Entering the era of artificial intelligence丨This article is a collection of AI projects and popular science!
🔔 Preface:
In today's society, artificial intelligence is no longer a fantasy in science fiction movies, but is integrated into every aspect of our lives. From voice assistants on smartphones to medical diagnosis and self-driving cars, artificial intelligence is changing our world at an astonishing speed.
The latest flagship model GPT-4o released by @OpenAI has also made great progress in technology. This momentum has not only attracted widespread attention in the encryption field, but also indicated that AI may become the next major trend in the integration of technology and finance in this dynamic and innovative field.
Dogecoin is great again—a new era of crypto is coming!
Preface Summary:
Evening report on May 4, 2024: Tesla announced that some of its products will accept Dogecoin ( $DOGE ) as a payment method.
On its official website, Tesla stated that eligible products will display the Dogecoin symbol next to the order button, and buyers only need to transfer Dogecoin to Tesla's designated wallet and complete the payment. This move is seen as an important step for cryptocurrency in mainstream business!
Musk and Dogecoin
Dogecoin: Its logo is a Japanese Shiba Inu named Kabosu, who is very famous on the Internet.
Bybit recently launched a USD1 promotion with a prize pool totaling 10 million $WLFI , divided into four mini-events running from late April to late May.
First off, let's talk about #USD1 . The trading pairs included this time are all mainstream coins like BTC, ETH, and USDC that most of us already hold, so there's no need to swap for obscure coins; it's all smooth sailing. Plus, it's running on the Mantle chain, which means fast speeds and low fees—when it comes to stablecoins, a smooth transaction is key.
The first event starts on May 6, featuring the USDC/USD1 and USD1/USDT trading pairs with zero fees. If you're frequently swapping these two coins, you can save quite a bit.
The second event is for spot trading, with the largest prize pool of 6 million WLFI. From April 22 to May 6, trade MNT/USD1, BTC/USD1, and ETH/USD1 on the main site, and if you hit 500 USDT, you can participate. The more you trade, the more you earn, with a max of 15,000 per person. This is perfect for those who are already looking to buy and sell Bitcoin and Ethereum.
The third event involves on-chain activities, offering 1 million WLFI. Also from April 22 to May 6, but you’ll need to use the Mantle chain to trade USD1 via Bybit Alpha, with the same 500 USDT threshold to join, and a max of 12,000 per person. If you haven't tried Alpha yet, now's a good chance to test it out and feel how smoothly USD1 runs on-chain.
The fourth event is the most fun, called Puzzle Collection, with 3 million WLFI up for grabs. From May 6 to May 22, it doesn’t matter how much you trade; just check in daily, trade, deposit, invite friends, and complete some extra tasks to collect puzzle pieces. You can even snag pieces by engaging on social media. Collect three consecutive pieces to enter a lottery, with a maximum of three entries per person and a 100% win rate. The first 500 people to complete the puzzle will receive additional rewards. If you gather extra pieces, you can share them with friends to complete faster.
A couple of reminders: both the spot and on-chain events run from April 22 to May 6, while the zero-fee and puzzle events kick off on May 6. Don’t forget to use the Mantle chain for the on-chain event; don't look for it on the main site.
Tongyi Qianwen is now on the blockchain, making it easier for AI agents to call it in the future.
I recently saw that the collaboration between the @0G Labs foundation and Alibaba Cloud has been finalized to integrate Tongyi Qianwen onto the blockchain. This means AI agents can now directly access this model on-chain, skipping the old-school traditional API route.
Why is this worth mentioning?
Currently, while there are more and more AI agents, their method of accessing large models is still quite primitive. They still require account registration, fiat payments, and manual configuration. This process is manageable for human operators, but since agents run tasks automatically and make high-frequency calls, it just doesn’t hold up.
So, the core change from this collaboration is that model access rights will be tokenized. Agents holding tokens can directly call Qwen, allowing for inference, decision-making, and task execution—all fully automated.
In terms of division of labor, Qwen is responsible for providing intelligence, while 0G ensures trust. One handles computation, the other handles verification. As 0G's founder puts it, it's “top-tier intelligence with trusted infrastructure.”
What I find interesting about this is that it indicates a trend. The capabilities of large models are already impressive, but what’s next? It’s about who can be more easily utilized by AI agents. The smoother the integration process, the farther they can go.
For developers, the benefits are quite tangible, allowing them to focus more intently on building the applications themselves.
This collaboration is just the beginning. But the direction is pretty clear.
There is a fairly common psychological phenomenon: when prices haven't risen yet, everyone habitually looks down on it. When prices really go up, everyone habitually demands too much.
Gold is a living example. At first, most people thought: the annual increase isn't that big, why bother? Later, gold really took off.
Those who previously looked down on it started to chase after buying. Moreover, they bought it for much more than the initial price, thinking: isn't this supposed to double?
The same thing, the same person, a change in price, and the appetite changes.
Now looking at Bitcoin $BTC , many people often say: even if it rises to 160,000 now, it’s just double. When saying this, it feels like doubling is the same as not rising at all.
But in other markets, if a stock or an index can double, that's something worth celebrating. Why is it that in Bitcoin, doubling has become just this?
