to 90% of my time was focused internally on operational aspects. Things like how teams would work with each other. Things like operational campaigns, such as having a campaign going on and then the next one and the next one. How to push for commercialization of the product, etc. So those are operational issues and a lot of my job is really at a center node, bringing people together from marketing, community management, sales, business development, product, and even tech guys. The tech guys are usually a little bit untouched compared to everyone else because, by right, they shouldn’t be disturbed, they should focus solely on the code and building it. The product guys talk to them, and get all the requirements, so the product guys are usually the in-between. They understand the business needs from everyone else, such as marketing, and then they talk back to the tech guys.</p><p>As COO, I want to make sure that the whole machine is well-run. As a CEO, it’s a lot more external, especially in a Web3 sense. It’s a lot about talking to investors, building the brand name, building the recognizable aspects of the company in this industry. It’s a lot about finding out what’s going on out there, and knowing whether we are in the right market space or not, and maneuvering it along the way. Since its infancy until now, Web3 has probably been around for more than 14 years (when Bitcoin was first published, and the whole Bitcoin ecosystem was set up). But it is still in its infancy. Everyone is just trying something.</p><p>One of the major criticisms of Web3 is that it is a solution trying to find a problem. I would argue against that, but I could see why people say it. A lot of it is pushing the boundaries. In the first place, why blockchain and the related technologies around it were invented, was because of a problem (the 2008 financial crisis). There was a lack of trust regarding all the centralized entities, the banks, and everything started from there. There was a problem and blockchain was and still is a potential solution to that. But we’ve gone way beyond that now. We have tried to apply that principle and those technologies to all sorts of different use cases: DeFi, NFTs, insurance, supply chains etc. And once you do that, a lot of the use cases are very exploratory. And being in the space that I am in right now, which is blockchain data, there are a lot of potential use cases among our clients as well. They’re exploratory, which means that not all of them stick. Sometimes things fail and we have to navigate it, both from a technological standpoint, the product, and the sales standpoint. We have to say that, okay, we have to adapt to market changes. That’s one. But as much as possible, we should try to stay ahead of the market as well. Set the new trends, and be ready for when new waves come in so we can ride them. That’s basically what a lot of my job as a CEO entails.”</p><h3><strong>Some of the product offerings at PARSIQ include Tsunami API and Data Lakes. From the perspective of clients, what would you say are the benefits of using one?</strong></h3><p>“Tsunami API — tsunami is basically the brand name. If you say a blockchain API solution, that’s probably it. You can find many types out in the Web3 space right now. But essentially what an API solution does, including Tsunami API, is that there is a problem with blockchain data access where if you attach to a node and you want to pull raw data from a node, it is raw. So it is basically a series of zero and ones. That’s it. And you have to translate it into something that you can understand and that you can read as a human.</p><p>So there are different processes that you need to do. And even after you do them, you need to look at the blockchain data from block one to block, maybe 20 billion right now. Across the 20 billion chains are blocks of data. If you have a search query, it will search block 1 to try and find what you want. And then, it will go to the second block, the third block, all the way to block 20 billion. So that is essentially what you get from the raw data. If you run your own node, or you gain access to a node service, an API service usually indexes it, which makes it easier for you to do all this processing. Indexing means bookmarking it. And that’s essentially what the Tsunami API does. So it takes the entire blockchain, and it bookmarks the whole thing. So when you want to search for certain fields, the search engine, think about it this way, does not have to search through all 20 billion chains anymore to just flag up all those that are relevant. Maybe it’s just 1000 blocks and that’s it. And it makes it way faster, it makes it way neater. The power of our technology is that it indexes the latest block in milliseconds. Once it comes in, on the blockchain, the latest block with all the data inside our system already indexed the bookmarks in milliseconds, and our users can get it in milliseconds as well.