<?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[Blockchain DYOR(2026) - Medium]]></title>
        <description><![CDATA[breifly explained blockchain world - Medium]]></description>
        <link>https://medium.com/blockchain-dyor?source=rss----d2e4ea0bbebc---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Blockchain DYOR(2026) - Medium</title>
            <link>https://medium.com/blockchain-dyor?source=rss----d2e4ea0bbebc---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 01:42:35 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/blockchain-dyor" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Ethereum Staking, Lido, and Eigenlayer — The Same ETH Working Three Jobs]]></title>
            <link>https://medium.com/blockchain-dyor/ethereum-staking-lido-and-eigenlayer-the-same-eth-working-three-jobs-c5830265a67b?source=rss----d2e4ea0bbebc---4</link>
            <guid isPermaLink="false">https://medium.com/p/c5830265a67b</guid>
            <category><![CDATA[restaking]]></category>
            <category><![CDATA[staking]]></category>
            <category><![CDATA[liquid-staking]]></category>
            <category><![CDATA[eigenlayer]]></category>
            <category><![CDATA[ethereum]]></category>
            <dc:creator><![CDATA[0x365home]]></dc:creator>
            <pubDate>Tue, 14 Apr 2026 17:09:00 GMT</pubDate>
            <atom:updated>2026-04-14T17:09:00.006Z</atom:updated>
            <content:encoded><![CDATA[<p>Most people treat staking as a set-and-forget yield strategy. Deposit ETH, earn rewards, done. But once you understand the full stack — from PoS mechanics to liquid staking to restaking — you realize this is one of the most architecturally interesting things happening in DeFi right now. And one of the riskiest.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/920/1*uH9zvOhalUvyq08KWnyWTA.png" /></figure><p><strong>1/ Why staking = security</strong></p><p>Ethereum’s PoS consensus requires 2/3 of validators to agree before a block is finalized. To attack the network, you’d need to control 2/3 of all staked ETH. With over $100B staked, that’s an astronomical cost.</p><p>The mechanism that makes this work is slashing. If a validator behaves maliciously — signing conflicting blocks, lying about transactions — a portion of their staked ETH gets burned automatically. The punishment is on-chain and non-negotiable.</p><p>So staking isn’t just yield farming. It’s putting skin in the game. More ETH staked = higher attack cost = stronger security. The two are directly linked.</p><p><strong>2/ The 32 ETH problem — why Lido exists</strong></p><p>Running a validator requires exactly 32 ETH. At current prices, that’s around $60,000–80,000. For most people, that’s inaccessible.</p><p>Lido solves this by pooling ETH from thousands of users, running validators on their behalf, and issuing stETH as a receipt. You deposit 0.1 ETH, you get 0.1 stETH. The stETH accrues staking rewards automatically — its value increases relative to ETH over time.</p><p>The key innovation is liquidity. Direct staking locks your ETH — unstaking takes time. stETH is tradeable. You earn staking yield while keeping the ability to exit whenever you want.</p><p>This is liquid staking. Yield without lockup.</p><p><strong>3/ The problem with stETH sitting idle</strong></p><p>Once you have stETH, you’re earning ~3–4% APY from Ethereum staking. But stETH is just sitting in your wallet, doing one job.</p><p>What if it could do two?</p><p>That’s the question Eigenlayer answered.</p><p><strong>4/ Eigenlayer — restaking</strong></p><p>Eigenlayer lets you deposit stETH as collateral and use it to secure additional networks called AVSs (Actively Validated Services). These are new protocols — bridges, oracles, new blockchains — that need validator networks but don’t have them yet.</p><p>Building a validator network from scratch takes years. You need validators, you need them to be credibly committed, and you need enough capital staked to make attacks expensive. Eigenlayer lets new protocols skip that bootstrapping problem entirely — they inherit Ethereum’s validator set and security.</p><p>In exchange, restakers earn additional rewards from the AVS on top of their Ethereum staking yield. One stETH. Two revenue streams.</p><p><strong>5/ But here’s the risk nobody talks about enough</strong></p><p>The same stETH is now collateral for multiple systems simultaneously. If you’re restaked across three AVSs, all three have a claim on your stETH if slashing occurs.</p><p>AVS A can slash you for bad behavior on its network. AVS B can slash you independently. And Ethereum can slash you for failures on its network. These events are not mutually exclusive.</p><p>This creates layered slashing risk — the same capital exposed to multiple independent failure conditions at once. If a poorly-designed AVS triggers mass slashing events, the stETH value drops for everyone who restaked — including people who had nothing to do with that AVS.</p><p>And because Lido pools ETH from thousands of users, a large-scale slashing event doesn’t just hurt restakers. It ripples through to anyone holding stETH.</p><p><strong>6/ The composability parallel</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GeXzu01Qha2Nz5xweT-YZA.png" /></figure><p>This is the same structural risk pattern we see in DeFi composability. Stack Aave → Curve → Convex and you get compounded yield. You also get compounded smart contract risk — a Curve exploit hits all three layers simultaneously.</p><p>Restaking is the same idea applied to security infrastructure. Stack Ethereum → Eigenlayer → AVS and you get compounded yield. You also get compounded slashing exposure.</p><p>More layers = more yield = more ways to get wrecked.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bhfDAQfeiME4mk_XLvJ49w.png" /></figure><p><strong>7/ The question worth asking</strong></p><p>Eigenlayer is solving a real problem. New protocols genuinely need bootstrapped security, and restaking is an elegant mechanism for providing it.</p><p>But the design assumes that AVS slashing conditions are well-constructed, that validators are rational about which AVSs they join, and that the aggregate slashing risk stays manageable.</p><p>Those assumptions deserve more scrutiny than they’re getting.</p><p>One badly designed AVS with a large restaker base could trigger cascading effects across the entire stETH ecosystem. The more interconnected the system becomes, the more a single failure point matters.</p><p>That’s the tradeoff at the core of restaking: you’re borrowing Ethereum’s security to bootstrap new networks. But you’re also importing new risks back into Ethereum’s security layer.</p><p>Whether that’s worth it depends entirely on how well the risks are priced.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lxw9rEYHiMqwb_FSDD2kLA.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c5830265a67b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/blockchain-dyor/ethereum-staking-lido-and-eigenlayer-the-same-eth-working-three-jobs-c5830265a67b">Ethereum Staking, Lido, and Eigenlayer — The Same ETH Working Three Jobs</a> was originally published in <a href="https://medium.com/blockchain-dyor">Blockchain DYOR(2026)</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[DAI, UST, FRAX — Three Stablecoins. Three Different Answers to the Same Problem.]]></title>
            <link>https://medium.com/blockchain-dyor/dai-ust-frax-three-stablecoins-three-different-answers-to-the-same-problem-0ad479c3040e?source=rss----d2e4ea0bbebc---4</link>
            <guid isPermaLink="false">https://medium.com/p/0ad479c3040e</guid>
            <category><![CDATA[makerdao]]></category>
            <category><![CDATA[stable-coin]]></category>
            <category><![CDATA[ust]]></category>
            <category><![CDATA[frax]]></category>
            <category><![CDATA[sky]]></category>
            <dc:creator><![CDATA[0x365home]]></dc:creator>
            <pubDate>Mon, 13 Apr 2026 16:13:20 GMT</pubDate>
            <atom:updated>2026-04-13T16:13:20.016Z</atom:updated>
