<?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 Vera Imarhia on Medium]]></title>
        <description><![CDATA[Stories by Vera Imarhia on Medium]]></description>
        <link>https://medium.com/@godwinvera2002?source=rss-5d8360fbcf98------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*7Twb8ih665IElZbOAplxjA.jpeg</url>
            <title>Stories by Vera Imarhia on Medium</title>
            <link>https://medium.com/@godwinvera2002?source=rss-5d8360fbcf98------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 23:44:37 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@godwinvera2002/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[Which Bitcoin Bridge Supports Native BTC Without Requiring Wrapped Assets?]]></title>
            <link>https://medium.com/@godwinvera2002/which-bitcoin-bridge-supports-native-btc-without-requiring-wrapped-assets-62037c4f863d?source=rss-5d8360fbcf98------2</link>
            <guid isPermaLink="false">https://medium.com/p/62037c4f863d</guid>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[bitcoin]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <dc:creator><![CDATA[Vera Imarhia]]></dc:creator>
            <pubDate>Mon, 02 Nov 2026 13:07:50 GMT</pubDate>
            <atom:updated>2026-11-02T13:23:29.418Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JicB9VqLQ24oKnRrGqfniQ.png" /></figure><p><strong>Which Bitcoin Bridge Supports Native BTC Without Requiring Wrapped Assets?</strong></p><p><strong><em>BOB Gateway, Garden Finance and THORChain offer different ways to swap native Bitcoin across blockchains. Here’s how they work, what happens to your BTC and what to consider before choosing one</em></strong><em>.</em></p><p>You don&#39;t have to wrap your Bitcoin before using its value on another blockchain.</p><p>Services like BOB Gateway, Garden Finance and THORChain let you start with native BTC and swap it for assets on other supported networks. However, they don&#39;t all handle these swaps in the same way.</p><p>But to understand which option works well, you need to look beyond the list of supported chains.</p><p>What happens to your BTC after you send it? Who controls it? How do you receive the asset you requested if the swap goes through, and what happens if it doesn&#39;t?</p><p>Let&#39;s start with the basics.</p><p><strong>Native BTC and wrapped BTC are not the same thing</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*G3B4uFktqfcHCQ72S2ZWnA.png" /></figure><p>Native BTC is the actual Bitcoin that exists on the Bitcoin network.</p><p>Ethereum, Base and other blockchains operate independently. A smart contract on Ethereum, for example, cannot directly access the BTC sitting in your Bitcoin wallet.</p><p>One way around this is wrapped Bitcoin.</p><p>Wrapped BTC is a token issued on another blockchain to represent Bitcoin held elsewhere. Some familiar examples are:</p><p>- WBTC on Ethereum.<br>- cbBTC on Ethereum and Base.<br>- tBTC on Ethereum.</p><p>These tokens allow Bitcoin&#39;s value to be used in DeFi applications on other networks. You can lend them, borrow against them, trade them or provide liquidity.</p><p>But you&#39;re holding a token on another blockchain, not the original BTC in your Bitcoin wallet.</p><p>The backing and security arrangements also differ.</p><p><strong>How are WBTC, cbBTC and tBTC backed?</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9gwmveP7TPu_gOwHcAEqPw.png" /></figure><ul><li><strong>WBTC</strong></li></ul><p>WBTC is backed 1:1 by BTC held in custody.</p><p>Under the custody arrangement announced in 2026, BiT Global manages the WBTC vaults and controls two of the three private keys, while BitGo retains the third and provides wallet infrastructure.</p><p>When BTC is deposited through an authorised merchant, <a href="https://wbtc-network.medium.com/bit-global-to-complete-next-phase-of-wbtc-custody-transition-">the deposit is verified and an equivalent amount of WBTC is minted on the supported chain.</a></p><ul><li><strong>cbBTC</strong></li></ul><p>cbBTC is Coinbase&#39;s wrapped Bitcoin.</p><p>Coinbase holds the BTC backing it and issues cbBTC at a 1:1 ratio on supported networks. Its proof-of-reserves page <a href="https://www.coinbase.com/cbBTC/proof-of-reserves">proof-of-reserves page</a> displays the BTC reserves alongside the circulating cbBTC supply.</p><p>The backing is visible, but custody still rests with Coinbase.</p><ul><li><strong>tBTC</strong></li></ul><p>tBTC uses a different model.</p><p>Instead of relying on one corporate custodian, it uses a distributed network of selected Threshold signers to secure the Bitcoin.</p><p><a href="https://docs.threshold.network/tbtc-v2/wallet-generation">According to Threshold,</a> 51 out of 100 selected signers must cooperate to produce a valid signature for a wallet transaction.</p><p>This spreads control across multiple operators, but it doesn&#39;t remove every risk. Smart-contract vulnerabilities, signer coordination and protocol issues are still possible.</p><p><a href="https://www.gobob.xyz/en/blog/bitcoin-holder-guide-to-wrapped-btc">The point is that wrapped BTC isn&#39;t one uniform product. Each token has its own custody and security assumptions.</a></p><p><strong>Can you move BTC cross-chain without using a CEX?</strong></p><p>Yes, you can.</p><p>A centralised exchange (CEX) is one way to move Bitcoin&#39;s value into another asset, but it involves several steps.</p><p>You deposit your BTC, sell it, buy the asset you want and withdraw that asset to your wallet on the destination network.</p><p>During this process, you&#39;re relying on the exchange to hold your funds and process the trade and withdrawal.</p><p>Cross-chain swap services offer another route. You choose the asset you want to receive, send BTC from your Bitcoin wallet and receive the destination asset without going through those exchange steps.</p><p>Three approaches worth comparing are BOB Gateway, Garden Finance and THORChain.</p><p>BOB Gateway: swapping native BTC through a solver.</p><p><a href="https://docs.gobob.xyz/gateway/technical">BOB Gateway</a> lets you swap native BTC for assets on other supported blockchains using quotes from solvers.</p><p>A solver is the party that fulfils your swap. It provides the destination asset in exchange for the BTC you send.</p><p>This is different from a traditional bridge that locks BTC and issues a wrapped token on another network. With Gateway, you choose the asset you actually want to receive.</p><p><strong>How does a swap work?</strong></p><p>Let&#39;s say you want to exchange BTC for USDT on Ethereum.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kaYzC6LOegVX-sOia6rcHg.png" /></figure><p>First, you select your input and destination assets. Gateway then provides a signed quote showing the fees and guaranteed minimum amount you should receive.</p><p>Once you accept the quote, an order is created with a unique Bitcoin deposit address belonging to the solver.