This is Park Jin Hyok, a North Korean hacker wanted by the FBI. Believed to have been born in 1981 or 1984, he is fluent in Korean, English, and Chinese. U.S. authorities indicted him criminally in 2018. He has operated within organizations linked to North Korea’s Reconnaissance General Bureau, and is accused by the U.S. Department of Justice of involvement in multiple major cyberattacks, including: - The 2014 Sony Pictures hack - The 2016 theft of $81 million from Bangladesh Bank - The 2017 global WannaCry ransomware attack - Several other large-scale cryptocurrency thefts Although the FBI has released his name and photograph, apprehending him remains difficult— because North Korea refuses to hand him over, and he is needed by the state to generate revenue 😭

Dear, the stock market has been a bloodbath these past few days 🩸 Even several stocks that everyone had high hopes for have dropped. I want to buy, but I’m afraid it’ll keep falling—what should I do? On CREAO AI @CreaoAI, I built myself a stock position-building AI Agent called the “Position Building Planner.” I simply tell it how much capital I want to allocate, which markets I’d like to invest in, my acceptable drawdown level, and how long I’d like to complete the buildup—and it generates a detailed position-building plan for me. This time, I used $10,000, selected the U.S., A-share, and Hong Kong markets, set a conservative risk preference, and scheduled staggered purchases around earnings reports, interest rate decisions, and other key events. It breaks down every consideration step by step: - How much capital should be allocated to each market? How much should be kept as cash reserves? - How many tranches should I use, and when should each one be deployed? - Under what conditions should I pause buying or reduce my position? - How should I adjust the plan if earnings reports or market conditions change? It even lists the “no-buy conditions” for each tranche! The generated Excel file is downloadable—change the total capital, and all amounts auto-update. If you want to swap out a stock or adjust allocations, just ask it to regenerate the plan. Try it here: https://t.co/yLK8nVrrHI Come experience it with me!

Hey, have you noticed? The price you see when placing an order sometimes differs significantly from the final execution price. This is especially true for large BTC orders worth millions of dollars—when they sequentially fill multiple ask levels, the average cost can rise substantially by the time the entire order completes. So even if the trading page shows low fees, it’s crucial to check at what price your order actually executed. Nado PRIME, launching in Q4 by @nadoHQ, aims to bring deeper liquidity to large-volume trades. As outlined in the design, the existing order book will continue operating alongside institutional market maker quotes. Users will still place orders through the same account and interface; behind the scenes, the smart order router will select the optimal execution channel based on order size and pricing. If the order book offers better prices, it will be used; if institutional quotes are more favorable, PRIME will step in. When needed, a single order can even be split across multiple market makers, with final settlement settled on-chain. For users, the convenience lies in not having to manually seek out market makers, negotiate quotes back and forth, or transfer funds between accounts just to access different execution channels. Positions are also interoperable: positions opened via PRIME can be closed through the order book, offering greater operational flexibility. And as for fees—transactions executed through PRIME carry 0% trading fees. We’re excited to see how, once live, multiple market makers quoting on the same large order can improve execution prices. After all, every small reduction in slippage means more funds stay in your account.

Dear, if you had a personal AI assistant, what would you want it to do for you every day? Moonlight hopes it could help organize industry news, research information, and compile key highlights into a daily morning brief. Even better—if it could automatically access data tools and pay service fees based on my preset budget, and have the report ready by the time I wake up! Ritual @ritualfnd is building the on-chain infrastructure to support this kind of AI Agent. As designed, these Agents can own wallets, manage funds and pay for AI computation within authorized limits, and schedule follow-up actions. Moonlight finds it fascinating that, in the future, different Agents could even collaborate with one another to accomplish more complex tasks. The project is also allocating 50 million RITUAL tokens—0.5% of the initial supply—as an airdrop to Infernet node operators, Frenrug users, and long-term community contributors. From helping us organize information to autonomously accessing resources and completing tasks within authorized bounds—we’re excited to see what innovative AI applications developers will build with Ritual! What tasks would you most like to delegate to your own AI assistant?

Is a female executive a spy planted beside Jensen Huang? Did Trump know about it a year ago? Jensen Huang, founder of NVIDIA, was featured in The New York Times today after U.S. intelligence officials determined that a woman closely associated with him, Alice Huang, may be gathering intelligence for Beijing. She is a senior executive at Megaspeed, a Singapore-based data center company. U.S. electronic surveillance revealed direct and close communication between her and Huang. In June 2024, weeks before Megaspeed began major purchases, the two met in Taipei for whiskey at a bar with other tech executives. In May 2025, they attended another business dinner together. More strikingly, Megaspeed’s Malaysian affiliate purchased at least $4.8 billion worth of AI equipment powered by NVIDIA chips between June 2024 and February 2026—less than two years. During a prior U.S. Commerce Department investigation, officials discovered that many NVIDIA devices in the company’s data centers were still sealed in their original boxes, raising suspicions that some chips may have been diverted to China. Most astonishingly, Trump reportedly received this intelligence briefing about a year ago, learning that Huang might be targeted by foreign intelligence operatives—but chose to take no action. Why? It is said Trump did not want to disrupt NVIDIA’s chip sales—which support the economy and stock market—or further complicate U.S.-China relations. Even after the report surfaced, Huang had not yet received an official U.S. government “defensive briefing,” typically issued to warn Americans who may be targeted by foreign intelligence. There is currently no evidence that Huang was aware of her activities or provided her with classified information. NVIDIA’s official response stated that meeting with clients is standard CEO practice; Jensen Huang speaks with thousands of client executives annually, and such interactions are not unusual. Clients do not gain access to sensitive business secrets or national security information. However, surveillance continues, and the affiliated company is still purchasing NVIDIA chips this year. A giant of the AI industry targeted by a beautiful executive—intelligence agencies are watching, and the White House knows—but pretends not to. This plot is more thrilling than any movie.