            <content:encoded><![CDATA[<p>Every stablecoin is trying to solve one thing: how do you keep a token worth exactly $1 without a central bank backing it?</p><p>Three protocols answered that question differently. One collapsed. Two survived. Here’s why.</p><p><strong>1/ The problem with stablecoins</strong></p><p>A stablecoin needs to hold its $1 peg even when markets go crazy. That’s easy if you just hold $1 in a bank — but then you’ve rebuilt the traditional financial system inside crypto. Pointless.</p><p>The real challenge is: how do you maintain the peg in a decentralized way, without depending on banks or governments?</p><p>Three different philosophies emerged.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gbEEtQMg0dlu1GvdZCpk2A.png" /></figure><p><strong>2/ UST — Pure algorithm. No collateral.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WWCopQTKdN9F__F30qSRQQ.png" /></figure><p>Terra’s UST maintained its peg through an exchange mechanism with LUNA. $1 of UST could always be swapped for $1 worth of LUNA, and vice versa.</p><p>When UST dropped below $1 → arbitrageurs bought UST cheap → swapped for LUNA → sold LUNA → UST supply decreased → price recovered.</p><p>Elegant in theory. Fatal in practice.</p><p>The problem: LUNA was Terra’s own token. When UST lost confidence, people sold UST → minted LUNA to cover → LUNA supply exploded → LUNA price crashed → UST collateral worth less → more UST sold.</p><p>A circular structure. Once confidence broke, there was nothing to stop the spiral.</p><p>$40B gone in a week.</p><p><strong>3/ DAI — Overcollateralized. External assets.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UyIvzJpYdWLNt_L0bfkclQ.png" /></figure><p>MakerDAO took the opposite approach. To mint $100 DAI, you need to deposit $150+ of ETH. Real, external collateral — not a token MakerDAO controls.</p><p>If ETH crashes, there’s still buffer before DAI becomes undercollateralized. Automated liquidations kick in before the system breaks.</p><p>And if the system takes a loss anyway? MKR token holders are the insurer of last resort. The protocol mints new MKR and sells it to cover the gap — diluting MKR holders in the process.</p><p>MKR holders accept that risk in exchange for the upside: when the system runs well, protocol fees are used to buy back and burn MKR — increasing its value.</p><p>The tradeoff: overcollateralization means capital inefficiency. You lock up $150 to borrow $100. A lot of value sitting idle.</p><p><strong>4/ MakerDAO’s recent evolution — RWA</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TkjeNANue8v5BcgKVjiPhQ.png" /></figure><p>To solve the capital efficiency problem, MakerDAO started accepting US Treasury bonds as collateral.</p><p>Treasuries earn yield (~4–5%). ETH just sits there. Suddenly the protocol has real revenue. MakerDAO’s income exploded after RWA adoption.</p><p>But there’s a catch — and it’s a fundamental one.</p><p>If the US government sanctions MakerDAO, the Treasury collateral gets frozen. A supposedly decentralized protocol becomes dependent on Washington’s goodwill.</p><p>That’s why MakerDAO rebranded to Sky. They’re wrestling with this tension out loud.</p><p><strong>5/ FRAX — The middle path</strong></p><p>Frax looked at both extremes and built something between them.</p><p>To mint $1 FRAX: $0.85 USDC (real collateral) + $0.15 FXS (Frax’s own token).</p><p>Not 100% algorithmic. Not 100% collateralized. The ratio isn’t fixed either — it adjusts automatically based on market conditions. When FRAX is stable and demand is high, the collateral ratio drops. When it’s unstable, the ratio rises. More conservative, more capital returned.</p><p>The excess USDC doesn’t just sit there either. It gets deployed into Aave, Curve, and other protocols to earn yield — making Frax a revenue-generating machine, not just a stablecoin issuer.</p><p><strong>6/ The core lesson</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mIeYbrd9Wq52C1d8VYhYXw.png" /></figure><p>All three were trying to do the same thing. What separated them was one question:</p><p><strong>What happens when confidence breaks?</strong></p><p>UST → nothing to fall back on. Circular collapse. DAI → overcollateralized buffer + MKR backstop. Survived. FRAX → real buffer + yield-generating collateral + auto-adjusting ratio. Survived.</p><p>The lesson isn’t that algorithmic stablecoins are impossible. It’s that stablecoins without real external collateral will always be one confidence crisis away from zero.</p><p>MakerDAO and Frax both understood that. UST didn’t.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0ad479c3040e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/blockchain-dyor/dai-ust-frax-three-stablecoins-three-different-answers-to-the-same-problem-0ad479c3040e">DAI, UST, FRAX — Three Stablecoins. Three Different Answers to the Same Problem.</a> was originally published in <a href="https://medium.com/blockchain-dyor">Blockchain DYOR(2026)</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Frax — If you understand why Terra/Luna collapsed, you’ll see why this is an improved approach]]></title>
            <link>https://medium.com/blockchain-dyor/frax-if-you-understand-why-terra-luna-collapsed-youll-see-why-this-is-an-improved-approach-89861096f955?source=rss----d2e4ea0bbebc---4</link>
            <guid isPermaLink="false">https://medium.com/p/89861096f955</guid>
            <category><![CDATA[frax]]></category>
            <category><![CDATA[stable-coin]]></category>
            <category><![CDATA[luna]]></category>
            <category><![CDATA[terra]]></category>
            <dc:creator><![CDATA[0x365home]]></dc:creator>
            <pubDate>Mon, 06 Apr 2026 16:36:28 GMT</pubDate>
            <atom:updated>2026-04-06T16:36:28.510Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/500/0*nIys6LEgJ5s-xL_3.png" /></figure><p>Remember UST? Dollar-pegged stablecoin with no real collateral. LUNA — its own token — was the backing. When LUNA price dropped, UST collapsed with it. A circular structure. $40B gone in a week.</p><p>Frax was built with that failure in mind.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*E1vnqtETQXn1phUGicTXnQ.png" /></figure><p><strong>1/ Core structure — partial collateral</strong></p><p>To mint $1 of FRAX:</p><ul><li>$0.85 USDC (real collateral)</li><li>$0.15 FXS (Frax’s native token)</li></ul><p>Not 100% algorithmic. Not 100% collateralized. Somewhere in between.</p><p>What’s different from UST? Even if FXS crashes, the 85% USDC acts as a buffer. For FRAX to fully collapse, USDC would have to go down with it. Structurally far more resilient.</p><p><strong>2/ CR — the collateral ratio moves automatically</strong></p><p>FRAX price above $1 → strong demand signal → stable → lower the collateral ratio. FRAX price below $1 → instability → raise the collateral ratio → more conservative.</p><p>This gets calculated on-chain automatically. No human decision required.</p><p>Why does lowering CR matter? The locked-up USDC gets freed. That USDC can now generate yield.</p><p><strong>3/ The idle USDC works</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7dqK1zaoAbm3UK0eX_w63Q.png" /></figure><p>Collateral doesn’t just sit in a vault. It gets deployed into protocols like Aave and Curve to earn interest. That yield flows back to FXS holders.</p><p>UST had no yield structure. The 20% Anchor APY was a foundation subsidy — not real revenue.</p><p>Frax is different. The collateral itself generates income. That’s what makes it sustainable.</p><p><strong>4/ Why Frax was the most aggressive player in Curve Wars</strong></p><p>For a stablecoin, liquidity is trust.</p><p>If the FRAX/USDC pool on Curve is thin — slippage grows — FRAX doesn’t trade at $1 — people lose confidence and sell — price drops further.</p><p>That’s why Frax was throwing bribes at vlCVX holders on Votium so aggressively. Deep Curve liquidity wasn’t a nice-to-have. It was existential.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ul_v0Uz9v532WJ7pgyc29Q.png" /></figure><p><strong>5/ What actually makes it different</strong></p><p>UST → no collateral, no yield structure, pure algorithm → collapse FRAX → real buffer, collateral generates yield, auto-adjusts to market → survives</p><p>Terra/Luna’s lesson wasn’t “algorithmic stablecoins are impossible.” It was “algorithmic stablecoins without real collateral will always die.”</p><p>Frax took that lesson and built it into the design.</p><p>reference : <a href="https://www.coingecko.com/learn/what-is-frax-crypto">https://www.coingecko.com/learn/what-is-frax-crypto</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=89861096f955" width="1" height="1" alt=""><hr><p><a href="https://medium.com/blockchain-dyor/frax-if-you-understand-why-terra-luna-collapsed-youll-see-why-this-is-an-improved-approach-89861096f955">Frax — If you understand why Terra/Luna collapsed, you’ll see why this is an improved approach</a> was originally published in <a href="https://medium.com/blockchain-dyor">Blockchain DYOR(2026)</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[[Curve Finance] How this liquidity war happened]]></title>