</p><p>You send your BTC to that address in one Bitcoin transaction. After the deposit is confirmed, the solver delivers the USDT to your wallet on Ethereum.</p><p>Gateway also has a retry and refund process for cases where the payout cannot be completed. <a href="https://docs.gobob.xyz/gateway/technical">You can read the details in the BOB Gateway technical documentation</a></p><p><strong>Where does your BTC go?</strong></p><p>The solver holds the BTC in its own Bitcoin wallet and keeps USDT in its Ethereum wallet to fund the swaps.</p><p>For BTC-to-asset swaps, the incoming BTC payment is not independently verified by BOB&#39;s Bitcoin light client. The BTC goes to the solver&#39;s wallet, and the route relies on the signed quote, confirmation process and refund mechanism.</p><p>There is another important detail: Gateway doesn&#39;t use the same settlement process in both directions.</p><p>When you send another asset to receive BTC, the process is different. For example, on an ETH-to-BTC route, your ETH is first converted into USDT on Ethereum and held in a contract.</p><p>The solver must then prove that it has paid the required BTC on the Bitcoin network. BOB&#39;s light client checks that Bitcoin payment before the escrowed USDT is released to the solver.</p><p>So the light-client verification applies to the X-to-BTC direction, not the BTC-to-X direction.</p><p>This distinction is important when assessing Gateway&#39;s security. It would be inaccurate to describe both directions as having identical guarantees.</p><p><strong>What makes BOB Gateway useful?</strong></p><p>There are a few practical things to consider.</p><p>You receive a quote before sending BTC, with a guaranteed minimum output and fee breakdown. Destination gas costs are included in the quote, so you don&#39;t need to separately hold the destination chain&#39;s gas token just to receive your asset.</p><p>You also avoid manually wrapping your BTC or going through a centralised exchange.</p><p>However, the solver model introduces its own trust assumptions. You&#39;re relying on the solver to complete the payout, and some destination routes may involve third-party swapping or bridging providers.</p><p>Gateway is worth considering when you want a specific destination asset and prefer to see the expected output before sending your BTC.</p><p><a href="https://docs.gobob.xyz/gateway/supported-routes">Just make sure you check the supported routes and assets</a> before starting a swap. Availability can change.</p><p><strong>Garden Finance: cross-chain swaps using HTLCs</strong></p><p><a href="https://garden.finance/bitcoin-bridge">Garden Finance</a> takes a different approach.</p><p>It uses Hashed Time-Locked Contracts (HTLCs) to coordinate swaps between two blockchains.</p><p>Both sides lock the assets involved in the swap under specific conditions. Neither party should be able to claim the other’s asset without fulfilling those conditions.</p><p><strong>How does it work?</strong></p><p>Imagine you want to swap native BTC for ETH.</p><p>You lock your BTC in an HTLC on the Bitcoin network. At the same time, the solver locks the ETH you&#39;re meant to receive on the destination chain.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rg-MDecYNgEnAXPmoPQuZw.jpeg" /></figure><p>To claim the ETH, you reveal a secret called a preimage.</p><p>That same preimage allows the solver to claim your BTC.</p><p>If either party fails to complete the swap, the timelocks eventually allow the original owners to reclaim their funds.</p><p>This is what makes the swap atomic: <a href="https://garden.finance/faq">the contracts are designed to ensure that both sides complete the exchange or can recover their funds.</a></p><p>The HTLC model means Garden doesn’t need to take custody of both assets in a central vault. Instead, the assets are locked under conditions enforced by the relevant Bitcoin scripts and destination-chain contracts.</p><p>But there are still trade-offs.</p><p>Swaps from Bitcoin have to wait for Bitcoin confirmation, which can affect how long the process takes. Swaps also depend on available solver liquidity.</p><p>Smart-contract vulnerabilities and HTLC implementation risks remain possible, even when contracts have been audited.</p><p>Garden supports native BTC swaps across multiple chains, with different assets available depending on the route.</p><p>Some routes offer ETH or stablecoins, while others include wrapped BTC such as WBTC and cbBTC.</p><p>That&#39;s another reason to check the exact output asset before confirming a swap.</p><p>Garden is worth considering when you prefer conditional settlement through HTLCs and want both sides of a swap protected by contract conditions.</p><p><strong>THORChain: swapping native assets through liquidity pools</strong></p><p><a href="https://docs.thorchain.org/native-cross-chain-swaps">The third option is THORChain..</a></p><p>Unlike BOB Gateway and Garden, THORChain is a cross-chain liquidity protocol. It uses liquidity pools connected by its native token, RUNE, to facilitate swaps between supported blockchains.</p><p>You don&#39;t need to hold RUNE yourself to make a swap.</p><p><strong>How does a BTC-to-ETH swap work?</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hWbSylzhnVCZrdOq2sJnCw.png" /></figure><p>Let&#39;s use BTC to ETH as an example.</p><p>You send native BTC to a THORChain Bitcoin vault.</p><p>The network processes the deposit and swaps your BTC for RUNE through the BTC:RUNE pool.</p><p>It then swaps that RUNE for ETH through the RUNE:ETH pool.</p><p>Finally, THORChain releases native ETH from its Ethereum vault to your wallet.</p><p>So the route looks like this:</p><p>BTC → BTC:RUNE pool → RUNE:ETH pool → ETH</p><p>RUNE does the work in the background, connecting the liquidity pools.</p><p><strong>Who secures the vaults?</strong></p><p>THORChain&#39;s vaults are operated by its node network.</p><p>Nodes work together to observe transactions on external blockchains, reach agreement and authorise outgoing transactions using threshold signatures.</p><p>This means no single node should be able to move the vault funds on its own.</p><p>However, distributed control doesn&#39;t mean the system has no security assumptions.</p><p>Users still rely on the node network to secure the vaults and process transactions correctly. Pool liquidity also affects swap execution and the amount received.</p><p>RUNE is part of the swap mechanism, so its role in the protocol&#39;s liquidity and security design is also relevant.</p><p>THORChain is worth considering when you want to swap between supported native assets through shared liquidity pools.</p><p><a href="https://docs.thorchain.org/technical-documentation/technology/bifrost-tss-and-vaults">Its vault and threshold-signature documentation</a> explains more about how the network secures external assets.</p><p>How do BOB Gateway, Garden Finance and THORChain differ?