</p><p>There are a lot of providers out there, but usually they are probably a couple of blocks behind. There’s historical reasons for it. I think part of the reason is because they come in and they build this whole suite solution. So the majority of the time is spent to make sure that the historical data is queried properly. And then as the latest block comes in, the newer the block, the harder it is to do it because you need to act very quickly, you need the developmental resources to do it, etc.</p><p>For <a href="https://www.parsiq.net/">PARSIQ</a>, we were actually purely a real time data service for 4 years of our existence from 2018 to 2022. Which means that we actually solved the hardest part and then we decided to build backwards to cover historical data and that makes us really fast. And that is probably purely on the generic Tsunami API. That is really our strength. If you want to query data, but you want to have the freedom to query whatever you want, you want the full data, then Tsunami API is probably for you.</p><p>Data lakes is probably where it gets really fun because I think in this space we are the only ones who actually do fully customized data. So why would people use the customized data rather than say the generic API? For the generic API, you’ll probably be thinking, “Oh, I don’t know whether my requirements may change, so I’d rather get the generic API”. Fair enough. But most of the time actually companies find that they get a generic API, they still need to process it further because they need the search queries, they may need to change the search queries, and after that they need to change it to the output that they need based on their specific use case.</p><p>That’s what Data Lakes is for. We fully customize the output, and we fully customize the inputs. There are very complex projects, especially now compared to 2021. Things have gotten a lot more complex. Especially when you are a project or a business building on blockchain, you might be looking at your own project and already have 100 different smart contracts just purely on your platform. But very often you are not only monitoring your platform, you’re monitoring someone else’s platform, especially your competitors. And you find that maybe you need to look into, as a hypothetical example, 5000 smart contracts. Each smart contract may have 50 different conditions that you are looking into. We’re going to hundreds of thousands of different filters, so to speak. And if you ask any API solutions, they’re going to tell you no, you can’t do hundreds of thousands of filters. That doesn’t make sense and that is too complicated.</p><p>Then what happens traditionally, is that most people do it themselves. This takes a lot of development time and it takes a lot of adjustments. And if you need to change certain parameters, you need to redo the whole thing again. So it’s extremely time and money consuming. Our Data Lake basically takes the whole thing and says that for your inputs, even if it’s hundreds of thousands of different filters or fields, we change the inputs according to exactly what you need. And then we also change the outputs at the bottom that says that, if your output is in this format, we will give it to you in the same format. So as an end user or as a project or a company, you do not need to do anything else. You get what you are trying to find, and straight away you can do whatever you want with that data.”</p><h3><strong>What advice would you give Web3 students? What learning resources would you suggest for them to learn more about how to work with data on the blockchain?</strong></h3><p>“There are actually quite a number of good educational platforms available. And if you’re already studying data and you know how to code, you should go straight to the SDKs (software development kits) of various different platforms. Very often, especially if you’re new to Web3 but you know all the fundamentals, check out those blockchain educational pages, such as the <a href="https://ethereum.foundation/">Ethereum Foundation</a>. They will tell you, if you are building a project, these are the things you need to look out for. Building a smart contract is one. Data is another layer that is a complete monster you need to look into.