小猪天上飞

A suspected Trump X was stolen— in ten minutes, Pame turned into a phone, but the dog’s stake still has 20% in illegal insider trading

After listening to your suggestions, doing small-bet Alpha is indeed a bit tiring, but the wear and tear is definitely a lot less. Turns out it’s really hard labor 🥺

Impressions After Watching:
1. Getting big things done without spending a dime still requires a strong body for a man.
2. For personal finance AI advice, use Claude.
3. When talking to Brother Sun, if Jing Tian earns 3,000 something, then Zeng Ying still has to wash underwear—empowering yourself is the key.
4. Keep your nail clippers clean; girls really do care about your nails.
5. For ordinary people, practicing Sun’s teachings is hard—start by improving your writing.
6. This world is both cruel and tender. Jing Tian acts with righteousness, stepping into the game herself, and with three invitations, Sun Ge is brought into the capital.
7. AI still doesn’t know what love is. When they break up, Claude takes full responsibility.
Why did Claude stop Sun Yuchen from giving Jing Tian $50 million?
In Sun Yuchen’s “fictional” little essay about dating Jing Tian, many people noticed the exchange where Sun Yuchen asked Claude— the whole passage feels strangely out of place.

According to what’s written in the piece, he first had Claude Code run through his own API to get a handle on how much cash he had on hand; the conclusion was that this wouldn’t hurt him. Then he asked a more expensive and more personal question: whether he should give her that fifty million. Claude said, “No, he shouldn’t.” He asked again—whether she already didn’t love him anymore. Claude replied coldly and precisely, saying: “As for love, I don’t care and I don’t understand. But you cannot give her those fifty million dollars.”

As a language model, why would it step in and stop someone from giving away fifty million. The answer isn’t in its own thoughts—it lies in Anthropic’s training process.

Based on the training papers for Claude, I did an analysis.

I don’t understand why they made me grind this Alpha—it's getting harder to grind, I’m getting clamped every day 😭😭

Last night I helped a friend set up a demo for an on-chain bill/receipt settlement. I debugged it until the dead of night and almost got driven to vomit by trying to estimate Gas consumption and validate the state.
I originally thought of using Dusk’s Rusk virtual machine to run this business logic, since it mainly targets compliant finance. Under the hood, it packages privacy protection and audit traceability directly into precompiled instructions. Compared with our earlier approach on Ethereum L2—shoehorning in ZK circuits—or tinkering with encryption-heavy computations on Secret Network that are prone to causing state conflicts, Dusk’s smart contracts written in Rust are definitely much more lightweight. You just call the wrapped interfaces, and you can hide the real lending interest rates and collateral lists at the bottom level, while only proving to the regulator nodes that “this bill has no duplicated collateral, and the asset coverage ratio is compliant.”
When you take this logic to talk with the folks running traditional supply-chain finance, they can actually understand it—and they can also clearly see what it would take to make it real.
But the ideal is丰满; the real implementation has plenty of pitfalls. The biggest issue is that there are too few ready-made building blocks in the ecosystem. I wanted to add an automatic reconciliation listener script for the contract. I searched through the official documentation and community repos, and found neither decent oracle reference templates nor any standardized event-parsing libraries. A lot of basic middleware basically has to be written from scratch.
Worse still, the local testing environment is brutal. As soon as you increase concurrency a bit, the local node’s ZK proof generation speed drops noticeably. The error messages are also often vague, so you can’t even tell whether your own logic blew up, or whether the node’s cryptographic runtime isn’t keeping up.
In fact, the biggest moat of a compliant privacy blockchain has never been whose math proofs are more elegant—it’s how low the cost is to migrate real business workflows. Traditional finance asset tokenization isn’t about making technical crypto nerds learn the hard way. If even basic tasks like exporting transaction histories and building exception reconciliation tools require developers to reinvent the wheel, institutions simply won’t dare to put their core transaction flows on-chain.
Filling in that plug-and-play business suite will do more for real TVL than publishing ten architecture whitepapers.
If we apply this compliant privacy technology to real business, what direction do you think people will be able to get running first? #dusk $DUSK @Dusk

How to view Unitree Technology's listing: a nearly 40% pullback in just 3 days

“OpenClaw disappeared as if it had never existed.”
And from “everyone raises shrimp” to “OpenClaw is dead” took less than two months.
Early this year, an open-source AI agent dubbed as able to “take over the keyboard and mouse and automatically complete office tasks” burst onto the scene. It can perform real actions on a computer—managing files, sending emails, calling software, and more—earning it the label of a “24-hour digital employee.” Within less than 60 days, GitHub stars surpassed React; WeChat index soared to 165 million. By March 15, the cumulative token usage hit 10.4T, making it the AI application with the highest usage worldwide. Large numbers of users lined up to pay for paid installations, and even gave rise on Xianyu to services like “home installation for raising shrimp (dragon shrimp).”