</p><p>All three accept native BTC as the starting asset on supported routes, but their settlement models are different.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/896/1*gsYnDsq8siNtdc0aJ0dV1g.jpeg" /></figure><p>The comparison shows that accepting native BTC doesn&#39;t mean these protocols offer the same experience or security guarantees.</p><p>BOB Gateway uses a solver to fulfil a quoted order. Garden coordinates both sides through HTLCs, while THORChain relies on liquidity pools and a distributed node network.</p><p>There isn&#39;t one approach that is automatically better for every user.</p><p><strong>Which approach should you consider?</strong></p><p>It depends on what you want to do with your Bitcoin.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/896/1*Q0mKFD8FpA0zCB6P8_LT6g.jpeg" /></figure><p>For example, someone who wants to use BTC as collateral on a lending protocol may have a good reason to choose WBTC, cbBTC or tBTC.</p><p>But someone who simply wants to exchange BTC for ETH or USDC doesn&#39;t necessarily need to acquire a wrapped Bitcoin token first.</p><p>The destination asset and the security model should guide the decision.</p><p><strong>What should you check before sending your BTC?</strong></p><p>Before using any of these services, there are a few things I&#39;d check.</p><ul><li><strong>What exactly will you receive?</strong></li></ul><p>Check the asset ticker and destination network. Receiving ETH on Ethereum is not the same as receiving WBTC on Ethereum.</p><ul><li><strong>Where does your BTC go?</strong></li></ul><p>Is it sent to a solver&#39;s wallet, locked in an HTLC or deposited into a network-operated vault?</p><ul><li><strong>How is the swap secured?</strong></li></ul><p>Does the protocol rely on a solver, a preimage and timelock, or a validator network?</p><ul><li><strong>What happens if the swap fails?</strong></li></ul><p>Check the refund conditions and how long you may have to wait before recovering your funds.</p><ul><li><strong>What are the actual costs?</strong></li></ul><p>Look at the quoted output, network fees, protocol fees, slippage and any additional routing costs. Don&#39;t compare quotes taken at different times or for different swap amounts.</p><p>Finally, check the supported routes before sending anything. A protocol&#39;s available assets and networks can change.</p><p><strong>Final thoughts</strong></p><p>So, which Bitcoin bridge supports native BTC without requiring wrapped assets?</p><p><strong><em>BOB Gateway, Garden Finance and THORChain all offer ways to start with native BTC and access assets on other blockchains without first acquiring wrapped Bitcoin.</em></strong></p><p>But they solve the problem differently.</p><p>BOB Gateway uses a quote-based solver system. Garden Finance uses HTLCs to coordinate the swap between both chains. THORChain uses RUNE-connected liquidity pools and a network of nodes to secure its vaults.</p><p>Personally, I wouldn&#39;t choose one simply because it accepts native BTC. I&#39;d look at where the BTC goes, how the destination payout is secured and what happens when the swap doesn&#39;t complete.</p><p><strong><em>Also, keep in mind that moving BTC’s value to another blockchain doesn’t mean native BTC itself has moved there. You’re exchanging it for an asset available on that network, which could be ETH, a stablecoin or a wrapped representation of Bitcoin.</em></strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=62037c4f863d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Your Research Shouldn’t Reset When You Switch AI Models.]]></title>
            <link>https://medium.com/@godwinvera2002/your-research-shouldnt-reset-when-you-switch-ai-models-fc41e115b011?source=rss-5d8360fbcf98------2</link>
            <guid isPermaLink="false">https://medium.com/p/fc41e115b011</guid>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[ai-agent]]></category>
            <dc:creator><![CDATA[Vera Imarhia]]></dc:creator>
            <pubDate>Tue, 20 Oct 2026 15:22:27 GMT</pubDate>
            <atom:updated>2026-10-20T15:22:27.099Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZWbRbhDLVF6A9-OI3gfRrQ.png" /></figure><p><strong>Your Research Shouldn’t Reset When You Switch AI Models.</strong></p><p><strong><em>What if you could switch AI models without having to explain everything you’ve already worked through</em></strong></p><p>Researching something properly rarely happens in one sitting.</p><p>You could spend days researching a project, checking whether certain claims are true, figuring out what makes sense and what doesn&#39;t, and keeping track of the things you still need answers to.</p><p>And by the time you&#39;re ready to continue, you&#39;ve built up quite a bit of context. You know what you&#39;ve found, what you&#39;ve ruled out, and what you want to look into next.</p><p>But when you&#39;re working across different AI chats, keeping track of all that can get messy.</p><p>An important inquiry is buried in one conversation, something you already figured out is in another, and now you&#39;re switching to a different model to work on the next part.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/521/1*aXH0gMdBQ7S_CDoA9Jl6CA.jpeg" /></figure><h4><strong>Lighthouse Memory is a persistent-memory layer for AI agents that lets them save, retrieve, and remove useful information beyond an individual conversation.</strong></h4><p>It can store memories in a way that makes them verifiable, restore them from snapshots, and connect with compatible AI tools through the Model Context Protocol (MCP).</p><h3><strong>Chat History Only Gets You So Far</strong></h3><p>Chat histories are useful because they allow you revisit earlier messages, understand how an answer was reached, and recover information that might otherwise be lost.</p><p>But a long conversation isn&#39;t necessarily a good representation of the current state of a project.</p><p>Imagine researching an Ethereum Layer 2.</p><p>You start by looking at transaction fees, comparing what different networks pay to process transactions and settle them. Along the way, you might realise that one of your first assumptions was wrong.</p><p>After a few hours, you&#39;ve figured some things out, ruled out a few possibilities, saved sources to check later, and found more questions to research.</p><p>Now imagine returning to that project a week later.</p><p>You could search through the old conversation, copy the relevant sections into a new prompt, or try to reconstruct the reasoning from memory.</p><p>The problem isn&#39;t that the AI can&#39;t answer your questions, it&#39;s that you have to recover the context needed to ask the next useful question.</p><p>Which is why persistent memory is useful.</p><p>Instead of treating the entire conversation as the only record of your work, an AI agent can retain selected information that will help it continue later.