Jeonlees

Moshi moshi!!! I won a prize he he he — 1 BNB!!
It seems like it's the first time I've received a reward of this kind.
Thanks, Binance 🥰
$BNB

My article was forwarded by the official account!! Thank you for the official recognition!! @Binance News I will continue to create 💪@币安广场

Go on, little pig, give out red envelopes

After reviewing $GRVT’s economic model, the one question I want to ask most is: why does a user necessarily need to buy the coin?
The official describes $GRVT as a “membership key” for the GRVT ecosystem. By staking, users can get lower trading fees, higher yield boosts, larger GLP allocation amounts, and priority access to new products. It sounds like the utility is extensive, and that same set of benefits could also be obtained through paying a monthly USD subscription fee.
That’s a bit awkward.
If ordinary users do the math and decide paying the subscription fee is simply more convenient, then $GRVT demand will mainly come from people who are willing to lock tokens long-term. The problem is that we still don’t know the parameters that determine real demand: how many coins are required to stake, how long they must be locked, and where the staking yield actually comes from. Writing up ten usage scenarios doesn’t mean ten separate buy orders will naturally appear.
Value being returned to token holders has a similar issue. The official says platform revenue is used first for R&D, expansion, and operations, and any remaining economic surplus is then used for buybacks. But what exactly counts as “remaining”? Is there a fixed ratio, a minimum buyback amount, and a publicly verifiable execution record? If revenue grows, operational spending would likely grow in parallel too—so it’s hard to judge how much token holders will ultimately receive in advance.
The supply side is actually clearer: total supply is 1 billion tokens, with community airdrops at 22%, and Season 2 briefly increased that to 18%. A new event starting July 10 sets aside 1.5 million $GRVT and notes that they will be distributed on the TGE day. The original TGE timeline at the end of June has already passed. What the market is seeing now is that rewards continue to accumulate; the buy-side mechanism that can genuinely absorb the released supply is still waiting to be validated.
So my take on @grvt_io is quite straightforward: you can continue to like the product, but the token value closed-loop system still can’t rely on stories alone for the time being. After the TGE, only these two data points will be the real answers: whether paid users ultimately choose a USD subscription or staking $GRVT, and whether buybacks can be executed stably.
#grvt

Can Newton stop dangerous trades, but is what it sees the same market at the same moment?
I’ve been watching the Newton Mainnet Beta recently, and found a problem worth traders’ attention beyond simple latency: when a strategy reads price, collateral status, and risk rating at the same time, are these data points actually from the same time slice?
On July 7, Newton announced an execution stack with RedStone and Credora. RedStone provides the price data, Credora outputs the risk ratings, and Newton combines the two signals into a single executable authorization decision. According to the official disclosure, RedStone covers 1,000+ assets across 100+ chains, while Credora already covers all ratings for Spark Savings and most of the Morpho Vault. The coverage is certainly large enough.
For high-frequency traders, the first reaction won’t be “how much it covers,” but whether the update times can be aligned.
Suppose the collateral price just drops below the threshold and RedStone has already updated, but Credora’s risk score is still from the previous round; or the risk rating is downgraded first, while the price data is still within its validity window. Newton can execute the strategy correctly, yet it may still make decisions based on a set of data that’s not synchronized in time. In calm market conditions this may not be obvious, but when stablecoins de-peg and on-chain liquidity is quickly drained, even a few seconds can be enough to completely separate actual slippage from the strategy’s expectations.
Newton Mainnet Beta is already live on Base and Ethereum. VaultKit can also bundle concentration, liquidity thresholds, and compliance checks into the same authorization decision, and will directly reject execution if any one requirement can’t be fulfilled. The mechanism is quite strict, and I appreciate this “check first before settling” approach.
But the next thing I want to see is for each data source: its timestamp, the maximum allowed staleness, and what happens when multi-source data is out of sync—whether the system rejects, degrades, or fetches the data again. For trade execution, verifying the decision process is only the starting point; stability truly comes from ensuring the input data reflects the same market moment.
@NewtonProtocol $NEWT #Newt