            <link>https://medium.com/blockchain-dyor/curve-finance-how-this-liquidity-war-happened-011d1e9fa03d?source=rss----d2e4ea0bbebc---4</link>
            <guid isPermaLink="false">https://medium.com/p/011d1e9fa03d</guid>
            <category><![CDATA[curve-finance]]></category>
            <category><![CDATA[uniswap]]></category>
            <category><![CDATA[stable-coin]]></category>
            <category><![CDATA[curve-wars]]></category>
            <category><![CDATA[amm]]></category>
            <dc:creator><![CDATA[0x365home]]></dc:creator>
            <pubDate>Wed, 01 Apr 2026 15:23:47 GMT</pubDate>
            <atom:updated>2026-04-01T15:23:47.341Z</atom:updated>
            <content:encoded><![CDATA[<p>Curve Finance held over $25B in TVL at its peak. Here’s how it was built, how it sparked a governance war, and why it created the very problem it was designed to solve.</p><p>① The problem: Uniswap wasn’t built for only for stablecoins</p><p>Uniswap uses the x*y=k formula. Swap USDC for USDT and the same math applies — even though both tokens are worth $1 and should trade 1:1.</p><p>Swap $10M and slippage explodes. You could lose hundreds of thousands of dollars on a trade between two assets that are supposed to be equivalent.</p><p>Curve was built to fix this. Its StableSwap formula recognizes that similarly-priced assets should trade near parity. In the range where both assets hover around $1, the curve flattens almost to a straight line. Slippage on that same $10M swap drops to near zero. That’s how Curve took over the stablecoin market.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/930/1*-F8HzucXprRJOZd1oQbZJg.png" /></figure><p>② The incentive problem: how do you attract liquidity?</p><p>Building a better pool isn’t enough. LPs need a reason to deposit.</p><p>Curve’s answer was CRV token emissions. Provide liquidity, earn CRV. But here’s the next question: which pools get how much? That’s where the ve model comes in.</p><p>③ The ve model: voting power determines incentives</p><p>Lock CRV (up to 4 years) and receive veCRV. Use veCRV to vote in Gauge elections. The vote determines how CRV emissions are distributed across pools.</p><p>Two design goals drove this:</p><p>First, dump prevention. Token incentives in DeFi Summer were sold the moment they landed. Locking forces commitment — only long-term holders get governance power.</p><p>Second, genuine participation. Short-term speculators out, real ecosystem stakeholders in.</p><p>④ Curve Wars: liquidity became survival</p><p>Think from a stablecoin issuer’s perspective. If your stablecoin can’t be swapped in large size without price impact, confidence collapses. Deep Curve liquidity isn’t a nice-to-have — it’s existential.</p><p>So FRAX, UST, MIM, and others went to war over veCRV. More veCRV → more votes → more emissions directed to your pool → higher APY → deeper liquidity → price stability. This was the Curve Wars.</p><p>⑤ Convex: the middleman takes the throne</p><p>Locking CRV directly means your capital is frozen for up to 4 years. No exit. No liquidity.</p><p>Convex solved this. Deposit CRV, Convex locks it on your behalf. In return, you receive cvxCRV — tradeable anytime. And on top of that, CVX tokens as an additional reward layer. Lock CVX to get even more yield.</p><p>Same CRV in. More yield out. Everyone moved to Convex.</p><p>The result: Convex accumulated over 40% of all veCRV.</p><p>⑥ Votium: bribes became a market</p><p>Locked CVX becomes vlCVX — Convex’s governance token. And vlCVX controls how Convex votes in Curve’s Gauge elections.</p><p>Stablecoin issuers figured out the arbitrage: instead of accumulating veCRV directly, just bribe vlCVX holders. “Vote for our pool. Here are tokens.”</p><p>Votium formalized this into an open marketplace. The full loop:</p><blockquote><em>CRV emissions → attract LPs → lock CRV → veCRV → Convex aggregates → vlCVX → Gauge control → stablecoin issuers buy votes on Votium → CVX stays valuable</em></blockquote><p>The Curve Wars weren’t really about liquidity. They were about voting power.</p><p>⑦ The paradox: decentralization created centralization</p><p>The ve model was designed to distribute governance across committed long-term holders. In practice, locking was inconvenient enough that a middleman emerged. That middleman absorbed 40%+ of voting power into a single entity.</p><p>The protocols that copied the ve model — Pendle, PancakeSwap, Balancer — all eventually abandoned it. The mechanism wasn’t the problem. What they failed to replicate was the underlying condition that made it work.</p><p>Curve survived because stablecoin issuers had structural, existential demand for Curve liquidity. As long as Ethena, FRAX, and others need deep stablecoin pools to maintain their pegs, the Gauge vote has real value — and the whole system holds.</p><p>Without that demand, the ve model is just a lock-up scheme with no one fighting over the keys.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=011d1e9fa03d" width="1" height="1" alt=""><hr><p><a href="https://medium.com/blockchain-dyor/curve-finance-how-this-liquidity-war-happened-011d1e9fa03d">[Curve Finance] How this liquidity war happened</a> was originally published in <a href="https://medium.com/blockchain-dyor">Blockchain DYOR(2026)</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Hyperliquid vs GMX — Same Perp DEX. Completely Different Architecture.]]></title>
            <link>https://medium.com/blockchain-dyor/hyperliquid-vs-gmx-same-perp-dex-completely-different-architecture-f1f2c43892e6?source=rss----d2e4ea0bbebc---4</link>
            <guid isPermaLink="false">https://medium.com/p/f1f2c43892e6</guid>
            <category><![CDATA[gmx]]></category>
            <category><![CDATA[hyperliquid]]></category>
            <category><![CDATA[clob]]></category>
            <dc:creator><![CDATA[0x365home]]></dc:creator>
            <pubDate>Tue, 31 Mar 2026 15:29:14 GMT</pubDate>
            <atom:updated>2026-03-31T15:29:14.804Z</atom:updated>
            <content:encoded><![CDATA[<p>Both are on-chain perpetual exchanges. But the mechanics couldn’t be more different. Here’s why that gap exists, what risks each protocol is holding, and what it actually means.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TwBh3RrRbRFM7HbUSmx5Dg.png" /></figure><p><strong>1/ They started from different problems</strong></p><p>Hyperliquid = CLOB. Real orders stack in an order book. Buy and sell get matched 1-to-1. Same mechanism as traditional finance — except everything is recorded on-chain.</p><p>GMX = no order book. The GLP pool automatically becomes your counterparty. Open an ETH long and the pool takes the other side. You win, money leaves the pool. You lose, money enters the pool.</p><p>Same category. Completely different architecture. Because each protocol was trying to solve a different problem first.</p><p><strong>2/ The core problem with CLOB — adverse selection</strong></p><p>Think from the MM’s perspective.</p><p>I’m providing liquidity on PEPE. Bid 100, ask 102 posted. Someone comes in with a massive buy and sweeps my asks. Right after — PEPE pumps 20%.</p><p>I sold cheap. They had information I didn’t.</p><p>That’s adverse selection. MMs want to provide liquidity only to uninformed traders — but informed traders use the same order book. The thinner the market, the worse this gets. So MMs don’t show up. Liquidity stays thin.</p><p><strong>3/ How GMX solved it</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*G_FqlVHZDBHUiRIWx1mRCQ.png" /></figure><p>Reference oracle prices. Execution always happens near fair value.</p><p>MMs don’t get picked off as easily. Adverse selection risk drops.</p><p>But there’s a tradeoff. If the oracle is slow — real market price has already moved but oracle hasn’t updated yet — informed traders exploit that gap. Oracle latency attack.