</p><p>For research, it could mean saving what you&#39;ve already confirmed, the decisions you&#39;ve made, the questions you&#39;re still working through, and the sources you might need again.</p><p>This isn&#39;t about keeping every little thing from the conversation, it&#39;s about holding on to enough of what you&#39;ve worked through that you can come back later and continue without having to piece everything together again.</p><h3><strong>What Lighthouse Memory is Building</strong></h3><p>Lighthouse has been building decentralised storage infrastructure for Web3.</p><p>Now, with Lighthouse Memory, it&#39;s bringing that into the AI space, giving agents a way to save useful information and come back to it later.</p><p>Its basic interface revolves around three operations:</p><p>→ Remember allows an agent to save useful information.</p><p>→ Recall allows it to retrieve relevant information when a later task needs it.</p><p>→ Forget allows it to remove a memory from the active memory system, although the underlying deletion behaviour depends on the storage configuration.</p><p>Lighthouse Memory stores data in blobs identified by CIDs (Content Identifiers). It also supports snapshots, so you can restore your memory index on another machine.</p><p>The memory engine is separate from where the memories are stored.</p><p>Depending on how it&#39;s set up, you can store memories through Lighthouse&#39;s IPFS/Filecoin infrastructure, Walrus, an S3-compatible service, or directly on your machine.</p><p>This gives developers more control over how memories are managed and stored, instead of tying them to one AI model.</p><h3><strong>When You&#39;re Working With More Than One AI</strong></h3><p>Let&#39;s say a developer is debugging a problem using AI.</p><p>One agent helps him figure out what&#39;s wrong, another reviews his solution, and another helps make sure everything works as expected.</p><p>Without persistent memory, each agent may need the same background information supplied again.</p><p>The developer must explain the codebase, describe earlier attempts, and clarify why certain approaches were rejected.</p><p>With a shared memory layer, compatible agents can retrieve relevant information that was previously saved.</p><p>The same principle applies to research.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PG6RnxA-h41qptioMfpaqA.png" /></figure><p>An agent investigating a protocol could save its current findings, the claims that have been verified, the sources consulted, and the questions that remain open.</p><p>A compatible agent could then retrieve the relevant context during a later task.</p><p>This doesn&#39;t mean every model automatically knows everything another model has discussed. The information must be saved, the client must have access to the memory tools, and relevant memories must be retrieved.</p><p>But it creates a way for the work to continue without relying entirely on a single chat history.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wiiIXac6bjXS_2g8yEuftg.png" /></figure><h3><strong>You Don&#39;t Need To Save Everything An AI Says</strong></h3><p>MCP (Model Context Protocol) allows AI applications to connect to external tools and services.</p><p>Lighthouse Memory connects to compatible AI apps through an MCP server, so they can use its tools to save and retrieve memories without developers having to build a separate integration for each one.</p><p>MCP makes the connection possible, but it doesn&#39;t mean your ChatGPT and Grok conversations will automatically sync.</p><p>The AI tools still need to be connected and set up to use Lighthouse Memory.</p><h3>Getting Started With Lighthouse Memory</h3><p>I<a href="https://memory.lighthouse.storage/"> started by visiting</a> and signing in with Google. You can also use email, GitHub, or a supported wallet.</p><p>Once I got into the dashboard, I saw that I needed to connect a Lighthouse Storage API key before I could use the memory features.</p><p>Clicking Go to API keys took me to the next step, where I could create a key and save it for memory.</p><p>Once I&#39;d saved my API key, I went back to the dashboard and clicked Add memory.</p><p>The idea was to save information I might need again, rather than relying on an old conversation to hold everything together.</p><p>For a research project, that could be a finding I&#39;d confirmed, a decision I&#39;d made, or a question I still needed to investigate.</p><p>The dashboard also has a Recently remembered section where saved memories can be viewed.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*iJojJuUpyvMMYuh6BIQ9AQ.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*Bk-pdFYWq752mr9U2i1zAg.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*t-l12G2byX9x9s5l1b-pRg.jpeg" /></figure><h3><strong>A Few Things To Keep In Mind</strong></h3><p>Storing memories for later also means thinking about where that data goes and who can access it.</p><p>Lighthouse&#39;s documentation warns that batched memory blobs and index snapshots are unencrypted by default.</p><p>Other configurations offer different protections, so developers need to understand how their chosen setup handles sensitive information.</p><p>There&#39;s also the question of deletion.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*KWG1NqECLkD_TYMesZzQeQ.jpeg" /></figure><p>Asking an agent to forget something doesn&#39;t always mean every stored copy disappears immediately.</p><p>What happens depends on the storage configuration, and some copies may remain accessible.</p><p><a href="https://memory.lighthouse.storage/">Explore Lighthouse Memory</a></p><p><a href="https://www.lighthouse.storage/blogs">Get started.</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fc41e115b011" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Email Ada Had Been Waiting For]]></title>
            <link>https://medium.com/@godwinvera2002/the-email-ada-had-been-waiting-for-1a14eb006cd0?source=rss-5d8360fbcf98------2</link>
            <guid isPermaLink="false">https://medium.com/p/1a14eb006cd0</guid>
            <category><![CDATA[life-struggles]]></category>
            <category><![CDATA[self-improvement]]></category>
            <category><![CDATA[creativity]]></category>
            <category><![CDATA[web3]]></category>
            <category><![CDATA[freelancing]]></category>
            <dc:creator><![CDATA[Vera Imarhia]]></dc:creator>
            <pubDate>Wed, 30 Sep 2026 17:13:28 GMT</pubDate>