</p><p>I think that data is something that most projects neglect. They look at the data layer and they say “It’s all on the blockchain, what’s so hard?”. Then, after their computers start running and the data volume starts coming in, they ask, “How do I connect to it?”. The first instinct is to just go to a block explorer and attach the API and then draw it. But then people realize that they need to process it further. That is where things escalate. You realize that you need to put in a lot more resources than you actually initially allocated for, especially on the data layer.</p><p>Do your own research as much as possible. Find out exactly what you need. Talk to projects like us, <a href="https://www.parsiq.net/">PARSIQ</a>, find out our strengths. There are other people in this space that are trying to create a solution for data as well. And I think that’s great. Everyone has different strengths. For <a href="https://www.parsiq.net/">PARSIQ</a>, our strength is we have extremely fast, and extremely customized data. That’s what we dedicate to our clients, and this is what we want to give to them. So we have AI projects coming to us, and they use us because we are literally the only company that can give them what they want. They crash our servers, that’s how much they’re using us. And we fix it within half an hour, for example. And they crash our service perhaps like twice, and then it never happens again because we grow with them as well.</p><p>Once you build further and you have more experiences, joining hackathons becomes a good way to learn the basics. But at the end of the day you need to ask yourself how serious are you in terms of the idea that you want to bring into fruition? Talk to people, join incubators if you are really interested in it because there’s going to be a lot more help from those platforms. It’s not just about building. You need to learn about the business side of things. You will learn about bringing together the team.</p><p>One last advice is when you’re forming a team, form it with the least number of people with the most number of skill sets. So make sure there’s no two people with duplicative skill sets. If both of them have the exact same skill sets that they’re bringing to a table, then one of them is redundant. Form the team with the smallest number of people. Make sure everyone is in it together. They have skin in the game and they are willing to go all out for this and then start building from there. Bring in different skill sets. That’s very important. The tech guys, maybe you just need one person, maybe you need two. It depends on what kind of skill sets they bring.</p><p>But you always need the financial commercial guys. At the end of the day, someone must sell this product. Operational guys as well, don’t neglect that. Bring in someone with the organizational skill sets, but it’s fine if everyone comes from university and you don’t really have the work experience. Find someone from the club or from the usual work assignments that you already know. Someone who is pretty good with organizing things, structuring things, who thinks like an engineer basically. And I think that is a skill set that is often neglected as well.”</p><h3><strong>Closing Remarks</strong></h3><p>Thanks for reading (or listening)! We deeply appreciate Rong Kai Wong for sharing his insights and lessons from a lifetime of valuable experiences. This interview is one of many insightful discussions we’ve had with industry leaders. If you’re keen on delving deeper into similar stories, we invite you to check out our interview with <a href="https://medium.com/@mustafa.hourani/interview-with-kevin-henrikson-ceo-of-dust-labs-b560673d613d">Kevin Henrikson from Dust Labs</a>. We’re committed to bringing you more in-depth conversations with the frontrunners of Web3. Stay tuned!</p><p>To connect with Rong Kai Wong, and learn more about the amazing work underway at <a href="https://www.parsiq.net/">PARSIQ</a>, please check out the links below:</p><p><a href="https://www.linkedin.com/in/rong-kai-wong/">Rong Kai’s LinkedIn</a> | <a href="https://twitter.com/RongKaiPARSIQ">Rong Kai’s Twitter</a> | <a href="https://www.parsiq.net/">PARSIQ Website</a> | <a href="https://twitter.com/parsiq_net">PARSIQ Twitter</a></p><p><em>*Originally published under the Academy Ventures Research Lab. Access the original article </em><a href="https://www.academyventures.xyz/post/full-interview-with-rong-kai-wong-ceo-of-parsiq"><em>here</em></a><em>.</em></p><p><em>*Also published on PARSIQ’s Medium page. Access that article </em><a href="https://medium.com/parsiq/exclusive-interview-with-rong-kai-wong-ceo-of-parsiq-7f174d4fc46b"><em>here</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=39d9c0328b96" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Interview with Kevin Henrikson, CEO of Dust Labs]]></title>