I studied @NewtonProtocol for three days, and for high-frequency traders who pursue millisecond-level responsiveness, Newton’s “pre-trade checks” mechanism is a thrilling paradox.
As soon as the mainnet Beta went live, the official story was that it “intercepts risk before settlement, just like Visa,” which certainly sounds solid. But as someone who builds algorithmic trading, what I’m watching is execution efficiency. Newton’s architectural logic is: your trading intent -> pre-authorization checks (TEEs + the RedStone oracle) -> on-chain settlement.
The problem is this: when market conditions turn extremely volatile, the latency introduced by this so-called “forced interception” becomes the biggest slippage killer.
At the moment, the project team emphasizes compliance and security, using Credora ratings and multiple real-time data validations to gate strategies. But as a trader, what I’m most worried about is: if the RedStone oracle feeds prices with even a few seconds of delay during extreme conditions, or if Credora’s risk calculations lag, does this so-called “authorization layer” truly protect funds—or does it lock my positions during a crash, leaving me staring at liquidity drying up while I’m unable to cancel orders?
The current logic is “authorization first,” but in high-frequency trading, speed is life. If the “checks” themselves become a blocking point, then does the stability brought by this Deterministic Execution come at the cost of survival rights during extreme volatility?
$NEWT is currently trying to define the “rules of passage” for on-chain assets, but before those rules take effect, I suggest the project team must clearly explain: in periods of extreme fluctuation, how will priority for canceling orders be guaranteed? If the data-source synchronization delay problem can’t be resolved, then these so-called “pre-checks” may become a burden on traders precisely when it matters most.$NEWT
Keep observing—see how future versions optimize the execution path. After all, compliance is for regulators; market depth and execution efficiency are what truly sustain traders’ lifelines.
#Newt #NewtonProtocol

This time, I’m not too inclined to count @grvt_io and how many new partners there are. I’m more concerned with one question: what exactly do these names bring to GRVT?
Right now, the two most worth-watching threads are Aave and Centrifuge with Janus Henderson. After Aave is integrated, users’ trading margin can continue to generate on-chain yield; Centrifuge brings the tokenized U.S. Treasury fund managed by Janus Henderson into the Earn product. And with the platform already expanding into gold, oil, and stock perpetuals, what GRVT wants to do clearly goes beyond a single Perp DEX—it’s more like stuffing trading, yield, and real-world assets into the same account.
The data isn’t exactly bad. In the recently disclosed reporting metrics, GRVT’s trading volume over the past 30 days is about $38 billion, TVL rose from roughly $10 million to $111 million, weekly active trading users at one point exceeded 16,000, and weekly retention reached 67%. This shows the product already has people using it continuously; it can’t be simply categorized as “an empty city propped up by points.”
But the contradiction I see is right here: Aave provides yield, Centrifuge provides assets, market makers provide the order books, and GRVT itself is responsible for putting all these pieces together. As partners get stronger, does GRVT’s own user distribution capability also get stronger at the same time? We still lack a clear set of data to prove it. For example, how many users enter the platform for the first time specifically because of the treasury product? How much Earn capital ultimately gets converted into trading margin? And how much incremental trading volume do stock and commodity perpetuals contribute? These numbers are more important than “partner announcements.”
My attitude toward GRVT is somewhat positive, but I won’t automatically give the ecosystem a high score just because the partner list is impressive. The truly convincing signal in the next phase is whether, after external assets come in, they can leave capital; after capital stays, it can generate trading; and after trading, it in turn improves liquidity. Only when this cycle runs can GRVT’s partnerships count as an ecosystem. If it doesn’t run, then it’s just a few great products put on the same page.
#grvt

What Newton Mainnet Beta is really meant to verify isn’t “whether AI can click a button for you,” but whether automated execution can be constrained, verified, and held accountable.
Newton Protocol’s core direction is to encode rules such as spending limits, identity and sanctions screening, and risk controls into the transaction authorization layer: after an AI agent or application submits a transaction intent, it must first pass predefined strategies before it can enter on-chain execution. For each strategy decision, verification credentials can also be generated to let external parties check whether the rules are truly enforced.
This is exactly why Newton Mainnet Beta is worth discussing. The Beta launch means the project is moving from “technical narrative” to real-environment testing: when facing different chains, real-time data, and complex transactions, can the strategies remain stable? Can operator nodes form consensus in a timely manner? Who decides when the rules are upgraded? If external data is wrong, the strategy misjudges, or the service goes down, do users have clear paths to appeal, pause, and exit?
Supporters will say that Newton uses decentralized operator nodes, BLS aggregated proofs, and EigenLayer restaking security, so the authorization result doesn’t have to rely on a single platform. This design is indeed more transparent than traditional back-end risk controls.
But the key scorecard for Mainnet Beta shouldn’t be only trading volume or the number of partnerships; it should include real strategy calls, verification success rates, fault handling, operator-node distribution, and whether developers are willing to integrate long-term. Only if these pieces of evidence keep showing up can the staking, fees, and governance uses of $NEWT potentially evolve from design into sustainable demand.
I’ll be watching the Beta progress, but for now I won’t equate “mainnet launch” directly with “commercial closed-loop completion.”
@NewtonProtocol $NEWT ， #Newt

$AAVE Took me a long time. It’s gone up, but... it’s only climbed back to break even 😭

After Codex and ChatGPT were merged, the original version of ChatGPT became “ChatGPT Classic.”