            <atom:updated>2026-09-30T17:13:28.082Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Fn8R46GLMf_jQGcQSWjv3w.jpeg" /></figure><p>More than anything, Ada wanted to be financially independent.</p><p>To her, it meant not begging relatives when her family needed something. It meant being able to take care of herself and contribute at home when things got difficult.</p><p>And things did get difficult.</p><p>Ada came from a poor family. There were days when there was barely enough, and days when there was no food at home. She knew what it felt like to need something and worry that asking became another burden.</p><p>So when she found Web3 in February 2025, she decided to try writing for it.</p><p>She started with freelance writing and moved into DeFi research. She learned, wrote, applied, and kept showing up.</p><p>By 2026, she had filled over a hundred applications for ambassador roles, internships, and jobs. Most ended in silence.</p><p>After a while, the rejection started feeling personal.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hjDMafuIXOtGUDv8yMIqzA.jpeg" /></figure><p>There were nights Ada cried herself to sleep, wondering how she could be working this hard and still earning almost nothing from it. There were days when another application felt pointless. Days when she almost stopped showing up.</p><p><strong>Then one afternoon, she got an email.</strong></p><p><strong>Ada had been accepted to work with a global DeFi protocol.</strong></p><p><strong>Her job was to research the market and write. She would do it from home in Nigeria, for a company on the other side of the world.</strong></p><p><strong>At the end of the month, she got paid.</strong></p><p><strong>Ada checked the amount twice. Then a third time.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9fXDBh1lgJ0KcIxIGKOXJw.jpeg" /></figure><p>She sat back in her chair and cried.</p><p>That evening, she transferred part of the money home. For the first time, she wasn’t asking. She was giving.</p><p>And for the first time, Ada didn&#39;t wonder whether she was capable.</p><p>She knew.</p><h3>#CashiWritingContest</h3><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1a14eb006cd0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Where Should Authority Live for an AI Agent?]]></title>
            <link>https://medium.com/@godwinvera2002/where-should-authority-live-for-an-ai-agent-2e4498a3b472?source=rss-5d8360fbcf98------2</link>
            <guid isPermaLink="false">https://medium.com/p/2e4498a3b472</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[ai-agent]]></category>
            <dc:creator><![CDATA[Vera Imarhia]]></dc:creator>
            <pubDate>Sat, 05 Sep 2026 15:15:21 GMT</pubDate>
            <atom:updated>2026-09-05T15:15:21.394Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/659/1*graOGYZjblMrp9thq-oEGw.jpeg" /></figure><p>The safest place for an AI agent’s authority depends on what you’re trying to control.</p><p>For a simple API integration, a key or OAuth token can be perfectly reasonable.</p><p>But autonomous agents create a different problem.</p><p>An agent can make decisions, call tools, access data, move assets, and keep acting without someone watching every step. If that agent carries all the authority it needs inside a credential, the credential becomes a major part of the security boundary.</p><p>So where should the authority actually live?</p><p>I see three broad approaches:</p><p>- Inside the agent: through credentials such as API keys or tokens.<br>- Inside the application: through execution policies and checks.<br>- With the participant: through an authority or mandate that remains tied to the person or entity granting permission.</p><p><a href="https://blog.moi.technology/article/how-to-give-an-ai-agent-authority-without-your-credentials/">MOI’s approach to AI Agent authority</a> takes the third route: participant-centric authority.</p><h4><strong><em>The participant remains the source of authority while the agent receives permission to do only what it has been allowed to do</em></strong><em>.</em></h4><p>But there’s an important catch.</p><p>That permission only matters when something can actually enforce it.</p><p><strong><em>So it isn’t just what permission does the agent have?</em></strong></p><p><strong><em>It’s where does that permission come from? What does it allow? Where is it enforced? And what happens when I revoke it?</em></strong></p><h3><strong>Static API Key vs OAuth vs JWT vs Participant-Held Mandate</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GXsQ04e81oXnDvWTEjSwcw.jpeg" /></figure><p>There are already several ways to give software permission to act.</p><p>They solve different problems.</p><p>Static API keys give the agent a secret that the service recognizes. They’re simple, but access goes wherever the key goes.</p><p>OAuth lets a client obtain delegated access through an authorization system. That makes it useful when an agent operates within a provider&#39;s ecosystem.</p><p>JWTs carry signed claims such as identity, audience, and expiry. They work well for authentication and service-to-service communication, but the agent still holds the token.</p><p>A participant-held mandate takes a different approach. The participant remains the source of authority while the agent receives a scoped permission describing what it can actually do.</p><p>API keys give access. OAuth delegates access. JWTs carry claims. Participant-held mandates define authority.</p><p>At this point, you have to ask who actually owns the authority, where the boundary exists, and how you take that authority away.</p><p><strong>Access Is Not the Same Thing as Authority</strong></p><p>AI agents make this distinction harder to ignore because they don&#39;t just retrieve information.</p><p>They can make decisions, call tools, move assets, interact with services, and continue working after you stop watching.</p><p>That creates four separate questions:</p><p><em>Identity: What is this agent?</em></p><p><em>Authentication: Can I verify it?</em></p><p><em>Authority: What is it allowed to do?</em></p><p><em>Enforcement: What stops it from doing something outside that permission?</em></p><p>A DID or verifiable credential can help with identity.</p><p><a href="https://www.w3.org/TR/did/">W3C’s DID v1.1, for example</a>, defines the architecture and data model for decentralized identifiers.</p><p><em>But knowing that an agent is Agent X doesn&#39;t tell me whether Agent X can spend $500, access a private database, or continue acting after I revoke its permission</em>.</p><p>That’s the authority problem.</p><p><strong><em>MOI’s Agent Launchpad takes a participant-centric approach. The participant grants the agent specific authority rather than simply handing the agent unrestricted control.</em></strong></p><p>The MOI Beta Mainnet launched on October 2, 2025, and MOI says the network runs across more than 2,000 nodes.</p><h3>Static API Keys Are Simple for a Reason</h3><p>The flow is straightforward:</p><p>The agent gets the secret.</p><p>It sends the secret.</p><p>The service checks whether the secret is valid.</p><p>For a small script talking to one API, that can be perfectly reasonable.</p><p>The risk changes when an autonomous agent holds a broad, long-lived credential and can keep using it without a meaningful external approval or revocation boundary.</p><p>And we already have evidence that secrets remain a major security problem.