</p><p>And from the GLP depositor’s side: if traders are collectively profitable, GLP loses even after collecting fees. Token incentives have to compensate — cut those incentives and liquidity leaves. Mercenary liquidity.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oyfKwKBAi7HgdP9c1QHJWw.png" /></figure><p><strong>4/ What Hyperliquid chose instead</strong></p><p>Accurate price discovery over adverse selection protection.</p><p>CLOB means the market finds price itself. No oracle latency risk. But when market price diverges from oracle price, basis risk emerges. Arbitrageurs close the gap — and HLP can get caught in that process.</p><p>That’s why liquidation triggers use mark price — a blend of CLOB market price and oracle. Prevents whales from manipulating the order book to force liquidations.</p><p><strong>5/ Bootstrapping — where they really diverge</strong></p><p>GMX flooded GLP depositors with token incentives. TVL stacked fast. But the structural weakness followed: cut the incentives, liquidity exits. Mercenary.</p><p>Hyperliquid was different. Jeff Yan is an ex-Alameda trader. Raised zero VC. Used his own capital to run the MM himself for two years. Distributed points to attract early traders. Volume came in, fees generated, HLP yield appeared, external liquidity followed, bigger traders arrived.</p><p>HLP yield comes from trading fees. Not token emissions. That’s the difference — and that’s why it’s sustainable.</p><p><strong>6/ Same problems. Different tradeoffs.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AACyrAFMz3UMQWGLHra-6g.png" /></figure><p>Adverse selection, slippage, bootstrapping — every perp DEX faces these. What differs is what each protocol chose to sacrifice.</p><p>GMX → reduced adverse selection, eliminated slippage. Took on oracle risk and mercenary liquidity. Hyperliquid → kept accurate price discovery, built sustainable liquidity. Took on basis risk and a harder bootstrapping problem.</p><p>There is no right answer. Only tradeoffs.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f1f2c43892e6" width="1" height="1" alt=""><hr><p><a href="https://medium.com/blockchain-dyor/hyperliquid-vs-gmx-same-perp-dex-completely-different-architecture-f1f2c43892e6">Hyperliquid vs GMX — Same Perp DEX. Completely Different Architecture.</a> was originally published in <a href="https://medium.com/blockchain-dyor">Blockchain DYOR(2026)</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Card vs Blockchain in Payment]]></title>
            <link>https://medium.com/blockchain-dyor/card-vs-blockchain-in-payment-e114cac86c5e?source=rss----d2e4ea0bbebc---4</link>
            <guid isPermaLink="false">https://medium.com/p/e114cac86c5e</guid>
            <category><![CDATA[cards]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[payfi]]></category>
            <dc:creator><![CDATA[0x365home]]></dc:creator>
            <pubDate>Wed, 25 Mar 2026 15:56:42 GMT</pubDate>
            <atom:updated>2026-03-25T15:56:42.658Z</atom:updated>
            <content:encoded><![CDATA[<p><em>The card payment system took 60 years to build. Here’s exactly which parts stablecoins replace — and which they can’t.</em></p><p>If you’ve read my previous piece on how card payments actually work, you know the system involves five separate institutions coordinating behind every transaction. Each one exists to solve a specific trust problem. Each one takes a cut.</p><p>Stablecoins don’t improve that system. They route around it entirely.</p><p>Here’s what that means in practice.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8KD62CnCR9lPHPW70WnUew.png" /></figure><p><strong>Quick recap: why card payments need five middlemen</strong></p><p>When you pay with a card, two strangers are exchanging value. The buyer doesn’t know if the merchant is legitimate. The merchant doesn’t know if the buyer can actually pay. Neither party’s bank knows if the other is acting in good faith.</p><p>The five-layer system exists to solve exactly that:</p><pre>Issuing bank      →  &quot;This buyer can pay — I&#39;ll front the money&quot;<br>Card network      →  &quot;I&#39;ll connect your bank to their bank, globally&quot;<br>Acquiring bank    →  &quot;This merchant is legitimate — I&#39;ll vouch for them&quot;<br>Payment processor →  &quot;Small merchants can&#39;t access this directly — I&#39;ll aggregate them&quot;<br>Merchant          →  &quot;I&#39;ll absorb all the fees to keep the system running&quot;</pre><p>Each layer takes on risk. Each layer earns a fee. The total comes to roughly <strong>1.9% per transaction</strong>, and your money takes <strong>1–2 business days</strong> to actually move.</p><p>This system works. It’s just expensive, slow, and built on institutional trust.</p><h4><strong>What stablecoins actually are</strong></h4><p>Bitcoin proved you could move value without a bank. But Bitcoin’s price swings wildly — worthless as a payment method if the “coffee” you bought this morning cost 0.000003 BTC and tomorrow it costs 0.000005 BTC.</p><p>Stablecoins solve this. One USDC = one dollar. Always.</p><p>The mechanism: Circle (the issuer) holds $1 in a bank account for every 1 USDC in circulation. When you want your dollar back, you burn the USDC and receive $1. The peg holds because it’s fully backed.</p><p>The result: <strong>a dollar that lives on a blockchain.</strong> No bank account required to hold it. No bank account required to send it. Just a wallet address.</p><h4><strong>How blockchain replaces each layer — technically</strong></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yanfXLko5mSh8GPC_VR8kA.png" /></figure><p>This is where it gets interesting. Blockchain doesn’t just “do payments differently.” It replaces specific functions that specific institutions used to perform.</p><p><strong>Issuing bank → On-chain balance</strong></p><p>The issuing bank’s core job is answering one question: <em>can this person pay?</em> They do it by querying credit databases and making an approval decision.</p><p>Blockchain answers the same question differently. Your wallet balance is public and permanently readable on-chain. No institution needs to vouch for you — the code checks your balance automatically before the transaction executes. If you don’t have the funds, the transaction fails. No human involved.</p><p>This works via two models:</p><ul><li><strong>UTXO model</strong> (Bitcoin): your “balance” is a collection of unspent transaction outputs — essentially digital change from previous transactions</li><li><strong>Account model</strong> (Ethereum, Solana): a single balance number that increments and decrements, like a bank account</li></ul><p>USDC runs on the Account model, which is why smart contracts can interact with it so easily.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NoqGgJ5YXdD646AbnX6vOQ.png" /></figure><p><strong>Card network → P2P network</strong></p><p>Visa’s job is connecting your bank to the merchant’s bank. They do it via ISO 8583, a messaging protocol that’s been around since 1987. It works, but it requires both parties to be members of the Visa network.</p><p>Blockchain replaces this with a peer-to-peer network. Any wallet can send to any other wallet — no membership required, no geographic restriction. A wallet in Seoul can send USDC to a wallet in Lagos in the same transaction it would take to send to someone next door.</p><p>The transaction broadcasts to the network, propagates through nodes, and gets picked up by validators for processing.</p><p><strong>Acquiring bank → On-chain finality</strong></p><p>The acquiring bank’s job is executing settlement — taking money from the issuing bank side and depositing it into the merchant’s account. This runs in batches, which is why settlement takes 1–2 days.</p><p>Blockchain replaces batch settlement with <strong>finality</strong>. Once a transaction is included in a confirmed block, it’s settled. Permanently. Irreversibly.</p><p>The speed depends on the chain:</p><ul><li>Solana: ~0.4 seconds</li><li>Ethereum: ~12 seconds</li><li>Bitcoin: ~10 minutes</li></ul><h4><strong>The fee math</strong></h4><p>The fees look different too.