jojo橘子

I couldn’t sleep yesterday, so I also browsed the TermMax risk architecture documentation. There’s a design that really left an impression on me: complete market isolation. Traditional lending protocols like Aave and Morpho use shared-pool models, where all collateral is mixed together. The upside is higher capital efficiency; the downside is that one “bad apple” can spoil the whole batch—if one collateral asset blows up, the entire pool suffers.

TermMax takes a different route. Each market runs independently under its own contracts, with collateral, debt tokens, maturity dates, and risk parameters all kept separate. Lenders can know with 100% certainty what the single underlying collateral for their funds is before they deposit. If something goes wrong in one market, it won’t spread to others. This design reminded me of SPVs (special purpose vehicles) in traditional finance—each asset pool is isolated on its own, and risks don’t cross-contaminate. TermMax brings this logic on-chain and enforces it through smart contracts.

But isolation also has a cost. Shared pools are more capital-efficient and liquidity is more concentrated. Isolated markets mean each market must build its own depth; early liquidity may be more fragmented. @TermMax

I don’t know whether to sacrifice efficiency for safety, or sacrifice safety for efficiency. TermMax chose the former. Whether that’s the right choice depends on what the market values more. After multiple shared-pool blow-up events, an isolation-based design may end up being repriced. #termmax

It’s been more than half a year since Dusk’s mainnet went live, and recently I came across a piece of on-chain data analysis that made me stop and think for a long time.

The report’s core conclusion is very straightforward: this privacy Layer1, designed specifically for institutional RWA, hasn’t seen much activity on-chain from institutions since the mainnet launched. TVL is low, and the depth of its liquidity pools is limited. Although Dusk’s ZK cryptography architecture is designed for the MiCA framework and can enable private settlement, there’s still a “liquidity gap” between technical milestones and actual commercial value.

That conclusion hit exactly the point I’ve been vaguely worried about.

Dusk’s narrative has always been impressive—compliant privacy Layer1, an NPEX partnership, and €300 million securities tokenized on-chain. But announcing a partnership and having real funds actually running on-chain are two different things. Without sufficient depth in the secondary market, tokenized securities can’t be effectively cleared and reallocated. In the end, the value of a protocol must be supported by real trading demand, not by past promises. @Dusk

Of course, the mainnet has only just been live for half a year, so it’s too early to draw firm conclusions. But transitioning from a speculative roadmap to a functional institutional settlement layer takes time, and the next few months for Dusk are extremely critical. When will NPEX’s assets truly start running on-chain? And can on-chain trading volume keep growing sustainably?

I don’t hold any positions—just observing. No matter how solid a chain’s technical foundation is, it only matters if real money actually flows through it. #dusk $DUSK

The other day I opened TermMax V2, and the biggest feeling wasn’t that the interface changed—it was that the order book finally started to move.

With V1 and fixed-rate markets, the depth looked quite good at first glance. But when it came time to place an order, people (including me) would suddenly feel uneasy. You don’t know whether the depth in front of you is real, especially for large orders. The worst part isn’t that there’s no liquidity—it’s that you think there is. You press the confirm button, and the price suddenly slips outward. People snap awake instantly. A lot of the liquidity is locked inside someone’s little boxes. The order book clearly has stuff sitting there, but when trades actually happen, it’s like an empty shell.

V2 solves this. Unified Orders aggregates the curator’s range orders and the user’s limit orders into the same trade. The official “One quote, One signature, Best execution” really matters here. Even more important is Atomic Orders: with the same piece of liquidity, you can place multiple markets at once, and only when the trade truly executes will it sync and disappear. Many of the old issues in fixed-rate markets weren’t about not having depth—they were about depth being locked too early.

Now borrowers can post the borrowing cost they’re willing to accept, and lenders can wait for the target rate. For the first time, it feels like the price is growing on its own in the market. Add the multi-chain dashboard, and Vault orders from different chains all show up on the same screen. @TermMax

Of course, it’s not magic. Thin markets will still be thin, and in extreme conditions the price will still slip. But at least fixed-rate markets aren’t as “stuck” as they used to be. A lot of people didn’t dare take big positions in fixed-rate markets before—not because they were afraid the returns were low, but because they were afraid that when they wanted to exit, the door wouldn’t open. If V2 can truly make exiting gradually work like part of the liquidity too, then fixed-rate markets might, for the first time, start genuinely retaining large capital. #termmax

Dusk has been making a lot of moves recently: the mainnet has gone live, NPEX has signed on, and the Boreas protocol has been upgraded. But as I keep tracking everything, there are a few issues that I feel I can’t get around.