I use Classic quite a lot.

Today I found I can’t use it anymore. It shows: “Unusual activity has been detected from your device. Try again later.”

At first I thought my account was banned or there was a network issue, but I found that Codex, and even ChatGPT on the mobile app and the web version, can all log in normally.

Then I wondered if it was because I’d downloaded an ancient version. I thought I’d uninstall and download it again, but it turns out there’s no way to download it again anymore.😭

When I log in, everything is the new version. For the old version, there’s only one webpage, and I can’t select the model either: chatgpt.com/g/g-YyyyMT9XH-chatgpt-classic

@OpenAI
Did you really discontinue ChatGPT Classic? 🥺

Today I went through the registration-to-order flow on @grvt_io in person. The most direct feeling is that GRVT wants to present itself as a professional trading platform, but in some places it assumes users already understand crypto trading by default.
When I first enter the page, the order book, positions, orders, and asset information in the trading area are shown pretty completely. Experienced users will find it neat and comfortable to look at, but new users may not.
Especially after topping up: once the money has already arrived, what to do next—where to transfer it, how to understand spot and perpetual accounts, and what other action is still needed before placing an order. If the page doesn’t provide continuous guidance across these steps, people are very likely to click back and forth between different modules. I know what to look for, so I paused and confirmed a few times; for someone using it for the first time, it’s even harder.
This feeling is even more obvious when withdrawing. With the same USDT, Ethereum has a fixed fee of 15 USDT, while Arbitrum and the BNB Chain only charge 1.01 USDT. The fees are displayed, but what users truly want to know is also: which chain my current address corresponds to, how much I’m expected to receive, and what happens if they choose the wrong network. If these reminders are hidden in the help center, and users are told to search on their own after something goes wrong, it’s a bit like distributing review materials only after the exam has ended.
Recently, $GRVT has already been confirmed to launch on July 21. Binance Wallet’s Booster campaign also put out 1.5 million $GRVT tokens. In the short term, it will definitely bring a batch of new users coming for the event. These people may not be familiar with GRVT and may not have the patience to study what each button does.
So my current judgment of GRVT is actually simple: putting aside whether the feature set is big enough for now, what I care about more is whether a newcomer can independently complete deposits, trading, and withdrawals—and whether they can immediately know how to fix things if they make a mistake. The true standard for a product being genuinely convenient is probably that users don’t have to go hunting for tutorials everywhere.#grvt

After Newton’s mainnet Beta went live, I’ve been waiting for more concrete implementation cases. On July 7, the official disclosure of its collaboration with RedStone finally got to the crux of the issue: on-chain vaults never lack risk data—what they lack is a clear authority to hit the brakes before a transaction is executed when the data becomes abnormal.
Many vaults have already integrated oracles and monitor price deviations, but these signals often remain at the dashboard and alerting layer. Once an administrator sees an alert and manually handles it, even a delay of just a few minutes could mean the position has already been adjusted. What Newton did this time is to write the price deviation data provided by RedStone into the VaultKit strategy. When the quote deviates beyond a preset range, operations submitted by the vault manager—such as rebalancing, raising limits, or enabling new markets—will undergo a check before execution; if the conditions aren’t met, the system blocks it directly.
This is also what’s most worth watching about Newton Mainnet Beta right now. Instead of creating another vault and migrating funds there, it wraps authorization checks around the existing management workflow, running first on Base and Ethereum. VaultKit is currently integrating multiple data sources, including vaults.fyi, Balancer, Webacy, Chainalysis, and Blockaid, so risk, liquidity, and sanctions screening can be composed into the same strategy.
What I care about more isn’t how long the collaboration list is, but rather: who sets the thresholds for each future strategy pack, whether updates are delayed, and how the system handles situations when a data source becomes invalid. Newton’s “fail the check means don’t execute” design has very strong life-safety properties, though the trade-off is that false positives could also block normal operations. What the mainnet Beta will truly need to prove next is whether this authorization layer can stay fast and accurate during volatile market conditions. It’s not surprising when on-chain rules are written beautifully—the real test is whether it still dares to intervene when the market goes wild.
@NewtonProtocol l $NEWT #Newt