</p><p>Card processing collects fees from each institutional layer:</p><pre>Issuing bank        $1.50  (1.5%)  ← compensation for credit risk<br>Card network        $0.10  (0.1%)  ← infrastructure maintenance<br>Payment processor   $0.30  (0.3%)  ← processing + security<br>─────────────────────────────────<br>Total fee on $100   $1.90  (1.9%)</pre><p>Stablecoin fees are just network validation costs:</p><pre>Solana gas fee     ~$0.00025 per transaction<br>Base gas fee       ~$0.001 per transaction<br>Ethereum gas fee   $0.50–5.00 (varies with congestion)</pre><p>On Solana, that’s <strong>0.00025%</strong> — roughly 7,600x cheaper than card processing.</p><p>This is why MPP (Stripe’s Machine Payment Protocol) can support transactions as small as <strong>0.01 USDC</strong>. At card rates, the fee would exceed the transaction value. On Solana, it’s negligible.</p><h4><strong>What actually happens when you pay in USDC</strong></h4><p>Here’s the full on-chain sequence:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IbVjuwx_Tk4u9J-v9x_Tvg.png" /></figure><p><strong>① Customer initiates</strong> — holds USDC in their wallet, initiates a payment</p><p><strong>② Sign the transaction</strong> — the customer’s private key signs the transaction cryptographically. This is the “I authorized this” proof. Without the private key, the transaction can’t be signed. Without a valid signature, it can’t be processed.</p><p><strong>③ Broadcast</strong> — the signed transaction is broadcast to the network. Every connected node receives it.</p><p><strong>④ Mempool</strong> — the transaction sits in the mempool (a waiting queue of unconfirmed transactions). Higher gas fees get prioritized. On Solana, this step is nearly instantaneous.</p><p><strong>⑤ Validator picks it up</strong> — a validator checks two things: is the signature valid? Does the sender have sufficient balance? If yes to both, the transaction gets included in the next block.</p><p><strong>⑥ Finality</strong> — block is confirmed. 100 USDC moves from the customer’s wallet to the merchant’s wallet. The USDC smart contract records the transfer. It’s done.</p><p>The entire process: <strong>~0.4 seconds on Solana.</strong></p><p>Compare that to the card equivalent: authorization in 300ms, actual money movement 1–2 business days later.</p><h4><strong>So why haven’t stablecoins replaced cards already?</strong></h4><p>Three structural problems remain unsolved.</p><p><strong>1. Irreversibility</strong></p><p>Every advantage of blockchain finality — instant, permanent, trustless — is also its biggest consumer protection problem.</p><p>Card payments have chargebacks. If a merchant commits fraud, if you’re double-charged, if a product never arrives — you can dispute the charge and get your money back. The issuing bank eats the loss and fights the merchant.</p><p>Blockchain has no equivalent. Once a transaction is confirmed, it cannot be reversed. There is no dispute resolution layer. If you send funds to the wrong address, they’re gone. If a merchant scams you, there’s no one to call.</p><p>For consumer-facing commerce, this is a dealbreaker at scale.</p><p><strong>2. UX complexity</strong></p><p>Using a card requires: a 16-digit number, an expiry date, a CVV. Most people have had this workflow memorized since they were teenagers.</p><p>Using a crypto wallet requires: downloading a wallet app, generating a seed phrase (12–24 words that you must never lose, never share, never store digitally), understanding gas fees, selecting the right network, and not making a typo in a 42-character address.</p><p>The cognitive overhead is enormous. And the penalty for mistakes — wrong address, wrong network — is permanent loss of funds. No undo.</p><p><strong>3. Regulatory fragmentation</strong></p><p>Card payments operate under decades of consumer protection law, international standards, and regulatory clarity. Merchants know exactly what they’re signing up for.</p><p>Stablecoins don’t have that yet:</p><pre>US          → SEC/CFTC jurisdiction debate ongoing<br>EU          → MiCA framework live since 2024<br>South Korea → Virtual Asset Act in force<br>China       → Banned entirely</pre><p>A merchant building on stablecoins today is building on regulatory ground that could shift under them. That uncertainty slows adoption.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e114cac86c5e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/blockchain-dyor/card-vs-blockchain-in-payment-e114cac86c5e">Card vs Blockchain in Payment</a> was originally published in <a href="https://medium.com/blockchain-dyor">Blockchain DYOR(2026)</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Does Paying With a Card Involve So Many Middlemen?]]></title>
            <link>https://medium.com/blockchain-dyor/why-does-paying-with-a-card-involve-so-many-middlemen-ac6a9976e8fe?source=rss----d2e4ea0bbebc---4</link>
            <guid isPermaLink="false">https://medium.com/p/ac6a9976e8fe</guid>
            <category><![CDATA[stripe]]></category>
            <category><![CDATA[payments]]></category>
            <dc:creator><![CDATA[0x365home]]></dc:creator>
            <pubDate>Tue, 24 Mar 2026 15:17:22 GMT</pubDate>
            <atom:updated>2026-03-24T15:17:22.877Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Every “Payment Successful” notification hides a surprisingly complex chain of trust</em></p><p>When you tap your card to pay $100 at a store, it feels instant. But behind that 0.3-second moment, four separate institutions are coordinating — each one essential, each one taking a cut.</p><p>This isn’t inefficiency. It’s a trust system built over decades.</p><p>Here’s how it actually works.</p><p><strong>The 5 Players</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eALNpXtVWFN1yOx6meVkSA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YiFtB2i4__D8_2aGaq-Zrg.png" /></figure><p><strong>1. The Issuing Bank — Where Trust Begins</strong></p><p>This is the bank that gave you your card. Chase, Bank of America, Citi.</p><p>Here’s what most people don’t realize: when you pay $100 with a credit card, <em>your money doesn’t move immediately.</em> The bank pays the merchant on your behalf first. You settle up later.</p><p>That’s the entire premise of credit.</p><p>Without an issuing bank: → No credit cards. Only prepaid or debit. → The concept of “buy now, pay later” simply doesn’t exist.</p><p>The risks they absorb:</p><ul><li>If you don’t pay, they take the loss</li><li>If your card gets stolen and used fraudulently, they’re on the hook</li><li>Every approval decision is a financial bet</li></ul><p>This is why they take the biggest cut — <strong>~$1.50 on a $100 transaction (1.5%).</strong> That’s over 80% of all fees collected. More risk, more reward.</p><p><strong>2. The Card Network — The Internet of Money</strong></p><p>Visa. Mastercard. American Express.</p><p>Most people call these “credit card companies.” They’re not. They don’t issue cards, they don’t lend money, and they don’t hold your funds. They’re closer to a <strong>protocol</strong> — a shared set of rules and infrastructure that connects every bank to every other bank.</p><p>Without a card network: → Your Chase card only works at merchants that bank with Chase → International payments become nearly impossible → Every bank needs its own proprietary payment system</p><p>The reason your Visa card works at 100 million merchants in 200 countries is because everyone agreed to speak the same language.</p><p>Their risk exposure: almost none. They’re just routing the message.</p><p>And yet — on that same $100 transaction, Visa takes <strong>~$0.10 (0.1%).</strong> Tiny per transaction. But multiply that by trillions of dollars in global payment volume and you get Visa’s <strong>operating margin of over 50%.</strong></p><p><strong>3. The Acquiring Bank — The Merchant’s Gatekeeper</strong></p><p>Merchants can’t connect to Visa or Mastercard directly. They need a bank on their side of the transaction — the acquiring bank.</p><p>The acquiring bank:</p><ul><li>Opens a merchant account so the business can receive card payments</li><li>Connects the merchant to the card networks</li><li>Executes settlement, transferring funds into the merchant’s account</li></ul><p>The flow at settlement looks like this:</p><pre>Issuing Bank → Card Network → Acquiring Bank → Merchant</pre><p>Without an acquiring bank: → Merchants literally cannot accept card payments → Every business would need to negotiate directly with each card network</p><p>The risks they absorb:</p><ul><li>If a merchant commits fraud, the acquiring bank is liable</li><li>If a merchant goes bankrupt mid-transaction, chargebacks land on them</li><li>This is why merchant underwriting is thorough and slow — getting a direct merchant account can take months</li></ul><p>Acquiring bank fees are typically bundled inside the payment processor’s fee.