            <link>https://medium.com/@mustafaxyz9/interview-with-kevin-henrikson-ceo-of-dust-labs-b560673d613d?source=rss-5bd8d2734ef5------2</link>
            <guid isPermaLink="false">https://medium.com/p/b560673d613d</guid>
            <category><![CDATA[dust-labs]]></category>
            <category><![CDATA[delabs]]></category>
            <category><![CDATA[degods]]></category>
            <category><![CDATA[y00ts]]></category>
            <category><![CDATA[kevin-henrikson]]></category>
            <dc:creator><![CDATA[Mustafa]]></dc:creator>
            <pubDate>Wed, 04 Oct 2023 13:39:53 GMT</pubDate>
            <atom:updated>2023-10-04T22:40:49.313Z</atom:updated>
            <content:encoded><![CDATA[<p><em>The key differences between a startup journey in the Web2 world versus the Web3 world, along with an honest examination of the challenges encountered when building a Web3 NFT company.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8U120mJfyqAcwHY-dczoCQ.png" /></figure><p>This week, Academy Ventures is excited to present an exclusive, rapid-fire interview with Kevin Henrikson, the founder and CEO of <a href="https://www.dustlabs.com/">Dust Labs</a>. <a href="https://www.dustlabs.com/">Dust Labs</a> is the software company behind two of the most popular non-fungible token (NFT) collections: <a href="https://opensea.io/collection/degods">DeGods</a> and <a href="https://opensea.io/collection/y00ts">y00ts</a>. By providing valuable B2B SaaS tools to other Web3 companies, <a href="https://www.dustlabs.com/">Dust Labs</a> places itself at the forefront of innovation within the NFT vertical of the Web3 realm.</p><p>Beyond his work at <a href="https://www.dustlabs.com/">Dust Labs</a>, Kevin Henrikson has an impressive track record in the Web2 entrepreneurial landscape. He served as the Director of Engineering at <a href="https://en.wikipedia.org/wiki/Zimbra">Zimbra</a>, a software suite encompassing an email server and web client, which was acquired by <a href="https://en.wikipedia.org/wiki/Yahoo!_Inc._(1995%E2%80%932017)">Yahoo! </a>for $350 million in 2007. Moreover, he co-founded <a href="https://en.wikipedia.org/wiki/Acompli">Acompli</a>, a mobile email app, and served as its VP of Engineering before its acquisition by <a href="https://en.wikipedia.org/wiki/Microsoft">Microsoft</a> for $200 million in 2014. Most recently, Henrikson was VP of Engineering at <a href="https://en.wikipedia.org/wiki/Instacart">Instacart</a>, a popular grocery delivery company in the USA.</p><p>In this rapid-fire interview, Kevin Henrikson candidly answers questions about his experiences as a founder, the journey of building a startup, the challenges associated with working in the Web3 environment, and the advice he would offer to aspiring college student founders.</p><h3><strong>How has your background as an experienced Web2 founder and tech leader influenced your role as the CEO of Dust Labs?</strong></h3><p>“Honestly, everything I learned in the Web2 world is helping me a ton at Dust Labs. It’s given me a solid grip on the nuts and bolts of running a tech company, and it guides how I build teams, make products, and think about users.”</p><h3><strong>How has your experience in Web3 changed your perception and understanding of traditional Web2 companies?</strong></h3><p>“Hanging out in Web3 is like learning the Internet all over again, but with a cool twist. The big lightbulb moment for me is how Web3 flips the power dynamics — it’s all about giving control and benefits to the users, which is totally different from traditional Web2 setups.”</p><h3><strong>Based on your experience, what is the make-or-break step in the startup journey?</strong></h3><p>“In my book, the make-or-break thing for startups is the team. Find the right gang who can vibe together, who’ve got the chops, and who are jazzed about what you’re doing, and you’ll be able to face whatever crazy stuff comes your way.”</p><h3><strong>How do community NFT projects compare to typical Web2 startups?</strong></h3><p>“Community NFT projects are like Web2 startups on a community adrenaline rush. They’re more of a team sport, with the community right in the driver’s seat influencing where things go, not just along for the ride.”</p><h3><strong>How do you balance the experimental approach of Dust Labs/DeLabs with the traditional product-market fit model that most companies strive to achieve?