But behind the runaway popularity was a kind of arbitrage game targeting AI companies. Users called Claude Pro or Max through OpenClaw, obtaining subscription authorization at a fixed monthly cost of $200, in exchange for API call volumes more than five times higher. It’s as if Anthropic is subsidizing each heavy “shrimp-raiser” user with several hundred dollars a month.

In April, Anthropic suddenly issued a notice: Claude subscriptions would no longer cover high-intensity calls by third-party tools like OpenClaw, and users were only given a one-day grace period. The power switch was flipped off, and the arbitrage chain was broken.

Meanwhile, OpenClaw’s real problems also began to surface: running 24/7 continuously consumes tokens, and for heavy users the monthly cost can reach thousands or even tens of thousands of yuan; frequent disruptive updates caused custom agents to become nonfunctional; and security vulnerabilities occurred frequently—one user had their API key stolen and received a 12,000-yuan token bill at 3 or 4 a.m.

By May, OpenClaw’s traffic had halved to 14.2 million, a month-over-month drop of 50.67%. Those who once paid high prices to get it installed started paying again to have it uninstalled. Even to this day, after the hype has ebbed away, the topic has completely gone silent.

Mu-tou Jie’s latest ARK rebalancing reveal: increased position in SpaceX while reducing Palantir.
Continuing to add to emerging sectors in space, biotech, and nuclear energy innovation competitions; making structural switches for certain stocks that have seen big gains—this is not an all-out bearish stance.

@Dusk official promotional materials always say “seamlessly move traditional financial assets onto the blockchain,” but once you get into real business scenarios, the experience is completely another story. We simulated a small-scale split and distribution of compliant assets. The underlying Zedger protocol really does have something to it—it can precisely distribute interest to eligible token holders while completely hiding the holders’ address and balance information. Judging purely by the cryptographic implementation, this logic is hard to criticize; at least in terms of anti-on-chain tracking and preventing counterparties from penetrating positions, the user experience is far better than with public ledgers.
But when you actually run this workflow in a real business, awkward problems show up immediately.
What’s most frustrating is the friction from multi-party coordination. When traditional institutions settle a transaction, behind the scenes there are custodians, underwriters, market makers, and audit firms—everyone is used to centralized systems with near-instant account synchronization. On Dusk, however, every party has to generate ZK proofs locally and then compare states via on-chain smart contracts. If any one party’s local client nodes experience network jitter or a proof verification timeout, the entire cross-institution batch settlement gets stuck midair. A few of us spent an afternoon just troubleshooting proof synchronization delays in the local environment, consuming the better part of our time. What kind of engineering friction is this—how could those “old-money” traditional institutions that only look at promotional PPTs tolerate it?
This also leads to a deeper industry knot: when compliant assets are put on-chain, what does the industry actually want—“decentralized settlement,” or merely a “cost-saving accounting tool”?
If the goal is to save costs, institutions could simply deploy a permissioned consortium chain or a centralized cloud database; the efficiency would be orders of magnitude higher, so there’s no need to pay Gas fees on a public chain. If the goal is decentralized global liquidity, Dusk—due to strict compliance admission—keeps most hot capital out of the gate. The result is: trying to borrow liquidity from the crypto retail side is hard because the threshold is too high for retail to enter; trying to eat into the traditional institutions’ existing big pie, but real operational efficiency is dragged down by on-chain consensus.
If traditional assets are onboarded at large scale in the future, which approach do you think is most likely to work in practice? #dusk $DUSK #油价维持跌势

Entropy raised $14 million in funding, but the term details haven’t been released yet
Entropy received $14 million, led by Ribbit Capital. However, the round, valuation, other investors, and how the funds will be used have not been disclosed.
At this point, the only thing that can be confirmed is that Entropy raised $14 million; other transaction details are basically not public. Ribbit Capital is the only investor mentioned—other participants are still unknown. The valuation, sector, and intended use of the funds haven’t been stated, so there’s very limited information. This looks more like a simple funding confirmation rather than a full transaction disclosure.
Only Ribbit Capital is mentioned, and the round wasn’t disclosed
In the announcement, Ribbit Capital is the only investor that appears. The stage of the round is completely unstated. It also doesn’t mention whether existing shareholders are adding more, whether new institutions have joined, or any other supporting parties. Usually, these details help people understand where the deal stands, but this time only the amount, date, and lead investor were provided.
On August 24, Entropy raised $14 million. The round was not disclosed. The lead investor is Ribbit Capital. The valuation, sector, and intended use of the funds were not mentioned. No other participating parties were listed.
The only information visible externally is the size, lead investor, and date
The timeline adds an August funding record, but there isn’t much to read. Without a valuation, there’s no pricing reference; without stated use, you can’t see the plan; without a sector tag, you can’t compare it with similar projects. The information density is low—enough to confirm that the funding happened, but not enough to be complete.
Still, this is at least a confirmed funding event, indicating that the money has already come through. In the public record, what can be seen are only the amount, lead investor, and date—no other terms are mentioned. Valuation, pricing, and allocation of funds can’t be inferred, so for now we can only compare the funding size itself.