Honestly, last time I missed out on that $ARTX 4x points boost,
and later I heard from a friend that they made quite a lot.
I regretted it for a while.
This time, no matter what, I can’t just be a bystander anymore.
As soon as the window opened, I rushed in with a small amount first.
Remember to DYOR
$ARTX #ARTX #Ultiland

The market is starting to debate GRVT’s token listing date again. As July 21 gets closer, questions like how much points are worth, and whether airdrops can cover the opportunity cost of time—those concerns are definitely real. But after reviewing GRVT’s data from the past year, I’m actually more concerned about another thing: whether it has the ability to continuously route revenue generated from trading back into the token ecosystem.
Last year, GRVT’s cumulative trading volume reached $177 billion, and its peak TVL hit $98 million. In the on-chain derivatives space, that’s not an empty project by any means—at least the platform has processed real trades. But impressive trading volume only proves that someone showed up; it doesn’t directly prove the token can sustainably capture and hold value. A common problem for derivatives platforms is that once incentives drop, trading volume shrinks too—faster than my wallet runs out of money at the end of the month.
So in GRVT’s latest design, I’m paying most attention to revenue buybacks, fee ownership rights, and margin efficiency. $GRVT has a fixed total supply of 1 billion tokens. In the future, holders can receive trading fee rates, strategy product allocation, and some platform benefits. If the buyback scale grows along with actual fee income, and if the buyback frequency, capital sources, and handling process are made public, then the token at least has a value pathway that can be observed and verified. Conversely, if after the listing it ends up relying only on points redemption and short-term incentives, that $177 billion of historical trading volume could easily become nothing more than background for marketing.
Recently, GRVT has also been extending into gold, crude oil, stock perpetuals, and RWA yield products, which suggests it wants to do more than just be a perpetual futures exchange—it’s aiming to become a full on-chain brokerage account system. The upside of this direction is indeed bigger, but the challenges are equally straightforward: the more products there are, the easier it is for liquidity to get fragmented, and compliance and risk-control costs rise as well.
So on July 21 I’ll look at the price—but I won’t look at price alone. What’s truly worth tracking is the natural trading volume after launch, fee revenue, buyback execution, and whether users are still willing to stay even without points incentives. Hype is there to attract people; retention is what settles the bill.
@grvt_io #grvt

After looking around the AI Agent space, I found an interesting shift: people are starting to move away from “how powerful the models are” toward focusing on “how these models are actually used and how they create value.” In the past, many projects talked about AI + Crypto, which sounded exciting—but once you get down to what happens on-chain, there’s often a sticking point: there’s a thick wall between model capability and real-world economics.
After the Newton Protocol Mainnet Beta went live, what I’m more interested in isn’t short-term data performance, but the long-term problem it aims to solve: whether the capabilities generated by AI Agents can be verified, invoked, recorded, and eventually form a sustainable on-chain collaboration system.
Based on the current design, what Newton wants to do isn’t just to put models on-chain for display. Mechanisms like the Model Registry are more like building a “model identity layer,” so that the models submitted by developers come with traceable information. When users and Agents call them, they can understand the underlying source, permissions, and status. This will be crucial for scaling up Agent operations in the future—because when AI starts performing transactions, searching, analysis, or even management tasks on behalf of people, trust issues will likely surface before model capability issues.
That said, I think there’s also something worth watching. During the Mainnet Beta phase, the real factor determining Newton’s ceiling isn’t just the technical architecture—it’s whether the ecosystem has enough developers willing to bring their own models and Agents in. If there aren’t real call-and-use scenarios, even a large number of registrations might amount to only a “model repository.”
So next, I’ll focus on several signals: the growth rate of the developer ecosystem, the frequency of on-chain Agent calls, how the models are actually used after being registered, and how Newton designs incentives for developers. The AI Agent space isn’t short of stories right now; what it lacks is long-term operational infrastructure.
Newton’s chosen direction is pretty challenging, because it’s tackling two of the fastest-changing areas: AI and blockchain. But if the Mainnet Beta can gradually prove that on-chain systems can make Agents easier to verify and more convenient to collaborate with, then it could become an important piece of the AI infrastructure puzzle.
Of course, the market will ultimately be decided by real usage. Keep watching and DYOR.
@NewtonProtocol l $NEWT #Newt