First is the definition of TVL. Data varies widely across different channels—some say it’s a few tens of millions, others claim it’s over a hundred million. If even the most basic locked-amount figure can’t be clearly stated, where does the credibility of an “ecosystem flourishing” narrative come from?

Second is the real value capture of the @Dusk coins. At the moment, DUSK is mainly used for Gas and staking, but the protocol’s actual income comes from the settlement fees of RWA assets’ trading fees on DuskTrade. The key question is whether that income can effectively flow back to the tokens. The design of Protocol Owned Liquidity is to buy back and inject liquidity into the pool—in essence, it’s equivalent to burning. The logic is sound, but the prerequisite is that there must be enough real on-chain trading volume. Given the current daily trading volume, it’s still a long way from having the conditions to kick off the deflationary engine.

Third is compliance—the double-edged sword. Dusk is deeply bound to MiCA and the EU regulatory framework. The upside is that institutions have clear rules to rely on. The downside is this: if regulatory details change, the underlying architecture may need to be modified accordingly. The deeper the compliance, the slower the turnaround.

I’m not denying Dusk’s direction. Privacy plus compliance really is a necessary condition for institutions to enter, and the direction is correct. But getting from “the direction is right” to “it can actually succeed” depends on countless execution details. Next, I’ll focus on three things: NPEX asset—whether the on-chain progress matches reality, changes in real on-chain trading volume, and network stability after the Boreas upgrade. Everything else is noise.#dusk $DUSK

Last month I moved, and I dug up a five-year-old time deposit certificate with an interest rate of 4.5%. I stared at that paper for a long time. Now, bank fixed deposits can’t even manage 2% anymore. It’s not that the money has gotten smaller—it’s that this thing called certainty is getting more and more expensive.

Later I came across TermMax and suddenly had a deja vu. What it does is essentially the same as that deposit certificate: it locks in a fixed interest rate, and you get your money back at maturity without having to worry about how the market changes in the meantime. The difference is that it brings this path on-chain and adds features that traditional time deposits could never have—leverage, RWA collateral, and more.

In January 2026, TermMax launched on BNB Chain the first fixed-rate lending market that supports tokenized stock collateral. Institutions holding Ondo tokenized stocks can borrow stablecoins without selling their assets. Isn’t that basically taking Wall Street’s stock pledge loans and putting them on-chain? @TermMax

But honestly, I was drawn in by the phrase “fixed interest rate,” and also trapped by it. Traditional finance took hundreds of years to get the fixed-income market to work smoothly. If DeFi tries to replicate it, the difficulty can only be greater. How does liquidity gather? How are markets with different maturities priced? And how do you exit when you suddenly need cash?
I like this direction, but it doesn’t mean it can run perfectly right now. Fixed interest rates are still a newborn in DeFi. How big it can grow depends on whether the market is willing to pay for certainty. Anyway, I’ll add TermMax to my watchlist for now, and see again after TMX’s TGE on August 25. #termmax

I recently reviewed Dusk’s architecture documentation and could clearly feel that the project has undergone a core upgrade: a brand-new three-layer modular design. Compared with the earlier architecture that was a confusing mix of components, the overall logic is organized in a way that’s not just a notch better, but in many ways a different level. @Dusk

At the very bottom is DuskDS, which serves as the core foundation of the entire public chain. It primarily handles data storage for the consensus mechanism and asset settlement. One particularly striking innovation here is the built-in MIPS-based pre-validation mechanism. Before any transaction is posted on-chain, it must first complete state verification at this layer. This allows it to completely avoid many of the typical issues seen in second-layer networks—there is no seven-day failure risk like Optimism’s, while still achieving both extreme security and excellent runtime efficiency. In addition, DuskDS only retains streamlined validity proof data, greatly lowering the operational threshold for nodes. Ordinary users can run a full node easily without needing high-end equipment.