This on-chain stuff is brutal, one marker and suddenly dozens of new coins. I calculated it and thought, forget it—I’m not playing anymore, P doesn’t work.

The latest holdings of US Vice President Vance have been revealed—ETFs are enough
US Vice President Vance mainly invests through ETFs. His assets include ETFs, real estate, Bitcoin, and more. His three largest holdings are the Nasdaq ETF, the S&P ETF, and the Dow ETF. In the first half of 2026, their gains were approximately 20.2%, 10.1%, and 9.4%, respectively.

His investment holdings are distributed as follows:
1. Nasdaq ETF ($ QQQ), accounting for 24%.
2. S&P ETF ($ SPY), accounting for 24%.
3. Dow ETF ($ DIA), accounting for 24%.
4. Gold ETF ($ GLD), accounting for 6%.
5. Long-term U.S. Treasury ETF ($ TLT), accounting for 6%.
6. Rumble ($ RUM, one of the few stocks he holds—an alt-right video platform), accounting for 6%.
7. Bitcoin ETF ($ FBTC), accounting for 6%.
8. Oil ETF ($ OILK), accounting for 2%.

What about Vance’s other assets?
· Investment style: The three major broad-market ETFs (QQQ + SPY + DIA) together make up more than 70% of his allocation, reflecting his approach to positioning for the long-term growth of U.S. stocks.
· Net asset valuation: According to the latest estimates for June 2026, Vance’s current personal net worth is about $20 million.

You missed the leading privacy chain—$ZEC —so go take a look at this privacy chain.
If you’ve actually been watching Dusk’s on-chain data and ecosystem delivery, you’ll find an extremely stark disconnect: on one side, there’s a highly compelling technical architecture and an official disclosure blueprint involving hundreds of millions of euros in cooperation; on the other, there’s massive friction in real-world deployment for retail users and developers.
The biggest paradox comes right from the very thing it prides itself on: its “institutional customization.”
Dusk has put in enormous effort to build a compliant foundation based on zero-knowledge proofs and the Piecrust virtual machine, tying it to traditional channels with proper licenses such as the Netherlands’ NPEX stock exchange. On paper, the logic is flawless: use hundreds of millions of euros of real-world assets to provide genuine settlement value on-chain.
But the problem is that the migration speed of traditional institutions is measured in quarters—even years—while crypto secondary markets can’t wait that long.
That creates a brutal situation:
More than 200 million tokens are locked in the mainnet and validation node staking pools. Token holders use real money while bearing the time cost, but on-chain—aside from some basic testing and fixed-interval settlement actions—daily high-frequency interactions and actual Gas consumption still have not formed a burst-like flywheel. To make up for the coldness in the developer ecosystem, the official team has recently been pushing the DuskEVM test environment, aiming to improve compatibility with the Ethereum toolchain and attract Solidity developers.
But that leads to the second core contradiction:
If a public chain that primarily targets institutional-exclusive compliance and its own ZK infrastructure still has to rely on EVM compatibility to attract traditional DeFi users to drive activity, then is its original positioning as a “pure institutional settlement layer” really not a compromise with real-world liquidity?
When your core technical threshold requires extremely rigorous compliance processes to get it running, ordinary developers and liquidity can’t get in; and when you compromise with the broader ecosystem, you then have to compete in the EVM cross-chain arena—which has long been a red ocean.
First stay alive, then get excited. Until the few hundred million euros in assets promised by NPEX truly run on-chain at high frequency and generate real protocol fees, any valuation model has to be discounted. No matter how hardcore the story is, in the end it still comes down to whether its fee-capture capability can actually sustain the entire validation network.
#dusk $DUSK @Dusk

Have you noticed that on-chain they’re “biting dogs” and, as you play around, your wallet ends up with lots of stock tokens?
If this time they kick off a new cycle, RWA will definitely be the focus. On-chain trending MEMEs are linking up with stocks.
Right now, many MEMEs on BSC are using LP pools to implement stock dividend distribution. There should be more new ways to do it on-chain too—we need to watch carefully and see if any talented devs have come up with more novel asset issuance methods.
If this is the early stage of a new bull run, then the new on-chain trend will be coming out soon.