</p><p><strong>4. The Payment Processor — The Only Door for Small Businesses</strong></p><p>Even with an acquiring bank in the picture, most small businesses can’t get a direct merchant account. Requirements typically include:</p><ul><li>Months of application review</li><li>Extensive documentation</li><li>Security deposits</li><li>Minimum revenue thresholds</li></ul><p>For a startup or indie shop, that’s a non-starter.</p><p>Payment processors solve this by aggregating thousands of small merchants under one acquiring bank relationship. They also handle the technical layer:</p><ul><li>Encrypting and transmitting card data</li><li>Maintaining PCI DSS compliance (the international security standard for handling card data)</li><li>Providing checkout UI</li></ul><p>Their risks:</p><ul><li>Responsible for vetting every merchant they onboard</li><li>Liable for security breaches</li><li>On the hook for system outages</li></ul><p>Cut on $100: <strong>~$0.30 (0.3%)</strong></p><p><strong>5. The Merchant — The One Who Pays for All of This</strong></p><p>The merchant provides the goods or services — but they’re also the ones absorbing the full cost of the fee stack.</p><p>Every fee from every layer above comes out of the merchant’s pocket, not the customer’s. Which leads to an important question.</p><p><strong>What $100 Actually Looks Like After Fees</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rpgcZ1q2t3fk5SWVIAtu4Q.png" /></figure><pre>Customer pays          $100.00<br>─────────────────────────────────<br>Issuing Bank     -$1.50   (1.5%)   ← credit risk<br>Card Network     -$0.10   (0.1%)   ← infrastructure<br>Payment Processor -$0.30  (0.3%)   ← processing + security<br>─────────────────────────────────<br>Merchant receives  $98.10  (98.1%)</pre><p><strong>Why Merchants Complain About “Just 1.5%”</strong></p><p>Because that 1.5% is taken from <em>revenue</em>, not <em>profit</em>.</p><p>Say a convenience store runs on a 25% margin:</p><pre>Revenue           $100.00<br>Cost of goods      -$75.00<br>─────────────────<br>Gross profit       $25.00   (25%)</pre><pre>Card fee (1.5%)    -$1.50<br>─────────────────<br>What&#39;s left        $23.50</pre><p>$1.50 out of $25.00 profit = <strong>6% of earnings gone.</strong></p><p>For a business running on 10% margins, that same 1.5% fee wipes out <strong>15% of their profit.</strong></p><p><strong>Here’s the Part Nobody Talks About</strong></p><p>That fee? <strong>It’s already in the price you pay.</strong></p><p>Merchants price their products knowing card fees exist. The $4 coffee at a café — the card processing cost is baked in. Every customer pays it, whether they use cash or card.</p><p>Which creates a quiet irony:</p><p><strong>Cash customers subsidize credit card rewards.</strong></p><p>Cash payers absorb the fee-inclusive price but get none of the benefits. Card users get the same price <em>plus</em> cashback, points, and miles.</p><p><strong>Why Credit Card Rewards Exist (The Real Reason)</strong></p><p>The issuing bank collects that 1.5% interchange fee.</p><p>Then they give a portion back to cardholders as rewards — cashback, airline miles, hotel points.</p><p>Why? Because rewards encourage you to spend more on the card, which generates more interchange fees.</p><p>The full loop:</p><pre>Merchant pays fee → Bank collects interchange<br>→ Bank returns portion as rewards<br>→ Cardholder spends more<br>→ Merchant pays more fees<br>→ repeat</pre><p>It’s funded entirely by the merchant, flowing through the bank, landing in your rewards balance.</p><p>Premium cards offer better rewards because they charge higher interchange rates — the math works out the same way.</p><p><strong>So Why Can’t We Just Remove the Middlemen?</strong></p><p>Because each one solves a real problem.</p><p>At its core, a payment is two strangers agreeing to exchange value. The buyer doesn’t know if the merchant is legitimate. The merchant doesn’t know if the buyer can actually pay. The banks don’t know if either party is acting in good faith.</p><p>The four-layer system is trust, institutionalized:</p><pre>Issuing Bank       → guarantees the buyer can pay<br>Card Network       → connects any bank to any bank, globally<br>Acquiring Bank     → vouches for the merchant<br>Payment Processor  → makes it accessible to everyone</pre><p>Each layer takes on risk. Each layer earns a fee in return.</p><p>This system took decades to build. That’s also why it’s hard to disrupt.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ac6a9976e8fe" width="1" height="1" alt=""><hr><p><a href="https://medium.com/blockchain-dyor/why-does-paying-with-a-card-involve-so-many-middlemen-ac6a9976e8fe">Why Does Paying With a Card Involve So Many Middlemen?</a> was originally published in <a href="https://medium.com/blockchain-dyor">Blockchain DYOR(2026)</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is an Oracle — how blockchains see the outside world]]></title>
            <link>https://medium.com/blockchain-dyor/what-is-an-oracle-how-blockchains-see-the-outside-world-cdbf96201e00?source=rss----d2e4ea0bbebc---4</link>
            <guid isPermaLink="false">https://medium.com/p/cdbf96201e00</guid>
            <category><![CDATA[oracle]]></category>
            <category><![CDATA[pyth]]></category>
            <category><![CDATA[chain-link]]></category>
            <dc:creator><![CDATA[0x365home]]></dc:creator>
            <pubDate>Fri, 20 Mar 2026 19:39:03 GMT</pubDate>
            <atom:updated>2026-03-20T19:39:03.387Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>1. Why they exist</strong></p><p>A smart contract can only read data that lives on the blockchain. It cannot fetch external data directly.</p><p>Aave wants to lend USDC against ETH collateral. To calculate the correct collateral ratio, it needs the current ETH price. But the Ethereum blockchain itself has no idea what ETH costs.</p><pre>The problem:<br>→ Hardcoding a price means it can never update<br>→ Using a single exchange API creates a single point of failure<br>→ Computing price on-chain from a single source is trivially manipulable</pre><p>An oracle is infrastructure that brings verifiable external data onto the blockchain.</p><p><strong>2. Chainlink — decentralized aggregation</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SXGXV1VhDK4pyxT65GxcPA.png" /></figure><p>Multiple independent node operators each collect price data from different sources. The results are aggregated on-chain using a median.</p><pre>Node 01: $65,012 (Binance + Coinbase)<br>Node 02: $65,008 (Kraken + OKX)<br>Node 03: $65,015 (Bybit + Bitfinex)<br>Node 04: $71,000 ← outlier detected, removed<br>Node 05: $65,010 (Huobi + Gate)</pre><pre>→ On-chain median: $65,010</pre><p>The OCR (Off-Chain Reporting) protocol lets nodes reach consensus off-chain first. One reporter node then submits a single transaction on behalf of all. This cut gas costs by ~90% compared to v1 where every node submitted individually.</p><p>Two trigger conditions exist for a price update:</p><pre>Deviation threshold: price moves more than X% from last reported value<br>Heartbeat:          maximum time between updates (e.g. 1 hour)<br>                    regardless of price movement</pre><p>If neither condition is met, the price doesn’t update. This is why staleness checks matter — a protocol that doesn’t verify when the price was last updated can operate on hours-old data without knowing it.</p><p><strong>3. Pyth — 1st party data and the pull model</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RT3HgbNjZPH6F4czEyQzpg.png" /></figure><p>Pyth’s architecture is fundamentally different from Chainlink’s.</p><p><strong>1st party data</strong>: Actual market makers — Jump Trading, Virtu, Jane Street — submit prices directly. No 3rd party node in the middle. The entity actually trading submits the price, which produces more accurate mid-price data than aggregating API feeds.</p><p><strong>Pull model</strong>: Chainlink pushes periodic updates to every supported chain. Pyth does the opposite. When a user submits a transaction, they attach the latest price proof (VAA) themselves. No automatic on-chain updates means no wasted gas between uses.