The middle layer is DuskEVM, officially scheduled to go live in 2026, which perfectly supports all Ethereum EVM standards. Decentralized applications across the Ethereum ecosystem can be migrated and deployed directly. Developers don’t need to modify code or learn any brand-new programming language. Regular users can simply connect and use it with a MetaMask wallet—maximum compatibility.

The most critical highlight is its native support for fully homomorphic encryption. Transaction data is processed end-to-end in encrypted form. Compliance validation can be completed without decryption. This both protects user privacy and satisfies the industry’s hard requirements for regulatory compliance audits.

The top layer, DuskVM, is still continuously iterated and refined. In the future, it will become a dedicated privacy transaction module, mainly supporting transaction scenarios with high privacy requirements—such as enterprise confidential payments and private fundraising fund transactions. The three-layer architecture achieves interoperability through native bridging. The DUSK token can move freely across different layers without needing wrapped or derivative tokens, and without relying on any third-party custody institutions. The entire end-to-end process is driven natively by the DUSK token. Staking, mining, and fee-based ecosystem governance are all unified throughout—binding the three-layer architecture tightly together.

In an objective sense, Dusk’s underlying technical foundation is very solid, and the modular architecture approach is forward-thinking. However, in my view, no matter how excellent a technical architecture is, it ultimately still has to land in the ecosystem! #dusk $DUSK

While reading the Dusk economic model documentation, one detail made me pause. The total supply is 1 billion coins, with 500 million released gradually over 36 years through geometric decay—cut in half every four years. This kind of long-cycle design is completely different from projects that are fully circulating at launch.

What’s even more worth pondering is the way the token captures value. After DuskEVM goes live, DUSK did not introduce a separate Gas token. In the end, execution costs flow back to DuskDS. When one token carries the demand of the entire ecosystem, having no additional Gas assets means less token fragmentation. Simply put: the more on-chain transactions there are, the more DUSK is consumed—so the deflationary logic becomes much stronger.

Staking data is also changing. APR dropped from about 27% in November 2025 to around 22.31% recently. The DUDE browser shows 206 active Provisioners, with about 1.6 million DUSK locked. A decline in APR usually has two explanations: either incentives for new stakers are weakening, or the network is maturing. Personally, I lean toward the latter. The mainnet has been running for over half a year, and inflation expectations are tightening. Of course, no matter how clever the deflation model is, it still has to be built on real transaction volume. Today, Dusk’s daily trading volume isn’t that large. The real variable is how much on-chain interaction the 300 million euros of assets on NPEX can generate after they’re on-chain. Every transaction consumes DUSK—if that flow really starts running, the token consumption rate will be completely different. DUSK’s core logic, in the end, can be summed up in one sentence: as the ecosystem comes alive, the token has its story. @Dusk

Looking across most crypto projects today, many of them mostly chase short-term circulation and speculation. Dusk’s ultra-long token release model, by definition, is geared toward long-term value. Its deflationary mechanism, enabled by each coin, has a strong moat. There’s no issue of token dilution or ecosystem fragmentation. The project’s only real shortcoming right now is insufficient real on-chain traffic. As long as the NPEX entity assets land on-chain and real ecosystem transactions keep growing, token value will keep rising over time through the combination of deflation and fundamentals. The long-term investment thesis is very solid. #dusk $DUSK

I’ve found that a lot of people who look at Memes nowadays decide at first glance whether they’re bullish or not.
But I think there’s something else worth looking at: whether this project has “playability.”
$niulai, relying on the film IP 《牛来》, already has a certain foundation in content.
In the future, how the community creates memes, how it does remixes/creative reinterpretations, and how it “plays” with this IP is actually more important than simply promoting it.
I’ll keep paying attention to this direction. #niulai #牛来

Dusk’s tokenomics model recently made me revisit the whitepaper. A few numbers are worth pondering.

The total supply is 1 billion DUSK, of which 500 million are released gradually over 36 years using a geometric decay model for staking rewards, halving every four years. Each block mints 19.86 DUSK, with 80% allocated to block producers. This isn’t pulled out of thin air—it has been audited by POL Finance’s economic model.