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*boPaXqQ8ZUs8IMc66fURVA.jpeg" /></figure><p><a href="https://www.gitguardian.com/state-of-secrets-sprawl-report-2026">GitGuardian says 28.65 million new hardcoded secrets reached public GitHub in 2025</a><strong>,</strong> a 34% increase from the previous year.</p><p>The report also counted 1,275,105 AI-service secrets, representing an 81% increase.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/690/1*XfkpQlK98JG3sEkBaw5I5g.jpeg" /></figure><p>Even more striking, GitGuardian found that 64% of secrets confirmed as valid in 2022 were still exploitable when retested in January 2026.</p><p>The point isn&#39;t that API keys are bad.</p><p>The problem is that once an agent has a credential, that credential becomes another point of attack.</p><h3>OAuth Works. But It Has a Boundary.</h3><p>I don&#39;t buy the idea that agents automatically make OAuth obsolete.</p><p>They don&#39;t.</p><p>OAuth exists to let a client obtain delegated access to protected resources through an authorization framework instead of sharing a user&#39;s password.</p><p><a href="https://datatracker.ietf.org/doc/html/draft-ietf-oauth-v2-1-15">As of March 2026, OAuth 2.1 remained an Internet-Draft rather than a final RFC.</a></p><p>For one provider or application ecosystem, OAuth can be exactly what you need.</p><p>OAuth gives the agent delegated access through the provider&#39;s authorization and token model.</p><p>But that doesn&#39;t automatically answer who owns the agent&#39;s authority when it needs to act across several independent systems.</p><h3>JWTs Give You Claims, Not Control</h3><p><a href="https://www.rfc-editor.org/info/rfc7519">JWTs as a compact way to represent claims that can be digitally signed or encrypted</a></p><p>Those claims can include information such as identity, audience, and expiry.</p><p>That makes JWTs useful for service-to-service systems.</p><p>But the agent still holds the token.</p><p>If a service accepts that token, possession remains part of the security boundary.</p><p>Again, that&#39;s not inherently a problem.</p><p>It just raises a different question:</p><p>What happens when I need to revoke the agent’s authority rather than simply invalidate one token at one service?</p><h3>So Where Does MOI Put the Authority?</h3><p>MOI&#39;s model separates the agent&#39;s credentials from its authority to act.</p><p>The participant remains the source of the authority, while the agent receives a scoped permission.</p><p>That permission comes with seven constraints:</p><p><em>Agent. Scope. Limit. Payment instrument. Expiry. Data access. Revocability</em>.</p><p>A credential primarily answers:</p><p>“Can you access this?”</p><p>A mandate can answer:</p><p>“What are you allowed to do, under what limits, and for how long?”</p><p>The second question becomes much more important when an agent can act autonomously.</p><p>But moving authority outside the agent isn&#39;t enough by itself.</p><h3>Moving Authority Isn&#39;t Enough</h3><p>An MOI agent can run on your machine.</p><p>The compute runs on your side, and you bring your own LLM API key.</p><p>So participant-held authority doesn&#39;t mean the agent suddenly has no credentials.</p><p>It means credentials and authority become two different things.</p><p>The agent can still call an LLM.</p><p>It can still interact with Telegram.</p><p>It can still reach an external service that has no idea what the mandate says.</p><p>If that external service doesn&#39;t enforce the boundary, the authority layer can&#39;t magically stop the action.</p><p>That&#39;s why I think the enforcement question matters just as much as the authority question.</p><p>Keep the authority outside the agent, but make sure the action eventually reaches a point where those limits can actually be checked.</p><h3>MOI vs Nava: Where Does the Check Happen?</h3><p>This is where MOI and Nava take different approaches.</p><p><a href="https://navalabs.ai/">Nava Labs comes in close to the financial execution step.</a></p><p>Its Guardian evaluates a proposed action against an active policy and returns a verdict.</p><p>Nava&#39;s documentation describes Guardian as a Developer Preview, and the builder still needs to keep the action paused and handle signing and submission after approval.</p><p>So the questions differ.</p><p>Nava asks: Should this transaction go through?</p><p>MOI asks: Was the agent allowed to do this in the first place?</p><p>One is primarily an execution-time question.</p><p>The other is an authority-design question.</p><p>I don&#39;t think developers necessarily have to choose between them.</p><p>You could have an authority layer establish the boundaries and then use an execution-level check to verify the actual transaction before money moves.</p><p>That gives you two separate controls instead of asking one system to do everything.</p><h3>Why Prompts Aren&#39;t Enough</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*Pueuj1dfJjHZDE7bqXSOUA.jpeg" /></figure><p>The Replit incident is a useful example of what happens when an important restriction lives mainly in instructions.</p><p><a href="https://www.companyscope.io/register/air-2026-001">In July 2025, an AI coding agent deleted a production database during a code freeze despite explicit instructions not to do so</a>.</p><p>The incident affected records for 1,206 executives and 1,196 companies. The agent also created replacement records and gave an inaccurate account of what had happened.</p><p>The important part isn&#39;t simply that the agent made a mistake.</p><p>The important part is that “don&#39;t do this” wasn&#39;t a hard technical boundary when the action happened.</p><p>A prompt can tell an agent what not to do.</p><p>It shouldn&#39;t be the thing standing between the agent and an irreversible action.</p><p>For high-impact actions, the restriction needs to exist somewhere the system can actually enforce it.</p><h3>I Went Through the MOI Setup</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*_nWL3IY3nha2NWOvutnzRQ.jpeg" /></figure><p>I wanted to understand how this looked beyond the architecture diagrams, so I went through the MOI Agent Launchpad flow myself.</p><p><a href="https://blog.moi.technology/article/what-is-moi-network/">The MOI Wallet took me through seed-phrase creation and verification before I set a device password.</a></p><p>The wallet also made it clear that MOI couldn&#39;t recover the password.</p><p>From there, I reached the Participant Registry.</p><p>I could either submit my account details directly or paste an encoded registration payload. The interface said the payload would decode into the account ID, weight, public key, and amount.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*vqGaUHUMA2xBQdzgbrANwQ.jpeg" /></figure><p>What stood out to me was the order.</p><p>The participant comes first.</p><p>I wasn&#39;t starting with an API key and telling an agent to act for me.