</p><p><strong>Confidence interval</strong>: This is Pyth’s core differentiator. Instead of a single price value, Pyth publishes price ± σ. When the market is calm, the interval is tight. When volatility spikes, it widens.</p><pre>Calm market:<br>$65,010 ± $50  → narrow band → high confidence</pre><pre>Volatile market:<br>$65,010 ± $800 → wide band → low confidence</pre><p>Protocols can use this to implement logic like “pause liquidations when the confidence interval is too wide.” Chainlink’s single-value feed doesn’t offer this kind of nuance.</p><p><strong>4. Chainlink vs Pyth</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*OLSt1_iu0-JXR_2aXbvP4w.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bYQHSCwvb17JtPB7tsW6jQ.png" /></figure><pre>Category            Chainlink              Pyth<br>Data source         3rd party nodes        1st party (actual MMs)<br>Update frequency    ~1 hour heartbeat      ~400ms continuous<br>Gas model           Oracle pushes updates  User attaches to TX<br>Decentralization    High (~30 operators)   Medium (few providers)<br>Confidence interval No                     Yes<br>Primary use cases   Aave, MakerDAO         Drift, Jupiter<br>Security risk       Node collusion         Wormhole bridge</pre><p>Pyth is significantly faster but routes through Wormhole to reach non-Solana chains. Wormhole was exploited for $320M in 2022. That dependency is a real, non-trivial risk.</p><p><strong>5. Oracle manipulation attacks</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*b8NzDsSbmZn657AvtxLTpw.png" /></figure><p>A significant portion of DeFi hacks trace back to oracle exploitation.</p><p><strong>Flash loan + oracle manipulation:</strong></p><pre>1. Borrow large amount via flash loan (no collateral required)<br>2. Execute a large buy on a single DEX<br>   → Price spikes on that DEX<br>   → If oracle sources from this DEX, on-chain price spikes too<br>3. Take out an over-collateralized loan at the inflated price<br>4. Extract borrowed assets, unwind position<br>5. Repay flash loan<br>→ Mango Markets: $117M (2022)</pre><p><strong>Staleness vulnerability:</strong></p><pre>Network congestion delays oracle update<br>→ Spot price moves significantly<br>→ On-chain price is stale<br>→ Liquidations don&#39;t trigger<br>→ Protocol holds undercollateralized positions without knowing</pre><p><strong>Defenses:</strong></p><pre>→ Multi-source aggregation (never rely on a single DEX)<br>→ Always verify when the price was last updated<br>→ Circuit breaker (pause protocol on extreme price movement)<br>→ TWAP (time-weighted average price to resist flash manipulation)</pre><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cdbf96201e00" width="1" height="1" alt=""><hr><p><a href="https://medium.com/blockchain-dyor/what-is-an-oracle-how-blockchains-see-the-outside-world-cdbf96201e00">What is an Oracle — how blockchains see the outside world</a> was originally published in <a href="https://medium.com/blockchain-dyor">Blockchain DYOR(2026)</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[DEX On-Chain Orderbook Implementation]]></title>
            <link>https://medium.com/blockchain-dyor/dex-on-chain-orderbook-implementation-d805347a924d?source=rss----d2e4ea0bbebc---4</link>
            <guid isPermaLink="false">https://medium.com/p/d805347a924d</guid>
            <category><![CDATA[order-book]]></category>
            <category><![CDATA[onchain]]></category>
            <category><![CDATA[dex]]></category>
            <category><![CDATA[cex]]></category>
            <dc:creator><![CDATA[0x365home]]></dc:creator>
            <pubDate>Thu, 19 Mar 2026 18:02:35 GMT</pubDate>
            <atom:updated>2026-03-19T18:02:35.053Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RMGWrGuJJ0fHEXxG4bjZqw.png" /></figure><p><strong>Why it’s hard</strong></p><p>The fundamental problem with an on-chain orderbook is that every state change is a transaction, and every transaction costs gas.</p><pre>CEX orderbook update:<br>→ In-memory pointer change<br>→ Cost: ~0, Latency: nanoseconds</pre><pre>Ethereum on-chain orderbook update:<br>→ SSTORE operation per storage slot<br>→ Cost: 20,000 gas/slot = ~$0.5~$5<br>→ Latency: 12 seconds (block time)</pre><p>Running a fully on-chain orderbook on Ethereum L1 is economically impossible. Every cancel, modify, and partial fill is a transaction. A market maker running a tight spread would burn thousands of dollars in gas per hour.</p><p><strong>Approach 1: Off-chain orderbook + on-chain settlement</strong></p><p>dYdX v3 and early Serum used this model.</p><pre>Order intake/matching: centralized server (off-chain)<br>Settlement/custody:    smart contract (on-chain)</pre><pre>Pros: CEX-level speed, low cost<br>Cons: orderbook operator is centralized<br>      censorship possible<br>      full outage if server goes down<br>      not truly decentralized</pre><p>The custody is non-custodial. The matching is not. This is the key distinction. Users own their assets, but the execution layer is a trusted third party.</p><p><strong>Approach 2: AMM (Uniswap, Curve)</strong></p><p>Eliminate the orderbook entirely. Price is determined by a mathematical formula.</p><pre>x * y = k  (Uniswap v2 CFMM)</pre><pre>→ No orderbook required<br>→ Entire state is just two reserve values<br>→ Swap completes in a single transaction<br>→ No market makers needed (LPs replace them)</pre><p>The tradeoff is price efficiency. AMMs are path-dependent — a large order moves the price along the bonding curve, creating slippage proportional to order size relative to pool depth. This makes AMMs unsuitable for perpetuals, where precision execution matters.</p><p>Uniswap v3 introduced concentrated liquidity to address this:</p><pre>v2: liquidity distributed uniformly across [0, ∞)<br>v3: LPs specify a price range [Pa, Pb]<br>    → capital efficiency improves dramatically<br>    → approximates a discrete orderbook<br>    → but LP position management becomes complex</pre><p><strong>Approach 3: App-chain / L1 optimization (Hyperliquid)</strong></p><p>Build a purpose-specific L1 where the cost model is designed around orderbook operations.</p><pre>Core idea:<br>→ Block time reduced to ~0.2 seconds<br>→ Gas structure designed for orderbook operations<br>→ Order cancel/modify cost near zero</pre><pre>Implementation:<br>→ Orderbook maintained as chain state<br>→ Validators reach consensus on orderbook state<br>→ All orders and fills recorded on-chain<br>→ Full auditability</pre><p><strong>HyperBFT internals</strong></p><p>HyperBFT is a modified Tendermint BFT consensus.</p><pre>Round structure:<br>1. Leader (proposer) broadcasts block proposal<br>2. Validators send PREVOTE messages<br>3. On 2/3+ PREVOTE → validators send PRECOMMIT<br>4. On 2/3+ PRECOMMIT → block finalized</pre><pre>Properties:<br>→ Single-slot finality (no reorganizations)<br>→ Deterministic leader rotation<br>→ Liveness requires &gt;2/3 honest validators<br>→ Safety holds as long as &lt;1/3 Byzantine</pre><p>Single-slot finality matters for orderbook execution. On Ethereum, a transaction has probabilistic finality — the chain can reorganize. This makes limit order placement risky because a fill might be reverted. HyperBFT’s instant finality removes this ambiguity.</p><p><strong>On-chain orderbook MEV exposure</strong></p><p>On-chain orderbooks are more MEV-exposed than CEXs.</p><pre>CEX:<br>→ Matching engine is centralized<br>→ Front-running requires latency advantage only<br>→ No mempool visibility</pre><pre>On-chain orderbook:<br>→ Transactions broadcast to mempool before inclusion<br>→ Large market orders visible pre-execution<br>→ Block proposer controls transaction ordering<br>→ Front-running is structurally easier</pre><p>The attack surface is different from AMM sandwich attacks. On an orderbook DEX, the attacker can:</p><pre>1. Detect large market buy in mempool<br>2. Place limit sell just above current ask<br>3. Cancel after market order fills<br>→ Effectively &quot;jumps the queue&quot; at the cost of<br>  cancel transaction gas</pre><p>Hyperliquid’s mitigations:</p><pre>Fast block time (~0.2s):<br>→ Reduces window for MEV bots to react<br>→ Smaller time delta between mempool observation<br>  and block inclusion</pre><pre>Validator set control:<br>→ Small validator set currently<br>→ Validators are known entities<br>→ Reduces anonymous MEV extraction<br>→ But this is also the centralization tradeoff</pre><p><strong>Storage optimization — order state encoding</strong></p><p>On EVM-based chains, storage cost dominates. Efficient order encoding is critical.