What really caught my attention is the protocol-owned liquidity setup and its deflationary design. It’s not just simple burning; instead, protocol revenue is used to buy back tokens and inject them into the liquidity pool. As tokens become scarcer, market depth also improves—two birds with one stone. Compared with projects that only know how to burn, this logic is more sustainable.

Of course, no matter how clever the deflationary design is, it has to be built on real on-chain trading volume. With a daily transaction count of @Dusk , the volume is roughly in the hundreds of thousands of dollars range. With this scale, achieving deflation purely through Gas consumption won’t happen quickly. The real variable is how much on-chain interaction the 300 million euros of assets on NPEX can generate. Every transaction has to consume DUSK as Gas. If that traffic ramps up, the token consumption speed will be completely different.

I’m not betting on short-term price. But the logic of this economic model is internally consistent; value capture ultimately depends on the size of the real assets that get brought on-chain. In the coming months, I’ll just keep an eye on NPEX’s actual on-chain progress. #dusk $DUSK

There’s an old pagoda tree in my hometown courtyard. Every summer it spreads a large patch of cool shade. When I was a kid, I always asked my grandfather why he didn’t cut the tree down to make the yard brighter. He said, “What’s the point of being bright? Stand under the scorching sun for ten minutes and you’ll want to come back anyway.” Privacy is the same—it's not about locking yourself away inside your house; it's about knowing when you should be under the tree’s shade.

Later, I came across the whitepaper by @Dusk Dusk and found that they call it programmable privacy. It’s not a choice between full transparency and full anonymity; it’s a knob you adjust according to the scenario. Some transactions don’t need to be public, so keep them hidden. In some places, being a bit more transparent can actually increase trust, so make it public. More importantly, it leaves a key for regulators—but they can only check what they’re supposed to check, without having to hand over the entire set of accounts.

In plain terms, it’s like giving each asset and each role its own different lock, instead of one lock securing the whole household—or not locking the door at all. This tiered approach doesn’t rely on people’s self-discipline; it’s written into the protocol and executed automatically by code.

I really agree with this direction. It pulls privacy out of the misunderstandings that keep it in the dark and turns it into a controllable foundational capability. But the more flexible privacy is, the more complex the rules design becomes. Whether regulators are willing to recognize this logic—that’s the real hurdle. Don’t just take my word about DUSK; do your own homework. #dusk $DUSK

I had a bit too much to drink today, so I’ll share something that surprised me about Dusk Network.
To be honest, when I first heard about this project, I didn’t feel much—just another privacy blockchain. But after it launched its mainnet on January 7th this year, I went through its whitepaper and technical documentation carefully. The more I read, the more there seemed to be something real here.

What truly impressed me about Dusk is that @Dusk solves a long-standing deadlock in the financial industry: how privacy and compliance can coexist. Traditional public chains are either fully transparent, so institutions don’t dare to use them; or they’re pure privacy coins that can’t properly support KYC and audits—regulators don’t recognize them. Dusk uses PLONK zero-knowledge proofs to connect these two sides. By default, transaction data is hidden, but when needed, cryptographic proofs can be generated to verify legitimacy. All of this is done without exposing any sensitive information. This logic fits precisely with the requirements of EU MiCA and MiFID regulations.

The technical architecture is also quite thoughtful. DuskDS handles settlement, DuskEVM is compatible with the Ethereum ecosystem, and Solidity developers can get started right away. There’s also a WASM-based DuskVM that supports writing smart contracts in Rust. The modular design makes it more extensible—unlike some blockchains where making changes is harder than climbing to the heavens.

I think the smartest thing about Dusk is that it didn’t treat itself as a speculative tool. Instead, it has quietly and diligently built compliant privacy infrastructure. This path isn’t easy, but once it’s made it through, it becomes a real moat.
#dusk $DUSK

I recently spent a bit of time learning about the #宇宙之心 project, and I feel its special quality lies in choosing a narrative direction with a longer time horizon.
It isn’t just focused on short-term market hot spots; instead, it draws on the ideas behind SpaceX’s interstellar exploration, aiming to build a decentralized digital civilization ecosystem.
Of course, any project ultimately depends on its actual construction and execution capabilities, but this kind of exploration into the future is definitely worth continued attention.
#宇宙之心 $SPCXB

Let’s talk about SanDisk’s current prices and the outlook going forward. I’ve been keeping an eye on the consumer end and the current market for chip inventory for this period.

