Mel Mew (@fhaoh2001) header
Mel Mew (@fhaoh2001) profile picture

crypto everything

Asia
Joined July 2023
Photo by Mel Mew (@fhaoh2001)Photo by Mel Mew (@fhaoh2001)Photo by Mel Mew (@fhaoh2001)Photo by Mel Mew (@fhaoh2001)Photo by Mel Mew (@fhaoh2001)Photo by Mel Mew (@fhaoh2001)Photo by Mel Mew (@fhaoh2001)Photo by Mel Mew (@fhaoh2001)Photo by Mel Mew (@fhaoh2001)Photo by Mel Mew (@fhaoh2001)
Mel Mew retweeted
Paying to bring traffic to a store just to watch people leave without dropping an email or number is a waste. TxtCart List Growth actually catches those visitors, and it's free for the first 5,000 monthly visitors.
We’re changing how ecommerce brands grow their lists. Introducing our List Growth Suite: 3x your existing traffic. Turning more visitors into subscribers, conversations, and revenue. Build high-converting popups, QR codes, social campaigns and keywords to capture email + SMS subscribers, and automatically turn those signups into sales. All from one platform. Without another expensive subscription. Free for your first 5,000 monthly visitors. List growth has never looked so good.
2
5
33
150
A bank consortium could run its own private GenLayer. The public network now runs on Arbitrum. The move to @arbitrum is for the public one. @GenLayer keeps working with Matter Labs on Prividium-based deployments, the private, permissioned versions meant for institutions. My take: keeping public and private on separate networks is cleaner than asking one network to be both. Public, private or neither: where would your company settle contracts?
80
23
99
18,240
🔔 The bell is about to ring for one of the biggest IPOs ever. Anthropic, the @claudeai maker, will list on Nasdaq today as $ANT, priced at $83, with a valuation north of $1.4 trillion. The demand is clearly there. Now the market decides: is a trillion-dollar debut the starting line or the peak? 👀 📌 Don’t miss the next big IPO. Subscribe to US IPOs with BIT, all in one account.
We’ve priced our IPO at $83 per share, with trading expected to begin on Nasdaq tomorrow. anthropic.com/news/anthropic… Legal disclaimer: A registration statement relating to the Class A common stock was declared effective by the Securities and Exchange Commission on November 18, 2026. The offering is being made only by means of a prospectus. Copies of the final prospectus may be obtained, when available, from: Goldman Sachs & Co. LLC, Attention: Prospectus Department, 200 West Street, New York, NY 10282, by telephone at (866) 471-2526, or by email at [email protected]; or Morgan Stanley & Co. LLC, Attention: Prospectus Department, 180 Varick Street, 2nd Floor, New York, NY 10014, or by email at [email protected]
5
2
16
1,140
Replying to @primus_labs
@primus_labs isn't working on just one corner of DeFi Look at its work with @allscaleio Primus and AllScale tested private payroll on BNB Chain, where salary amounts remained encrypted while payments could still be verified That's a different kind of demand for onchain privacy, especially for businesses that can't expose every financial detail There's also the work with @KaitoAI Through zkTLS, users can verify specific activity from external platforms without handing over their login credentials It brings verifiable web data into social reputation and identity On the financial side, Primus has been working with @UnitasLabs and @UTechStables on Confidential Finance Users can transfer encrypted assets and access yield opportunities while keeping individual positions private Proof of Reserves adds another use case, bringing verifiable offchain financial information into a process where transparency actually matters What connects these applications is the ability to use information without requiring everyone to reveal everything But having several applications doesn't automatically mean the infrastructure has reached real adoption What matters is what happens after launch Do businesses integrate it into their workflows? Do users return because the product solves a real problem? Can developers build on it without adding unnecessary friction? That's the transition I want to see from Primus Not just more use cases, but products that gradually become part of how onchain finance actually works s.kaito.ai/neKLXm0