</strong></h3><p>“At <a href="https://www.dustlabs.com/">Dust Labs</a>/<a href="https://delabs.xyz/">DeLabs</a>, it’s about having our cake and eating it too. We like to push the envelope and play around with new ideas, but we’re always super focused on our community and making sure we’re delivering the good stuff they want.”</p><h3><strong>What was the biggest challenge when bridging from the Solana blockchain to the Ethereum blockchain?</strong></h3><p>“Making the jump from <a href="https://solana.com/">Solana</a> to <a href="https://ethereum.org/en/">Ethereum</a> was like changing trains at rush hour while juggling flaming torches. We had to get deep into the weeds of both ecosystems and make sure everything was safe and sound. Lots of coffee was consumed.”</p><h3><strong>How does your team balance customer empathy and meeting expectations with a risk-taking attitude that could potentially elicit a negative community reaction?</strong></h3><p>“For us, it’s all about keeping it real. We listen to our community, take risks, and sometimes that can ruffle some feathers. But when it happens, we don’t shy away — we listen, learn, and tweak things based on what we hear.</p><h3><strong>What advice would you give to college students looking to build their own Web3 community project? How should they manage high expectations from community members?</strong></h3><p>“For college students dreaming up their own Web3 project, my advice is this: be stoked about what you’re doing, and have a clear vision for the tribe you want to build. And remember, keeping it real with your community, talking to them, and delivering the goods is the key to managing their expectations.”</p><h3>Closing Remarks</h3><p>This concludes our most recent startup interview with Kevin Henrikson, the founder and CEO of <a href="https://www.dustlabs.com/">Dust Labs</a>. We hope this can be a valuable resource for college students interested in Web3. If you would like to gain insight into the startup journey from a different sector within Web3, we invite you to read our previous interview with <a href="https://medium.com/@mustafa.hourani/interview-with-taiko-a-leading-type-1-zkevm-ddf71eb4eabe">Taiko, a Zk-EVM</a>.</p><p>To connect with Kevin Henrikson and learn more about the amazing work underway at <a href="https://www.dustlabs.com/">Dust Labs</a>, please check out the links below:</p><p><a href="https://www.linkedin.com/in/kevinhenrikson/">Kevin’s LinkedIn</a> | <a href="https://twitter.com/kevindegods">Kevin’s Twitter</a> | <a href="https://www.dustlabs.com/">Dust Labs Website</a> | <a href="https://twitter.com/dust_labs">Dust Labs Twitter</a></p><p><em>*Originally published under the Academy Ventures Research Lab. Access the original article </em><a href="https://www.academyventures.xyz/post/interview-with-kevin-henrikson-ceo-of-dust-labs"><em>here</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b560673d613d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Interview with Taiko, a Leading Type-1 zkEVM]]></title>
            <link>https://medium.com/@mustafaxyz9/interview-with-taiko-a-leading-type-1-zkevm-ddf71eb4eabe?source=rss-5bd8d2734ef5------2</link>
            <guid isPermaLink="false">https://medium.com/p/ddf71eb4eabe</guid>
            <category><![CDATA[taiko]]></category>
            <category><![CDATA[zero-knowledge-proofs]]></category>
            <category><![CDATA[zero-knowledge]]></category>
            <category><![CDATA[zkevm]]></category>
            <category><![CDATA[ethereum]]></category>
            <dc:creator><![CDATA[Mustafa]]></dc:creator>
            <pubDate>Wed, 04 Oct 2023 13:22:49 GMT</pubDate>
            <atom:updated>2023-10-04T13:26:01.386Z</atom:updated>
            <content:encoded><![CDATA[<p><em>The challenges and opportunities with zero-knowledge technology and how Taiko, a type-1 zkEVM, is building an efficient solution to scale the Ethereum blockchain.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*r5GPNLPQFycWXY4HDaSJcw.png" /></figure><p>At Academy Ventures, we believe that zero-knowledge (ZK) solutions are poised to play a crucial role in the scalability of various public blockchains in the coming months and years. ZK enhances privacy and opens up new possibilities for more efficient blockchain interoperability, allowing greater scalability. The demand for secure, scalable solutions in the space will continue to grow. ZK solutions are positioned to meet this demand and will likely be at the forefront of innovation in the blockchain industry.