Now, original manufacturers have fully redirected their capacity toward AI enterprise orders, squeezing supply of consumer SSDs and storage cards. As a result, terminal retail prices continue to rise, and the price increases for 1TB solid-state drives and large-capacity SD cards are especially noticeable. Although the stock price has recently pulled back significantly, this is profit-taking after the earlier surge and doesn’t mean that physical storage will get cheaper. The supply-demand situation in chip inventory remains tight.

Looking ahead to the second half of the year, NAND chip prices will likely continue trending upward, but the rate of increase will slow compared with the first half. Consumer products are unlikely to see a sharp drop. The turning point will likely only be visible when new capacity is released in 2027.

The risk is also very real. It is able to support profits with large long-term AI orders, but storage is a strong cyclical industry. Once cloud providers reduce their capital expenditures and contract prices loosen, both enterprise business and consumer-side pricing will quickly come under pressure.

Ordinary users don’t need to gamble on a big drop. If you have a genuine need, buy in batches rather than hoarding and betting on price increases. #TradFi晒单

While I was going through Babylon’s technical documentation, there was one piece of data that made me double-check it again and again. Previously, deploying a ZK verifier on the Bitcoin network would cost roughly $15,000. But now, after the BABE protocol has successfully been deployed and run, that cost has been compressed to just $10–$20.

The cost difference is a full 1,000x—that gap is really staggering.

This data was publicly shared by Babylon co-founder David Tse at the Consensus 2026 conference. Let me briefly organize it. BABE stands for BAbylon-BErkeley. Its core is a Groth16 proof verification protocol. Its most crucial breakthrough—@BabylonLabs_io —is simplifying what used to be tedious and complex multi-pairing operations into a single scalar multiplication on an elliptic curve. This dramatically reduces computational cost.

In my view, the value of this technical breakthrough matters far more than the TVL surge that everyone has been discussing. Because the underlying logic of the entire TBV system is very clear: users lock BTC on the Bitcoin mainnet, then transmit the corresponding state proofs to the Ethereum network to complete lending operations. In the whole workflow, the step that transmits and verifies the proofs is absolutely critical. If the cost of this step stays high, then high-frequency verification simply becomes economically unworkable—no one will be willing to spend more than ten thousand dollars to verify a loan amount that isn’t that large.

Now that the verification cost has been pushed down to the level of around $20, it means the verification step has fully escaped the project bottleneck. The BABE protocol officially went live as early as February 2026 alongside the TBV alpha testnet. This also shows that Babylon is no longer just a project that can make the technical logic run—it genuinely achieves internal consistency in its economic model, and has real value for real-world commercial deployment. #baby

In fact, you can understand it with a very simple, straightforward analogy: even if the infrastructure pipeline is laid, whether it can keep running smoothly ultimately depends on the cost of using it. Now Babylon is effectively reducing the core usage cost by 1,000x, and only then does the entire ecosystem pipeline become truly unobstructed. $BABY

When I was reading the collaboration announcement between Babylon and Aegis, the point that really made me stop and think wasn’t actually the collaboration itself. It was one crucial sentence in the announcement: Aegis will provide fixed-rate credit products. In traditional finance, fixed rates are something everyone takes for granted. But in the DeFi space, this feature has long been a scarce resource. Loan rates on Aave fluctuate in real time with the utilization rate. If your borrowing rate is 5% today, tomorrow it could very likely jump straight to 15%. I’ve always felt that using floating rates for financial planning is basically like using a weather forecast to create an annual budget—it has zero stability.

At the moment, Babylon has built a native BTC-collateral infrastructure. Aave V4 fully handles the execution of lending and borrowing, while Aegis builds a fixed-rate credit layer at the top level. Users simply need to lock their BTC into TBV, then borrow stablecoins on Aave V4, and finally—through Aegis—have the borrowing cost fully fixed. In my view, this change brings a qualitative shift, especially for institutional users. For any institution’s finance department to approve a loan, the key prerequisite is that it can clearly understand the total interest cost. Floating rates add an extra uncertain variable—interest rate volatility—to the risk-control model. And precisely because of this hidden risk, many institutions simply give up on participating in this kind of lending market.