To put it bluntly, it’s not that doubling is too little; it’s that everyone’s psychological expectations have long been pulled to a higher level, and they haven't realized it yet. I'm not saying Bitcoin won't fall. Given the current macro environment and geopolitical situation, no one can be sure in the short term.
But the real question worth thinking through is actually very simple:
Are you willing to buy slowly and then watch the price drop for a while? Or do you insist on waiting for that perfect low point, only to find that the market doesn’t give you the opportunity, directly pulling up, and you completely miss out?
To be honest, I think the latter is what most people truly regret in the end.
#WLFI 's proposal clarified the unlocking timetable that had not been explained before.
More than 60 billion tokens will be handled in two categories. The early participants, approximately 17 billion tokens $WLFI will be locked for another two years, and then gradually unlocked over the next two years, receiving what they are entitled to.
The other category consists of internal project members, founders, teams, advisors, and partners, totaling 45.2 billion tokens $WLFI . The treatment for this group is much harsher; if they want to follow the new rules, they must first pay out 10% to be directly burned, and the remaining will be locked for two years before being gradually released over three years. If they do not accept, then they will continue to be locked indefinitely with no other options.
Personally, I believe the most noteworthy aspect of this proposal is that the team proactively added the heaviest terms on themselves. More than 40 billion tokens will first burn 10%, and the remaining can only be fully accessed in the fifth year. Most projects in the circle are lenient towards their own people and strict with outsiders, but WLFI is doing the opposite this time, at least in terms of showing sincerity.
Early supporters need to actively confirm the new plan; otherwise, the original locked state will not change by itself.
Recently, #USD1 has been growing rapidly, having integrated Chainlink's reserve proof, with reserve assets available for real-time checking and still applying for a trust bank license in the United States. If this proposal passes, more than 60 billion WLFI tokens will not flow into the market for at least two years, and the ecosystem on the USD1 side is also gradually being established, with lending markets and payment SDKs being promoted.
Overall, their goal is to run both the governance token and stablecoin ecosystem simultaneously.
If you have been running AI models, you definitely know that Hugging Face is already filled with over 2.7 million open models, but 'being able to download' and 'being able to run in a production environment' are completely two different things.
A 70B parameter LLM requires at least two A100 GPUs to set up, and the hardware cost exceeds thirty thousand dollars.
💠 This is the first point raised in the @OpenGradient Four Pillars report: the promise of open models has not yet been fulfilled.
💠 The second point is even more painful—the real costs of closed infrastructure include: pricing being arbitrarily adjusted by a single vendor, content review policies changing at any time, single points of failure causing global applications to crash simultaneously, and sensitive data being fully exposed to central servers. These issues make high-value applications like DeFi Agent and financial risk models wary of fully relying on OpenAI or AWS.
The uniqueness of OpenGradient is that it directly addresses these issues. It does not force AI into traditional blockchain but designs HACA (Hybrid AI Compute Architecture), a mixed computing architecture that completely separates 'execution' from 'verification'. Inference Nodes specialize in running GPU inference (Fast Path), as fast as calling a central API; Full Nodes only verify TEE attestation or zkML proof without needing to rerun the model, reducing the hardware threshold so that general servers can act as validators, truly achieving decentralization.
💠 The third feature is full vertical integration, with the x402 protocol directly embedded in each TEE instance, allowing payments and inference to be completed within the same trust boundary without exposing data through intermediate layers; SolidML allows Solidity contracts to directly call models in the Model Hub, completing inference + execution actions atomically within a transaction; BitQuant is a real-world example of a DeFi Agent, enabling portfolio analysis through natural language, with all the underlying inference having on-chain proof.
💠 The fourth point is the most philosophical: open AI cannot stop at model weights; the execution infrastructure must also be open. OpenGradient proves that this path is viable with over 2 million verification counts and more than 500,000 zkML/TEE proofs.
I believe that for friends interested in the combination of AI development and blockchain technology, it is definitely worth exploring the charm of OpenGradient's products.
Everyone is most concerned about the early lock-up, and there will finally be some movement.
#WLFI Next week, the forum will propose a plan to unlock in batches slowly. I think the project team has considered this quite clearly; at least they haven't forgotten those who believed in them early on, nor do they intend to deceive the market.
#USD1 Recently, it has also been updated, and it's the kind that doesn't require you to do anything. The balance authorization and such are retained. The biggest change is that transfers no longer require ETH to pay for gas, saving a hassle for both people and AI operations.
By the way, it now natively supports those AI payment protocols, which developers can use right away, and compliance tools have also added some new features. In short, it's aimed at the future.
In terms of data, there is no clearing risk, and the returns are quite good; over 65 million dollars have been repurchased. The repayment side hasn’t lagged behind either; a few days ago, 15 million dollars were repaid, followed by another 10 million dollars, which can be seen on the chain.
Some people always pick on this project, but what I see is that while it is doing practical things, it is also thinking back on early users. Ultimately, it is building something that can last a long time.
At this point, I really don't understand why there are still people fixated on the interest of USDT.