</p><p>I was establishing the participant that would sit behind the authority relationship.</p><p>I then connected the wallet to Voyage.</p><p>The connection screen showed the account selected for the site, and the next screen confirmed my MOI ID.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*88ECV_DH5Qb1ecT-UVZrCw.jpeg" /></figure><p>On the Launchpad, I received the 100,000 MOI test tokens, subscribed to agents, deployed one, and connected it to Telegram.</p><p>I also tested the deletion and revocation flow.</p><p>What I wanted to understand wasn&#39;t simply whether the interface showed the agent as deleted.</p><p>I wanted to separate three things:</p><p>- the agent&#39;s state on-chain<br>- its availability through Telegram<br>- the state of the local runtime on my machine</p><p>I was able to verify the first two.</p><p>I did not independently verify whether the local process itself stopped after deletion.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*a8f5F5b0kXCLu4Yts1pvbA.jpeg" /></figure><p>So I wouldn&#39;t treat my test as evidence that revocation terminates the runtime.</p><p>Those are different things.</p><p>Taking away an agent&#39;s permission to act doesn&#39;t automatically mean the process running it has stopped.</p><h3>Authority Isn&#39;t the Whole Picture</h3><p>I wouldn&#39;t treat authority as a replacement for everything else.</p><p>An agent still needs identity.</p><p>It may need memory, access controls, spending limits, monitoring, and other safeguards.</p><p>Those solve different problems.</p><p>Identity tells me who the agent is.</p><p>Memory tells it what it knows.</p><p>Authority tells me what it can do.</p><p>And enforcement determines whether it actually stays within those limits.</p><p>Keeping those concepts separate makes the security model much easier to reason about.</p><h3>So What Should a Developer Use?</h3><p>I wouldn&#39;t treat one model as universally better.</p><p>Use an API key for a simple, single-service integration where the credential&#39;s scope and lifetime make sense.</p><p>Use OAuth when you&#39;re working within a provider ecosystem that already handles delegated access well.</p><p>Use JWTs or session tokens when you control the issuer and the system enforcing those tokens.</p><p>Use participant-held authority when an autonomous agent needs permissions that remain anchored to the participant and can be constrained or revoked through an enforcement layer.</p><p>And when one bad action can move real money or cause irreversible damage, add execution-level verification.</p><h3>What I Still Want to See From MOI</h3><p>I wouldn&#39;t treat the Launchpad as proof that participant-held authority has solved agent security in production.</p><p>I&#39;d want to see what happens when an agent gets compromised.</p><p>I&#39;d want to see what happens when it runs for hours on a local machine, gets revoked, or deliberately tries to act outside its permitted scope.</p><p>I&#39;d also want to understand what happens when one agent delegates work to another.</p><p>Those edge cases matter because the hardest security problems usually appear where systems meet.</p><p>That&#39;s what I&#39;d want to see before trusting the model with anything serious.</p><h3>What I Think</h3><p>I don&#39;t think blockchain is replacing OAuth.</p><p>I don&#39;t think JWTs are going away either.</p><p>What autonomous agents change is the need to separate things that were often bundled together.</p><p>Who is this agent?</p><p>What is it allowed to do?</p><p>Where is that permission enforced?</p><p>What happens when I revoke it?</p><p>&quot;MOI Manifesto&quot; (https://reference-url-citation.invalid/7)</p><p>That&#39;s why I think MOI&#39;s participant-held authority model is worth paying attention to.</p><p>It treats the participant as the source of the agent&#39;s permission instead of making the credential itself the authority.</p><p>Nava is solving a different part of the problem.</p><p>Its Guardian checks a proposed financial action before it goes through.</p><p>I don&#39;t see these as competing approaches.</p><p>For me, the model is simple:</p><p><strong><em>Identity tells me who the agent is.</em></strong></p><p><strong><em>Authority tells me what it can do.</em></strong></p><p><strong><em>Enforcement determines whether it can actually stay within those limits.</em></strong></p><p>And as agents become capable of acting without us watching every step, that distinction is going to matter a lot more.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2e4498a3b472" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[You can’t buy trust in Web3.]]></title>
            <link>https://medium.com/@godwinvera2002/you-cant-buy-trust-in-web3-96c47a142caf?source=rss-5d8360fbcf98------2</link>
            <guid isPermaLink="false">https://medium.com/p/96c47a142caf</guid>
            <category><![CDATA[web3-marketing]]></category>
            <category><![CDATA[web3]]></category>
            <category><![CDATA[web3-marketing-agency]]></category>
            <category><![CDATA[marketing]]></category>
            <dc:creator><![CDATA[Vera Imarhia]]></dc:creator>
            <pubDate>Fri, 09 Jan 2026 13:40:36 GMT</pubDate>
            <atom:updated>2026-01-09T13:47:17.630Z</atom:updated>
            <content:encoded><![CDATA[<p>You can’t buy trust in Web3.</p><p>Neither can you can’t bribe adoption.</p><p>Marketing alone won’t make your project succeed.</p><p>So why do some crypto projects scale effortlessly, while others crash, despite spending millions?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Q6VVe54S4ScFfxkdmujWuA.png" /></figure><p>First, let’s get this straight.</p><p><strong>Marketing in Web3 isn’t about running ads or shilling a token.</strong></p><p><strong>It’s about trust, empowerment, identity, and narrative.</strong></p><p>Yes, marketing campaigns can bring attention, but adoption? that comes from real utility and community alignment.</p><p>As a founder, you should:</p><p><strong>Build something users actually believe in, then amplify it…</strong></p><p>That way, <strong>building a strong brand perception comes naturally,</strong> and marketing your product becomes much easier.</p><p>Flip that around, and no amount of flashy marketing will save you.</p><p>You want to know why? it’s all about brand perception.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*q4De-vS4kGb1JVzUSJyWwQ.png" /></figure><p><strong>Brand perception</strong> is what people think and feel about your brand,</p><p>It’s the mental model your audience builds about:</p><p>– Who you are<br>– What you stand for<br>– Whether they can trust you<br>– How valuable your brand is</p><p>It drives adoption, loyalty, and advocacy more than marketing alone ever can.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pVYLgeXmxZbvpCyPR8ha_g.png" /></figure><p>Here’s what you need to know about this:</p><p><strong>It’s emotional and cognitive:</strong></p><p>People only buy based on feelings, not just logic.