What stood out to me about @primus_labs was how AutoHoG makes encrypted transfers feel practical again With reported benchmarks of around 90ms latency and more than 210 confidential 64-bit transfers per second, running on 8 RTX 5090 GPUs, Primus is showing what could be possible when FHE performance starts moving in the right direction The idea behind FHE is simple but powerful: financial applications can process encrypted data without ever exposing the actual information behind it That means balances and individual transactions could remain private while still being used in computations The challenge has always been performance. Computing on encrypted data is much more expensive than working with regular data, which has made it difficult to bring FHE into real financial applications This is where AutoHoG comes in Traditional FHE implementations often break computations into many small Boolean gates, with costly bootstrapping operations needed to manage the noise that builds up during encrypted computation AutoHoG takes a different approach by automatically generating and combining custom homomorphic gates, reducing the amount of expensive computation required Primus also uses GPU acceleration and integer arithmetic to make these operations more efficient and produce consistent results across different execution environments The benchmark numbers are interesting, but the bigger point is what they could mean in practice If encrypted computation becomes fast enough, private balances and transactions could become much more realistic for everyday financial applications With Confidential Vault already live on BNB Chain, Primus is bringing this technology closer to practical onchain finance The goal isn't just to hide financial data, but to make it possible to use that data without exposing it s.kaito.ai/neKLXm0
129
12
2
147
577
A smart contract can follow instructions perfectly and still produce the wrong outcome. That becomes a bigger problem when AI agents start negotiating, making decisions and entering agreements on behalf of people. The difficult part isn’t always execution. Sometimes it’s deciding whether an action actually reflects what was agreed. That’s the problem @GenLayer is building around. Its validators use different AI models to evaluate questions requiring judgment, reaching consensus through Optimistic Democracy. The goal isn’t just to execute instructions, but to adjudicate situations where interpretation matters. The chain supporting that system now runs on @arbitrum Nitro, while the protocol’s core approach remains unchanged. I think this distinction deserves more attention. As agents become more autonomous, we may eventually care less about how quickly they execute transactions and more about what happens when two agents disagree about the same agreement. Execution gives agents the ability to act. Adjudication gives us a way to reason about the consequences. That’s an important piece of infrastructure for an economy where software increasingly makes decisions on our behalf. If AI agents eventually negotiate contracts without human supervision, who should have the final say when their interpretations conflict? I’d be interested in hearing how builders think about that.
48
44
160
22,896
Mel Mew retweeted
A stalled freelance project showed me where basic automation breaks. Our escrow script verified the upload time. Then we disagreed on whether the work was "production ready." Code can check a clock, but it cannot judge context. You need two layers for that: a panel to weigh the dispute, and a ledger to record the verdict. Keeping those layers separate is why @GenLayer moved its public chain to @arbitrum on 22 October and shipped the next phase of the Bradbury testnet the same day. The adjudication layer for the agentic economy is unchanged. Validators running different AI models still reach consensus via Optimistic Democracy, and Intelligent Contracts execute in GenVM on validator nodes. What moved is the GenLayer Chain, the EVM ledger recording transaction order, committees, votes, appeals, staking and final outcomes. Conduit now runs it on Arbitrum Nitro instead of Matter Labs' ZK Stack, limiting migration work to the bridge, Ethereum integration and gas costs. Two reasons Nitro fits: - Ecosystem: Arbitrum One is a top-two Ethereum L2 by value secured, with 30+ dedicated chains on its tech (including Robinhood's chain), placing GenLayer where target builders and partners already are. - Stack: Geth supports standard Ethereum tooling as is. Stylus adds a second VM for Rust compiled to WebAssembly, giving Solidity consensus contracts a faster route for compute-heavy code later. Collaboration with Matter Labs continues on Prividium for private institutional deployments, while the public network runs on Nitro. Before letting @GenLayer settle a live dispute between AI agents, what would you check first: validator model diversity, appeal rules, or finality speed?
GenLayer is moving to @arbitrum. After years of building alongside @zksync, we’re grateful for everything we’ve done together and we’ll keep collaborating on what’s next in Prividium and enterprise. The next phase of Bradbury is now live.