This feature is expected to go live in 2026 Q4.

I believe that @BabylonLabs_io isn’t just a simple product-stacking upgrade. Once BTC collateral comes with predictable borrowing costs, Babylon’s entire system can truly shed the label of a niche geek toy and become a legitimate financial tool that can be verified and implemented in practice. #baby $BABY

Let me tell everyone a shameful story. Last month, when SanDisk was pulling back from its high, I kept averaging down all the way, and all the way I got trapped.

I bought at 1400, then it dropped to 1300. I added at 1300, and it fell again to 1200. In the end, I couldn’t hold on any longer, so I cut my position—only to see a violent rebound just two days later. A textbook case of a retail investor’s move: averaging down in the middle of the mountain, and selling out on the floor. After that, I made a rule for myself: when building a position for the first time in any stock, never exceed 10%; any add-on must wait until it drops 10% or more. This market isn’t about who can make money faster—it’s about who can stay alive longer. #TradFi晒单

Tomorrow SanDisk will release its earnings report, and I’m so nervous I can’t sleep.

Market expectations are revenue of $8.3 to $8.4 billion, and EPS roughly $3.4 to $3.5. Goldman previously raised its target price from 1,200 to 2,200, calling it a “very strong quarter.” But what worries me is that even if results beat expectations, if the forward guidance isn’t impressive enough, investors will still sell off and drive the stock down. Last quarter, data center revenue jumped 233% quarter over quarter—can it keep that momentum this time? I’m not sure. My position has already been cut to just 20%; I’ll wait it out for now. #TradFi晒单

Yesterday I used native BTC collateral to borrow and lend on Aave V4. I personally tested the testnet flow end to end. What shocked me the most when switching over to Aave V4 testnet was that, in the entire process, there was actually no <approve> button for wBTC.

In the past, when borrowing stablecoins on Aave, you had to first swap BTC into wBTC, then approve it, and only then deposit it. Now with Babylon, this whole thing runs a different path: you lock your BTC into Taproot UTXOs on the Bitcoin chain. On the Ethereum side, it automatically generates an accounting credential called vaultBTC. But it’s not a token like wBTC that you can transfer around freely—the transfer function is effectively locked. You can only interact with whitelisted contracts from Aave.

The proposal deployed two Spokes: Babylon Core Lending Spoke handles borrowing and lending, while the BTC Vault Swap Spoke handles the settlement after liquidation. The former lets you use vaultBTC as collateral to borrow stablecoins. The latter, when you can’t repay, swaps the collateral BTC into WBTC to complete the liquidation.

I ran through the testnet. From locking the coins on the Bitcoin chain to seeing the collateral credited on the Ethereum side, I waited about 6 confirmations. The redemption logic is also straightforward: after you repay the loan, you submit a ZK proof. Once verification passes on the Bitcoin chain, the UTXO is unlocked. Throughout the whole process, my private key never left my own wallet. @BabylonLabs_io

It’s currently still in the Aave governance ARFC phase. The auditing parties include multiple organizations such as Coinspect, Sherlock, Zellic, and others. The testnet is already live—if you want to experience it, you can go to the Babylon Discord and join the support channels for the testnet. As for the mainnet launch time, it depends on the progress of the Aave governance voting. #baby $BABY

SanDisk No. 5 was released ahead of its Q4+2026 fiscal year full-year performance. Based on the recent U.S. stock earnings reports, it feels like another wave of declines is coming. People are expecting a lot from it—so it will definitely be the extreme opposite. This is a curse! But I still feel confident about the latter part of August. If Old Ted gets elected, they will surely roll out a lot of positive news to stabilize their political achievements. Otherwise, the loss from the current geopolitical issues is just too embarrassing! #TradFi晒单

Every order you get a few cuts and you run. You don’t take hits—you fold when the direction is wrong. Let me try this kind of style—will it work? Every day is losing money. I’m really losing too much! Every time it’s loss, it’s because I take the position and hold it through the drawdown. Today I’ll try doing a Sandisk (SNDY) wave/interval trade—will I make money or lose? The news flow today also feels okay, and the geopolitical issues have eased a bit. We’ll just see how things go when the US stock market opens tonight. #TradFi晒单