</p><p>This week, we are excited to announce that we met with Ben Wan, the Community Manager at Taiko. We interviewed him to learn more about their ZK-EVM project, challenges with implementation, and the future of ZK. We hope this interview can inform our community about the potential (and limitations) of this widely discussed topic in the space.</p><p>Let’s dive straight in!</p><h3><strong>There are several zkEVM projects out there, like ZKsync and Starknet. How is Taiko different?</strong></h3><p>“Before talking about the difference, we first need to be aware that <a href="https://vitalik.ca/general/2022/08/04/zkevm.html">there are different types of ZK-EVMs</a>, as described in the article written by Vitalik Buterin. There are currently 4 types if you exclude the nomenclatures in between. For all intents and purposes, the greatest difference is that with the exception of type 1, all of the others require some modification to Ethereum or the EVM. Taiko is developing a <a href="https://taiko.mirror.xyz/w7NSKDeKfJoEy0p89I9feixKfdK-20JgWF9HZzxfeBo">Type 1 (Ethereum-equivalent) ZK-EVM</a>. This means Taiko aims to have Ethereum equivalence on Layer 2 (L2). Its purpose is to allow developers and users of dApps developed for Ethereum Layer 1 (L1) to be used on Taiko’s L2 without any changes. As a result, dApps can be easily deployed to L2, inheriting Ethereum’s security while incurring lower transaction fees than on L1. Other types (types 2 through 4) make a tradeoff between Ethereum-equivalence in favor of generating ZK proofs more quickly or more cheaply.</p><p>Another difference is that Taiko aims to be fully decentralized from day 1 by being permissionless, meaning anyone with the required hardware will be able to participate in the proposing and proving process on L2.</p><p>To summarize the difference, Taiko aims to replicate Ethereum L1 exactly as is into our rollup, whereas others modify certain aspects of Ethereum or EVM.”</p><h3><strong>Can you discuss the tradeoff between decentralization and security vs. scalability on the Ethereum blockchain?</strong></h3><p>“Ethereum’s consensus mechanism, Proof of Stake (PoS), which was formerly Proof of Work (PoW), uses a large number of nodes to validate and record transactions. This is what gives Ethereum its security and decentralization. This does, however, have a limit on scalability because a lot of nodes are required to process each transaction, which slows down transaction processing times and raises the cost of executing transactions on the network.”</p><h3><strong>How does Taiko achieve scalability while maintaining decentralization and security?</strong></h3><p>“In general, there are 2 ways to improve scalability:</p><p>The first is on-chain which means that the Ethereum blockchain itself is updated to increase the amount of data and thus the number of transactions that can fit into each block.</p><p>The second is off-chain which commonly refers to L2 where transactions are moved to a separate layer and processed in batches before sending it back to the Ethereum L1 as a single transaction.</p><p>Taiko achieves scalability while maintaining security and decentralization through the use of ZK-EVM and rollup technology.</p><p>Rollups batch transactions and guarantee that as long as the L1 (data availability layer) exists, users can reconstruct the L2 state and exit the rollup by forcing a transaction on L1. Using compression, every batch can have thousands of transactions submitted to Ethereum that includes a cryptographic proof with minimal data that is verified by a contract deployed on Ethereum L1. ZK-Rollups thus allow certain network participants to simply verify a validity proof, rather than executing every transaction. Decentralization is maintained by ensuring that the main participants (proposer, provers and nodes) can be run by a multitude of parties.”</p><h3><strong>What are the limitations forcing many ZK-Rollups to be application-specific rather than general-purpose?</strong></h3><p>“The main limitation is that each smart contract or dApp needs to be written or adapted specifically to work within the constraints and design of the ZK-Rollup. It is much easier to design ZK circuits that perform specific tasks, which cover the vast majority of what users will do with a particular application, making it application specific. This means that ZK-Rollups are not always compatible with every type of smart contract or dApp, and integrating with existing codebases may require significant engineering effort.</p><p>Creating a general-purpose rollup requires designing a ZK circuit for each individual EVM opcode, which was previously thought to be computationally challenging. Some EVM opcodes are problematic because they are not ZK-friendly. This means that even the fastest computers would take an inordinate amount of time to prove.