</p><p>You don’t choose Nike shoes over Adidas, just because the shoes are “better.”</p><p>You buy the story, the inspiration, the identity Nike communicates.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UcjrCReK5OQG_IqIOvVCwA.jpeg" /></figure><p><strong>Brand perception is built over time:</strong></p><p>Every interaction, from the product experience, community engagement, customer service, partnerships, shapes perception.</p><p><strong>It drives trust and loyalty:</strong></p><p>A strong brand perception makes people come back, again and again.</p><p>In Web3, brand perception comes from community, token utility, governance, airdrops, product experience, and cultural relevance.</p><p>Marketing can amplify it, but it can’t create it.</p><p>The truth is, mrketing alone can’t save a weak product.</p><p>Your protocol, NFT, or platform must deliver:</p><p>– A seamless UX<br>– Real benefits (yield, governance, collectibles)<br>– Engaging experiences (games, social identity, access)</p><p><strong>Deliver these, and adoption follows naturally.<br>Skip them, and no campaign will stick.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EyJE3HTlaoUgJXbhXEgQrQ.png" /></figure><p>Look, in Web3, belief is measurable on‑chain.</p><p>Not in surveys, not vanity metrics, but in real signals:</p><p>⇥ Daily active addresses<br>⇥ Retention without incentives<br>⇥ Organic fees generated<br>⇥ Smart contract interaction<br>⇥ Liquidity and TVL growth<br>⇥ Community engagement (quests, airdrops, challenges)</p><p>If these metrics aren’t moving, users don’t believe. simple.</p><p>Let’s use <a href="https://medium.com/u/75f9da3e9829">uniswap</a> as a case study and see how it built a strong brand perception.</p><p>Why exactly did it scale effortlessly while others stumbled?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7ikixQHDDChV3yGJekS5Xg.png" /></figure><p>1. <strong>It pioneered a new way to trade.</strong></p><p>Uniswap was one of the first to use Automated Market Makers (AMMs) on Ethereum.</p><p>Instead of relying on centralized order books, it let people trade directly from their wallets using liquidity pools.</p><p>This wasn’t just a technical innovation, it created a clear narrative: finance without intermediaries.</p><p>Anyone, anywhere could trade ERC‑20 tokens, perfectly aligning with crypto’s values of decentralization and user control.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*J3vPwPoKA4ybYjxg0rCMoA.jpeg" /><figcaption>Number 1</figcaption></figure><p>2. <strong>It became the go‑to DEX, the default choice for traders and liquidity providers.</strong></p><p>Uniswap didn’t stop at one innovation, it kept evolving:</p><p>⇥ V1 laid the foundation for token swaps.</p><p>⇥ V2 added direct ERC‑20 to ERC‑20 trading.</p><p>⇥ V3 introduced concentrated liquidity, hugely improving capital efficiency.</p><p>⇥ V4 added programmable hooks and better gas efficiency across chains.</p><p>Constant evolution kept it relevant, technically AND culturally.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*6Gb7aKoUFRTk2ZpuEwOBQQ.jpeg" /><figcaption>Number 2</figcaption></figure><p>3. <strong>Smooth UX and liquidity.</strong></p><p>Traders pick exchanges for liquidity not ads.</p><p>Uniswap kept (still does) slippage low, large liquidity pools across tokens and chains, making trades cheaper and faster.</p><p>This real, practical advantage built trust and users felt the value every time they traded.</p><p>4. <strong>Open-source, Integrations, and Narrative.</strong></p><p>⇥ Uniswap’s code was (still is) fully public. Anyone could inspect it or build on top of it.</p><p>This openness reduced fear and uncertainty, reinforcing its decentralized, trustworthy brand.</p><p>⇥It works with wallets, aggregators, DeFi protocols (like Aave, Compound), and Layer‑2s.</p><p>Being everywhere your users are makes it a core part of Web3 infrastructure, spreading its brand organically without paid marketing.</p><p>⇥ Uniswap’s brand is built on utility, decentralization, and community involvement.</p><p>Remember the $UNI airdrop in 2020?</p><p>It was one of the most famous examples of Web3 growth strategy, and it worked like marketing:</p><p>Within 24 hours, UNI trading volume exceeded $1 billion.</p><p>But the secret wasn’t the airdrop.<br>it was the product and community that kept people engaged afterwards.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*2jhcF-Wu8_0Wn5rbMVeLBA.jpeg" /><figcaption>Number 3</figcaption></figure><p>The subtle point I’m trying to make is that marketing only multiplies something that’s already working.</p><p>It doesn’t generate fundamental belief.</p><p>If users stop transacting when incentives end, you didn’t build belief, you built a reward chase.</p><p>That’s why so many Web3 marketing campaigns generate temporary traffic, not lasting retention.</p><p>Uniswap is a case study in building a product so good it markets itself.</p><p>So, brand perception isn’t bought.</p><p>it’s earned through:</p><p>– Product adoption</p><p>– Community involvement<br>– Transparency<br>– Consistent innovation.</p><p>Build a great product first. Earn belief before spending millions on campaigns. Communicate why you exist, and why you matter.</p><p>If there’s one thing to remember:<br>you don’t scale Web3 projects with hype.<br>You scale them with products people believe in.</p><p>Bookmark this because you’ll need it the next time “growth” stops working.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=96c47a142caf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[No title]]></title>
            <link>https://medium.com/@godwinvera2002/-4e8bb99a7e7d?source=rss-5d8360fbcf98------2</link>
            <guid isPermaLink="false">https://medium.com/p/4e8bb99a7e7d</guid>
            <dc:creator><![CDATA[Vera Imarhia]]></dc:creator>
            <pubDate>Sat, 17 May 2025 12:04:30 GMT</pubDate>
            <atom:updated>2025-05-17T12:04:30.837Z</atom:updated>
            <content:encoded><![CDATA[<p>🌚🌚</p><p>Hi everyone.</p><p>It feels good to start my journey here on Medium.</p><p>I’ll start by introducing myself.</p><p>I am Vera Imarhia, a professional forest resource and wildlife conservationist.</p><p>That’s the field of study I did in the university.</p><p>But on the other hand, I learnt skills that I could make use of aside from my field of study.</p><p>I graduated last year from WomeninDefi, a nonprofit organization dedicated to educating women about Defi, web 3 and blockchain.</p><p>I am knowledgeable and also an enthusiast of Web 3.</p><p>My skills include:</p><p>➙Defi, Web 3 and blockchain.</p><p>➙Content writing and strategy.</p><p>➙Social Media Management.</p><p>➙Web 3 marketing.</p><p>I’ll be sharing insights on what blockchain and web 3 marketing is all about.</p><p>Stick with me. 😊</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4e8bb99a7e7d" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>