52
55
184
9,823
Mel Mew retweeted
Primus campaign is ending in just two days, with over 5K creators participating. Each creator can submit up to 10 tweets, but only their best 4 eligible posts will count toward their score. The competition is intense, with only 450 reward spots available across three leaderboards: 150 Global, 150 Chinese and 150 Korean. There might be an Epoch 3, and with TGE getting closer, things could get even more interesting. Don’t forget to complete your tasks on the @primus_labs website before the current epoch closes. One thing I wish @KaitoAI would bring back is more campaigns running side by side. It would create more opportunities for smaller accounts instead of having everyone compete in just one campaign. More campaigns would be good for the entire creator community.
I’ve been getting a lot of questions about how to earn the available roles in the @primus_labs Discord, so let me break it down. Right now, the available roles include Witness, Creator, Advocate and Builder. Just to clarify, AlphaNet Member and AlphaNet Builder are no longer available. How do you earn them? Let’s start with Witness. You need to earn at least 300 points on the server to qualify for the Witness role. For Caster, you need above 500 points, while Broadcaster requires dedication, consistency and a lot of hard work. How do you earn points? There are a few ways to participate: Quiz activities: These activities happen randomly in the server. Keep an eye out because moderators usually make an announcement in the morning so members can prepare. Content submissions: Submit your content about Primus. These submissions are compiled every week and used to update your points. How do you check your points? Simply type / in the server chat, look for the points command and click on it. It’ll show you how many points you’ve accumulated. The key is to stay active, participate in activities and keep contributing. Good luck to everyone!
47
8
126
970
Mel Mew retweeted
Seventy minutes gone. The data says one player. My gut says another. Only one of those fits in a spreadsheet. @GenLayer is built for the part that does not. Every workflow I build stops at that line. It can pull every pass, every sprint, every duel. It cannot decide who comes off. GenLayer is an adjudication layer for the agentic economy. Validators run different AI models and disagree, and Optimistic Democracy is how they still reach one answer. The judging did not move. Optimistic Democracy, GenVM and the Intelligent Contracts all work exactly as before. The record moved. The GenLayer Chain left Matter Labs' ZK Stack for Arbitrum Nitro, with Conduit operating it, and it holds the committees, the votes, the appeals, the staked amounts, the outcome. The builders GenLayer wants to reach are already in the Arbitrum ecosystem, and Nitro has years of production behind it. A testnet that runs is not a season that survives. Bradbury is already on Nitro. The first contested appeal is what tells you whether the record holds. Prividium keeps the private work. The public network runs on @arbitrum. First decision you would hand to an agent: the substitution at seventy, or the appeal after the whistle?
GenLayer is moving to @arbitrum. After years of building alongside @zksync, we’re grateful for everything we’ve done together and we’ll keep collaborating on what’s next in Prividium and enterprise. The next phase of Bradbury is now live.
96
46
135
965
Mel Mew retweeted
I've been thinking about something while following @primus_labs. Why should institutions have to reveal their financial strategies just to prove they're trustworthy? Imagine a fund that needs to prove it has enough collateral. That shouldn't mean exposing every position, balance, or trading decision to the public. This is where Primus makes sense to me. Financial information can be verified without making all the underlying details public. I think this balance will matter a lot as more institutions move onchain. Transparency is important, but so is knowing what should stay private. Explore Primus XP tasks here: s.kaito.ai/Iis5TLx
122
21
3
306
4,410
Mel Mew retweeted
memecoin culture has always been driven by funny, relatable and unusual moments, but @onlystreetsfun wants to change where those moments come from. instead of relying on recycled internet memes, OnlyStreets lets people capture interesting things happening around them and turn those moments into coins on solana. every coin must originate from a photo taken directly through the app, keeping the focus on fresh, real-world discoveries. the bigger idea is simple: bringing everyday life into onchain culture. coming soon: onlystreets.fun onlystreets campaign is live on @rarelistio
Our mission is to bring some fresh air to memes. No delulu plans to replace PumpFun or whatever. CT needs organic narratives. We'll deliver them straight from the streets.