</p><pre>Naive implementation:<br>struct Order {<br>  address trader;    // 20 bytes → 1 storage slot<br>  uint256 price;     // 32 bytes → 1 storage slot<br>  uint256 quantity;  // 32 bytes → 1 storage slot<br>  uint256 timestamp; // 32 bytes → 1 storage slot<br>  bool isBuy;        // 1 byte  → 1 storage slot (padded)<br>}<br>→ 5 storage slots = 5 × 20,000 gas = 100,000 gas per order</pre><pre>Packed implementation:<br>uint256 packed = (uint256(trader) &lt;&lt; 96)<br>               | (price &lt;&lt; 64)     // assuming price fits in 32 bits<br>               | (quantity &lt;&lt; 32)<br>               | (timestamp &lt;&lt; 1)<br>               | (isBuy ? 1 : 0)<br>→ 1 storage slot = 20,000 gas per order<br>→ 5x cost reduction</pre><p>Hyperliquid sidesteps this entirely. As a non-EVM chain, it’s not bound to the EVM storage model. The orderbook can be maintained as an in-memory data structure on validator nodes, with only event logs committed to blocks — similar to how a CEX persists trade logs to disk asynchronously while keeping the live orderbook in RAM.</p><p><strong>The settlement layer problem</strong></p><p>Even with an optimized L1, cross-asset settlement introduces friction.</p><pre>CEX:<br>→ All assets exist as database entries<br>→ Settlement is a single atomic DB update<br>→ No bridge latency</pre><pre>On-chain:<br>→ USDC collateral deposited from Ethereum/Arbitrum<br>→ Bridge latency: minutes to hours<br>→ Bridge security risk<br>→ Capital fragmentation across chains</pre><p>This is why Hyperliquid’s TVL growth is closely tied to bridge activity. Onboarding friction directly caps addressable liquidity.</p><p><strong>Current state and trajectory</strong></p><pre>Fully on-chain orderbook on Ethereum L1: not viable<br>Off-chain orderbook + on-chain settlement: viable, but centralized matching<br>AMM: viable, but not suitable for perps<br>Purpose-built L1: viable, CEX-level throughput, full on-chain transparency<br>  → Tradeoff: self-sovereign security, small validator set</pre><p>The design space is converging toward purpose-built execution environments — either L2s with fast finality (like StarkEx for dYdX v4 predecessor) or dedicated app-chains. The fundamental constraint is that general-purpose blockchains optimize for decentralization and security at the cost of throughput, which is the wrong tradeoff for an orderbook.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d805347a924d" width="1" height="1" alt=""><hr><p><a href="https://medium.com/blockchain-dyor/dex-on-chain-orderbook-implementation-d805347a924d">DEX On-Chain Orderbook Implementation</a> was originally published in <a href="https://medium.com/blockchain-dyor">Blockchain DYOR(2026)</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Ethereum PoS — How does it actually work?]]></title>
            <link>https://medium.com/blockchain-dyor/ethereum-pos-how-does-it-actually-work-02f3287c75b5?source=rss----d2e4ea0bbebc---4</link>
            <guid isPermaLink="false">https://medium.com/p/02f3287c75b5</guid>
            <category><![CDATA[pow]]></category>
            <category><![CDATA[eth]]></category>
            <category><![CDATA[btc]]></category>
            <category><![CDATA[pos]]></category>
            <category><![CDATA[ethereum]]></category>
            <dc:creator><![CDATA[0x365home]]></dc:creator>
            <pubDate>Wed, 18 Mar 2026 17:15:07 GMT</pubDate>
            <atom:updated>2026-03-18T17:15:07.241Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>1. The essence of PoS</strong></p><p>PoW gives block creation rights to whoever does the most computation. PoS gives it to whoever puts up the most collateral.</p><p>The security model is fundamentally different.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gl8Ow-VPk2Tp1bWCrYjlCQ.png" /></figure><pre>PoW: electricity + hardware → hardware survives a failed attack<br>PoS: ETH collateral        → ETH gets burned on a failed attack</pre><p>Both are designed to make betrayal cost more than it earns. But PoS goes one step further — attempting an attack is an act of destroying your own assets.</p><p><strong>2. How to become a validator</strong></p><pre>Stake 32 ETH<br>→ Generate validator key → join the queue<br>→ Every slot (12 sec), roles are randomly assigned:<br>   · Proposer: propose a block (1 per slot)<br>   · Attester: vote on blocks (everyone else)<br>→ Honest participation → ~4% APR reward<br>→ Malicious behavior → Slashing (collateral burned)</pre><p>Why 32 ETH? Too low and the validator count explodes — network communication overhead becomes unmanageable. Too high and decentralization breaks down. 32 ETH is the balance point.</p><p>There are currently over 1 million active validators.</p><p><strong>3. Gasper — Ethereum’s two engines</strong></p><p>Ethereum PoS combines two mechanisms.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QRhBrRDZShEzml3UazwQhA.png" /></figure><p><strong>LMD-GHOST (fork choice)</strong></p><p>When multiple chains exist simultaneously, which one do you follow? PoW follows the longest chain. Ethereum does it differently.</p><blockquote><em>Follow the chain where validators’ most recent votes have accumulated the most weight.</em></blockquote><p>It’s not about length — it’s about the weight of validator votes.</p><p><strong>Casper FFG (finality)</strong></p><p>LMD-GHOST alone can’t prevent the chain from changing. Casper FFG permanently finalizes specific checkpoints.</p><pre>2/3+ of total staked ETH votes → Justified<br>Two consecutive epochs Justified → Finalized ✅</pre><p>A Finalized block can never be changed. Bitcoin says “6 confirmations = safe” — but theoretically it can still be reversed. Once Ethereum finalizes a block, it’s mathematically impossible to reverse.</p><p><strong>4. Slots and Epochs — Ethereum’s time units</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SWPZVeC3xX3Py8Och9f-WA.png" /></figure><pre>Slot:  12 seconds → one block production opportunity<br>Epoch: 32 slots   → ~6.4 minutes</pre><pre>Last slot of each epoch = checkpoint<br>→ Two consecutive epochs Justified → Finalized<br>→ Mathematical finality in ~13 minutes</pre><p>Bitcoin’s 6 confirmations take ~60 minutes. Ethereum gets there in ~13 minutes — and the guarantee is stronger. Bitcoin’s is probabilistic safety. Ethereum’s is mathematical finality.</p><p><strong>5. PoW vs PoS — core differences</strong></p><pre>PoW                  PoS<br>Security       Computation          Collateral (ETH)<br>Energy         Extremely high       99.95% reduction<br>Failed attack  Hardware survives    ETH gets burned<br>Finality       Probabilistic (~60m) Mathematical (~13m)<br>Centralization Mining pools         Lido etc.</pre><p>PoS solved the energy problem. But centralization didn’t disappear — it just changed its form.</p><p><strong>6. Liquid Staking — participate without 32 ETH</strong></p><p>32 ETH is roughly $100,000+ at current prices. Most people can’t participate directly. Liquid Staking solves this.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mcTI1CezMqPhHr8sON228A.png" /></figure><pre>Deposit any amount of ETH (from 0.01 ETH)<br>→ Receive stETH (proof of staking)<br>→ Use stETH in DeFi (collateral, LP, etc.)<br>→ Staking rewards accumulate automatically<br>→ Redeem stETH → ETH anytime</pre><p>Main protocols:</p><pre>Lido:        issues stETH, ~30% market share<br>Rocket Pool: issues rETH, focused on decentralization<br>Coinbase:    issues cbETH, easiest access</pre><p>But Lido now controls over 30% of all Ethereum staking. If it crosses 33%, it could theoretically disrupt Casper FFG’s 2/3 consensus — enough to block Finality.</p><p>Liquid Staking lowered the barrier to entry. But it created a new centralization risk in return.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=02f3287c75b5" width="1" height="1" alt=""><hr><p><a href="https://medium.com/blockchain-dyor/ethereum-pos-how-does-it-actually-work-02f3287c75b5">Ethereum PoS — How does it actually work?</a> was originally published in <a href="https://medium.com/blockchain-dyor">Blockchain DYOR(2026)</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>