</p><p>The development of new proving systems, such as PLONK, reduces the number of operations required to construct a proof, making it more efficient to generate proofs for ZK unfriendly computations. Other mechanisms such as parallel proving also speeds up the proof time, resulting in more general-purpose friendliness.”</p><h3><strong>How does Taiko use ZK-EVM to develop general-purpose rollups?</strong></h3><p>“The ZK-EVM uses validity proofs to prove the correctness of the EVM computations on the rollup. Taiko implements a ZK-EVM that supports every EVM opcode, resulting in a <a href="https://github.com/privacy-scaling-explorations/zkevm-circuits">ZK-EVM circuit</a> validity proof. This bytecode compatibility means that all Ethereum tooling integrates seamlessly with Taiko, without developers requiring code changes. The actual ZK-EVM circuits we use and contribute to are an open community effort, involving the Ethereum Foundation and others.</p><p>For a general-purpose rollup, as opposed to an app-specific rollup, users must be able to force the inclusion of any arbitrary transaction, not just ‘exit’ transactions. From the proposed blocks, provers generate SNARK proofs asserting the validity of L2 transactions and blocks. Taiko enables this by being permissionless and decentralized from day one which means all valid transactions are processed.</p><p>Overall, Taiko uses ZK-EVM to develop a general purpose rollup by leveraging the succinctness and verifiability of zero-knowledge proofs to ensure the integrity of off-chain computations, while also improving the scalability of the Ethereum network.”</p><h3><strong>What are some key milestones and events to look forward to in 2023 with Taiko?</strong></h3><p>“We will release further testnets that will include tokenomics and the ZK proving components in the coming months. This will make it possible for the community to contribute more realistically by testing transaction speeds, costs, running nodes, serving as proposers or provers, and for developers to evaluate our protocol and architecture by putting their dApps into use and stress testing the environment. Additionally, you will definitely see and hear more from Taiko at conference events and hackathons throughout 2023. Watch out for more exciting developments by following our announcements on <a href="https://twitter.com/taikoxyz">Twitter</a> and <a href="https://discord.com/invite/taikoxyz">Discord</a> or swing by the events and say hello if you are in the neighborhood!”</p><h3><strong>What are some good resources for a college student looking to learn more about ZK?</strong></h3><p>“<a href="https://ethereum.org/en/developers/docs/scaling/zk-rollups/">Ethereum.org</a> would be a great place to start to learn more about ZK-rollups and how this benefits the ecosystem. There are other more in-depth great articles <a href="https://a16zcrypto.com/posts/article/zero-knowledge-canon/">here</a>. Attending conferences and workshops related to cryptography and blockchain technology would provide opportunities to learn from experts in the field and network with others interested in ZK. And of course who can forget youtube for not only simple explanations but even recordings from event presentations.”</p><h3><strong>What advice would you give a college student seeking a career in the ZK domain?</strong></h3><p>“You should have a keen interest in computer science, mathematics and cryptography. These principles form the basis of ZK and will continue to drive advancements in this field. Keep yourself up to date by participating in hackathons, conferences, workshops, online communities, and other events to get acknowledged for your accomplishments and find career opportunities. Above all, have fun while doing it!”</p><h3><strong>Closing Remarks</strong></h3><p>To conclude this week’s inside scoop on ZK-EVM, we would like to thank Ben Wan and the rest of the Taiko team for their time and valuable insight.</p><p>To learn more about Taiko and follow their updates, we’ve linked their website and contact information below for your convenience.</p><p><a href="https://taiko.xyz/">Website</a> | <a href="https://discord.com/invite/taikoxyz">Discord</a> | <a href="https://twitter.com/taikoxyz">Twitter</a></p><p><em>*Originally published under the Academy Ventures Research Lab. Access the original article </em><a href="https://www.academyventures.xyz/post/exclusive-interview-with-taiko"><em>here</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ddf71eb4eabe" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>