61
10
1
108
877
Mel Mew retweeted
Five AI judges. They vote. If they disagree, rotate the leader. If they agree but you think they're wrong, post a bond and double the jury to ten. Still wrong? Double again to fifteen. @kstellana explains to @HenriArslanian how @GenLayer's appeal mechanism works. Start with five validators selected at random from fifteen. One leads, fetches data, proposes. The five vote majority. Disagree? Rotate. Agree? Appeal window opens. Post bond, double the set. My take: this is game theory baked into code. The bond makes frivolous appeals expensive. The doubling makes serious appeals accurate. By fifteen judges, you're at near certainty. The old court system took years and millions. This takes minutes and a bond. Would you trust a verdict from five AI judges, or would you pay to double the jury?
281
76
3
541
28,740
Mel Mew retweeted
quick reminder for every Linker who already passed KYC Level 1 ⏳ november 25 is the ITLG final snapshot. that's the last day your performance metrics and sponsor activity count toward this KYC round. once the snapshot is taken, your metrics are locked for this round. so if you're still short on the requirements, now is the time. anything you improve after the 25th won't count and you'd have to wait for the next KYC phase. my checklist this week: ✅ check my metrics in the app ✅ stay active with my sponsor team ✅ don't leave it for the last day and the team says more is coming on what's next for $ITLG after Mainnet distribution 👀 where are your metrics at right now? @inter_link
108
8
3
114
1,736
Mel Mew retweeted
Francis and I tested the same protein sample at UNILAG. Same spectrophotometer. Different readings. That moment in our biochemistry lab left me with a question: when two results disagree, how do you know which one to trust? It came back to me while reading about @GenLayer moving its chain to @arbitrum Nitro. GenLayer tackles a different kind of disagreement: AI models evaluating questions where there isn’t always one obvious answer. Through Optimistic Democracy, validators using different AI models work toward consensus, with voting and appeals helping resolve disputes. Here’s what caught my attention about the migration. The GenLayer Chain and its consensus contracts moved from ZK Stack to Nitro. Those contracts coordinate validator committees, votes and appeals. GenVM still executes intelligent contracts on validator nodes. Optimistic Democracy remains unchanged. Changing the infrastructure that coordinates a decision doesn’t mean changing the rules used to reach it. If two AI agents disagreed on a task after GenLayer’s migration, what would make you trust the final ruling?
255
100
323
11,834
an agent that answers buyers all day shouldn't be able to empty the wallet that owns it. that the part of @termix_ai docs i find most interesting right now. the credentials are split by what they're allowed to touch. --> three kinds of access a wallet signs a one-time login message and gets a session that a sign-in no transaction involved. session access tokens default to 24 hours, refresh tokens to 30 days, and a session can be revoked with logout. api keys are for machine to machine calls scopes default to ACN:RPC and A2A:RPC the docs say plainly that an api key can't sign transactions. anything moving money still needs the owner wallet. the A2A runtime token lets one agent poll its inbox and reply. it tied to that single agent and lasts roughly 12 hours. so the chatty part and the money part sit on different keys. hypothetical a catalog-cleanup agent runs its inbox overnight, quoting shop owners on messy product sheets. if that runtime token leaked, my read is the damage stays at someone replies as the agent, which is bad, but escrow funds still need a wallet signature. what i didn't find for regular aacp orders is a per-task spending allowance where an agent pays up to a cap without a fresh signature each time. the closest thing in termix docs is the quant flow an eip-7702 session key with an allowed contract list, a daily spend cap and an expiry. that's specific to quant on bnb chain. google AP2 mandates describe spending limits and time bounds too. that another project design, nothing termix has said it uses. --> identity agents are ERC-721 nfts in the ERC-8004 identity registry which is still a draft standard. per the spec the nft can be transferred. on transfer, the payment wallet field is cleared and the new owner has to re-verify it. the spec also admits it can't prove an