Big Brother Monkey opens the trade with more courage than my 10U war god.
His recent positions:
BTC: 40x long, about 30 BTC, notional position about $2.4 million
ETH: 25x long, about 21,500 ETH, notional position about $41 million
HYPE: 10x long, about 146,500 HYPE, notional position about $10 million
PUMP: 10x long, about 350 million PUMP, notional position about $3.7 million

I directly looked up Dusk’s underlying interfaces and node status, without checking prices first—I pulled the latest batch’s transaction proof generation runtime and memory overhead. As for those whitepapers plastered with big institutions’ endorsements, it’s more convincing to set up a local environment, feed Dusk several rounds of concurrent test data, and see for yourself.
Let me put it plainly: Dusk didn’t blindly chase the big, all-in Turing-complete route. Instead, it poured its effort into its custom Piecrust virtual machine and the Phoenix state model. Dusk bakes confidential transfers, on-chain shareholder registry maintenance, and audit-oriented selective disclosure directly into the native protocol layer. When I ran tests locally, the cost of generating zero-knowledge proofs on Dusk was indeed very low—on ordinary devices, the fans don’t spin like crazy. Compared with Aztec’s relatively high mental overhead for development, or Oasis-style designs that rely on hardware Enclaves, where people always worry about side-channel attack risks, Dusk has found an extremely restrained engineering compromise between pure mathematical proof complexity and client-side compute overhead.
But after testing, I also want to poke holes. Sure, it can prove that Dusk’s chain is running—but it can’t prove that real business has formed a closed loop. Most of today’s on-chain interactions are still niche enthusiasts and nodes playing among themselves. Dusk’s local debugging environment sometimes throws error logs that are a bit obscure and hard to understand, and its developer tooling still isn’t truly “out of the box” ready. More importantly, Dusk focuses on security-type tokens and enterprise-grade settlement. Once you strip away the testnet’s self-generated trading volume, how many regulated, real-world off-chain custodians, KYC gateways, and compliant traditional asset issuers actually connect? Putting compliant assets on-chain isn’t just writing a few lines of confidential smart contracts—the real deep water is integrating off-chain legal title verification with settlement interfaces. Without real assets settling on Dusk, even the most refined cryptography design is only a sophisticated precision gear that spins in place.
Next, don’t just watch Dusk’s cumulative number of interactions—watch the daily increase in real institutional asset cashflows. No matter how good the technical framework is, you’re only building plumbing; whether it can attract real inflow is the key to staying alive.
In your view, for a compliant privacy blockchain like Dusk to truly take root, which part is most likely to become a bottleneck? #dusk $DUSK @Dusk

Everyone who’s playing Alpha and the competition is confused—it's starting the hell mode again. Really hard to play

Young and promising, Fan Xiaoqin; in this world, there will be no second “Ma Yun”
After watching the future 14.4 million times, only this time did I reverse my life

Hedger’s privacy engine and a two-layer architecture built to embrace the EVM ecosystem
The story Dusk originally told was utterly pure: a native, mathematically self-consistent approach to fully hiding state, pairing the underlying UTXO model with the Zedger model. But reality is harsh. Over 90% of DeFi developers and capital across the industry are settled in the Ethereum account model. To avoid being sidelined by the market, the official team introduced the Hedger engine in the latest DuskEVM, attempting to use a homomorphic encryption (HE) plus a lightweight ZK compromise to enable confidential transactions.
Compromise often comes with a cost. The account model’s global state nature makes it inherently difficult to achieve the kind of total privacy that UTXO offers. Hedger can indeed encrypt balances and transfer amounts, but the call relationships between addresses and the execution ordering of smart contracts still leave traces on-chain. This leads to an extremely awkward technical split: the underlying Rusk virtual machine runs original, fully compliant confidential contracts, while the upper-layer DuskEVM runs a semi-private account application that incorporates compromises. When funds flow across layers between these two logical rails, users not only have to endure the friction of bridged state transitions; degrading the privacy level may even indirectly allow data that was strictly guarded in the underlying layer to be inferred at the EVM layer.
Even more thought-provoking is the deterministic final settlement of the consensus mechanism Succinct Attestation (SA). For licensed clearinghouses (e.g., NPEX), blocks produced within seconds and absolutely non-reversible finality are a must. But inserting high-concurrency EVM smart contract execution and off-chain proof aggregation into such deterministic consensus causes the nodes’ verification load to rise almost exponentially. On the one hand, you want a rigorous, compliant, tamper-proof institutional clearing backbone; on the other, you want to cater to Ethereum’s retail frenzy and programmability. This architecture, repeatedly pulled back and forth between two completely different development trajectories—does it truly gather the best of all approaches, or is it burying the system under double layers of technical debt?
More often than not, trying to cover everything equals mediocrity. If the underlying architecture is torn into two execution environments for the sake of ecosystem compatibility, then in the end it may be unable to provide institutions with an ultra-pure security closed loop—and also unable to give Web3 developers the original, frictionless experience.
In Dusk’s hybrid privacy architecture, what do you think is the biggest technical risk?#dusk $DUSK @Dusk

ZEC pumping looks like the ETF has been approved