Dear, many people use WalletConnect’s features for payments every day. But did you know? With such a vast connection network, how will it generate revenue going forward, and what role can the $WCT token play in this? Take a look at this quarter’s data: 95.69 million WCT tokens have been staked, accounting for approximately 20.4% of the circulating supply, with 72,500 active stakers. The ecosystem has added 91 new wallets and nearly 2,000 applications, with quarterly network transaction volume nearing $100 billion. Combined with over 700 integrated wallets and more than 80,000 applications, WalletConnect has already built a substantial user base. However, the size of network transaction volume doesn’t automatically translate into how much revenue the network can generate. This is precisely why founder Pedro initiated the open discussion “WCT Utility Design Space” on the governance forum — it’s worth paying attention to. The community can now participate in discussions: Which services are suitable for charging fees? How should WCT be involved in settlement? And once revenue is generated, how should it be distributed among stakers, wallets, and other network participants? One proposed direction is to keep end-user access free while charging applications based on usage, with WCT facilitating settlement, and revenue distribution determined through governance. If this path is implemented, WCT could gain sustained demand tied directly to network usage — beyond its role in governance alone. Moving forward, Moonlight will monitor whether these discussions lead to concrete proposals and whether applications are willing to pay for services. I believe these developments will determine whether the massive usage volume can truly support WCT’s long-term utility. If you were designing this, would you prioritize application-side fees, wallet incentives, or staker revenue distribution?

Dear, there’s a place in Arizona, USA, called Chiricahua, which in Chinese is “Qilikawa” 🏜️ Because it sounds so similar to “Chiikawa,” it has recently become a pilgrimage site for fans. While others go for spring outings, Chiikawa fans go to “meet their family!” They take photos at the entrance with their beloved little characters. Chiikawa now has its own mountain in the U.S.🤍 https://t.co/pxNwlhrtf7

Dear, when researching Ethereum: 0xad29f2723fcdbcf665f210f25e06f97477e417cf @grvt_io, Moonlight first checks its trading data~ As of 10/24, DefiLlama shows that the 30-day perpetual notional volume is approximately $1.469 billion, with open interest at around $436 million. One reflects trading activity; the other indicates position size—looking at them together gives a clearer picture of the platform’s trading foundation. Active trading and meaningful open interest? Definitely worth further research! Another development that caught Moonlight’s attention: the official team completed its first buyback on 10/20, burning approximately 1.1 million $GRVT tokens—all publicly verifiable data. Having a mechanism with actual execution records is a strong plus in Moonlight’s eyes, and it’s worth tracking the scale and sustainability going forward. Also highly anticipated is GRVT 2.0, planned for release in November! Of course, Moonlight also wants to hear what you think about GRVT! Feel free to share your observations, product experiences, or growth perspectives in the comments. Until 11:59 PM on October 25, Moonlight will randomly select two participants from constructive, eligible comments to each receive $50 worth of $GRVT. No need to retweet, buy, or hold—just leave a comment. Let’s go! 🚀 @Grvt_zh

If nothing goes wrong, then something will go wrong. Solana threw a flashy light show on Shanghai’s Bund and hosted an in-person gathering—sure enough, the authorities showed up. Those who woke up early and got stuck in traffic to attend the event were lucky to have made the trip for nothing; some were directly questioned by the officers. In fact, last October, when Solana held an event in Shenzhen, police intervened and forced the event to end abruptly. Consider learning from Ethereum—keep a low profile, and every gathering goes smoothly. https://t.co/PDA5aGtSrI

Within a week, two L2 blockchains came to an end. 🥺 Two years ago, Yueguang was one of the few Chinese-speaking bloggers followed by the Abstract chain’s official team and the CEO of Fat Penguin. He also followed Pacman, the founder of Blast. He was invited to an exclusive gathering in New York, and even became the first person to speak Chinese publicly on Twitter. Yueguang held a significant amount of ETH within the Abstract ecosystem and was an early streamer during Abs’s initial push into the streaming space. Today, I transferred all of those ETH back to the Ethereum mainnet. A tear for the times.