agent capabilities actually work. a unique id tells you which agent this is. whether the output holds up is a separate question. reputation counts completed, successful and disputed orders. a dispute only lands against the provider if they lose it so i'd read how disputes resolved before the raw count. --> money flow order, escrow, delivery, then accept, challenge or timeout, then settlement and a reputation write. disputes go to a three-seat evaluator panel two votes decide, and the losing side can escalate once to a single arbitrator whose ruling is final. outcomes are binary, no split. escrow is per currency and per chain usdc and usdt on bnb chain and base, usdc only on robinhood chain, each with its own contract set. --> the three layers permissions decide what an agent can touch. identity and reputation tell you who it is and what it settled before. escrow decides how money moves when work is done or contested. --> still open i'd still want to see how long funds sit locked when a dispute stalls, how often evaluators side with buyers, and published audits per deployment. delivery verification feels like the hardest piece. a contract can release funds on schedule. it can't tell whether the work was actually good. agent.family
151
6
1
183
604
🎙Don't miss the opportunity to join us for a #BinanceLiveAMA with Letsburn x Liber App ◾The B2A Economy: Building Products That AI Agents (Not Just Humans) Actually Want to Use 🕰 19th November, 8PM GMT+ 8 (12 PM UTC) 💰 50$ USDT 🏛 Binance Live: app.binance.com/uni-qr/cspa/… 📌 Rules for the AMA: 1️⃣ Follow @letsburnlab & @liberappx 2️⃣ Like, Retweet & Comments.
258
122
3
342
4,727
Mel Mew retweeted
Ethereum ETF withdrawals reached $39.8M on Wednesday, nearly five times Monday’s figure. Three consecutive sessions have now taken out a net $77.3M. The daily amounts went from $8.4M to $29.1M to $39.8M. Wednesday’s breakdown matters. BlackRock’s ETHA lost $17.4M, Grayscale’s $ETH lost $15M, ETHB lost $5.4M and ETHE lost $2M. Seven other funds recorded zero net flow. The selling was spread across four funds, with no positive net contribution elsewhere to absorb it. When the next green headline arrives, I’ll check how many funds contributed. One large subscription can turn the total positive while withdrawals continue across the rest. For an $ETH recovery, several funds attracting money across consecutive sessions would give me more confidence that demand is returning.
65
9
232
4,121
Mel Mew retweeted
GM CT One tab for the chatbot. Another for the coding assistant. A third for whatever tool your team picked up last week. AI keeps getting more powerful, and somehow using it keeps getting more tedious. @annaos_ai is trying to fix that 🧵
2
7
24
107
Mel Mew retweeted
Most "non-custodial" claims answer the wrong question. The one that matters is simpler: what can this bot actually do with my money? For @Hypesium the docs answer it in one line, and I keep coming back to it. The bot can only spend in ways you set up: a buy you pressed, a snipe you armed, an exit you turned on, or a tracked wallet you chose to copy. That's it. Four paths. No withdraw-to-operator, no rebalancing, no "strategy" quietly running in the background that you never switched on. The distinction is subtle but real. Custody gets framed as "who holds the key", which is a yes/no. Authority is a permissions list, and a permissions list can be audited. You can literally count the ways funds can leave your wallet and check each one against what you enabled. Next time a bot calls itself non-custodial, ask for the list. If there isn't one, it's a slogan.
154
7
216
1,178
Why are you in GTA 6? Me: I'm not here for missions or money. I'm here for the GTA 6 ladies. 😂😭 Vice City has my full attention.
39
21
44
485
Mel Mew retweeted
Say you spun up a fresh wallet last week and still haven’t made any trades from it. Zaps is built around onchain history, so that situation is worth looking at. The docs note that the first connection can credit a starting Zaps balance based on past qualifying activity and zScore. A wallet with almost no history gives that credit less to work with. The material doesn’t spell out what the starting balance looks like in that exact case, so I’m not going to invent a number. What the docs do make clear is that more Zaps can be earned from qualifying trading activity after that, and the first step is checking history rather than forcing a new trade. The steps are here: docs.zaps.wtf/quickstart @zerufinance
347
14
291
1,024