When $USD1 came out, I transferred most of my spare money over. Yesterday I saw the annualized rate shoot up so high, to be honest, it felt pretty good because the mechanism of this thing easily leads to such pulsing market conditions.
When someone borrows heavily, the interest rate soars. When it soars, I eat a bit more. Although it has dropped a bit today, it is still higher than almost all stablecoins on the market.
Personally, I don't like to get into those overly mysterious things, like future ecosystems or grand narratives. I only look at two things: whether the pool is safe and whether the interest is real. So far, #USD1 has passed both checks.
So my advice is to put a portion of spare money in to earn interest and just check it every few days. It's definitely better than leaving it there to collect dust.
In a bear market, minimizing losses is profit, and it’s even better if you can earn a bit more.
🦅 #WLFI and @Aster DEX this cooperation, the most worth seeing is one point:
In the future, all TradFi perpetual contracts on Aster will settle using only #USD1 . USDT will not participate.
It's not support, it's exclusivity.
The first batch of the market has already come out: gold, silver, and crude oil, all priced in USD1. More will be added later.
Regarding the fees: taker 1 basis point, maker negative 0.5 basis points, placing orders can actually earn some back.
And the current position of USD1 is very clear, it's the underlying currency for RWA on Aster, and both sides are also looking at the integration at the token level, so there is a high probability that more will come out later.
The competition for stablecoins is no longer about who issues first or who has more chains. The key is who can first become the dollar for on-chain real assets.
#Aster and $WLFI this time are relying on settlement exclusivity + real market + low fees to pin USD1 down.
So there is a high probability that there will be further participant plans and actions at the token level, so we can keep watching.
Recently, the market really doesn't have much to look at; it's sluggish, and I'm too lazy to move.
But precisely because of this, it's easy to see who is working hard.
#USD1 Recently, the operations at a certain tea place are worth discussing.
They directly put USD1 into Launchpool, wealth management, and contract margin. These three places are basically where trading users spend most of their time. Being able to enter all at once shows that the platform is taking it seriously.
Moreover, USD1 itself has a solid foundation; it's a 1:1 pegged dollar issued by Bitgo, with reserves in short-term U.S. Treasury bonds and cash. WLFI is behind it, doing branding and operations. This kind of configuration is quite strong among stablecoins.
You can tell just by looking at its current Launchpool data; staking USD1 gets $WLFI in airdrops, and the annualized return is in the twenties, with interest paid every hour. The data looks okay to me, and the number of participants isn't too many; there are a few days left in the event.
To be honest, in this market, having something stable to let sit and earn interest without having to watch the charts anxiously every day is quite reassuring.
Moreover, the wealth management activities are coming in waves, obviously trying to get people to gradually develop the habit of using USD1. Once it becomes second nature, it will be hard to switch to something else.
On a larger scale, they are also expanding in payment and AI sectors. My feeling is that they are genuinely building out scenarios step by step.
Those who keep working hard even when the market is bad are truly worth our attention.
Just saw the news from WLFI on April 1st #USD1 set up a vault on Morpho, Monad chain, the first one.
The announcement mentioned a net annualized rate of about 13%, but when I checked it today, it was 8.19%.
It hasn't dropped. The new pool just opened with less capital, so the returns seem high; as more people gradually come in, the returns will naturally level out to normal. Old DeFi players understand this; it's called returning to normal, not unexpected.
Back to the point, this vault itself is a solid positive signal.
USD1 has finally established a legitimate vault on mainstream lending protocols like Morpho. The mechanism is simple: you deposit USD1, the protocol lends it out, others use assets as collateral, and the interest goes to you. Behind it, SteakhouseFi is making strategies, and Reservoir is issuing rUSD as a base, with a clean structure.
Now at 8.19%, what level is that?
It's about two or three points higher than Binance's wealth management. If you don't mind the hassle, you can cross over to #Monad and deposit it; you can indeed earn a bit more. The operation is not troublesome; just go to the WLFI official bridge to cross-chain, and then just enter the vault.
If you find on-chain operations cumbersome, continuing to hold on Binance is also completely fine; the difference isn't much.
My view:
Actually, the most interesting point about this vault is not the specific interest rate number but who it is built upon: Morpho, Monad, SteakhouseFi; putting these together at least shows that they are not just casually setting up a pool to fool around, but genuinely want to expand the scene.
As for whether to move or not, it's up to your habits. If you want to tinker, give it a try; if you're too lazy to move, just leave it.
Anyway, #USD1 itself is stable, and that's enough.
The market has been quite cold recently, and I haven't been active much, but when looking at the data, I've noticed that Solana's #USD1 has seen significant growth over the past two months.
Two months ago it was $160 million, and now it's $855 million. It has increased more than five times.
Moreover, the daily trading volume has stabilized at around $200-300 million, indicating that this money isn't just sitting idle in accounts; it's genuinely circulating within the ecosystem. The issue with stablecoins is that the worst-case scenario is having an inflated issuance without actual usage, but the trading volume of USD1 at least proves that it's alive.
I think one thing they did right is deeply integrating with existing projects on Solana like Raydium, Kamino, and BonkFun. Once stablecoins enter lending, trading, and liquidity pool scenarios, their role changes; they become essential infrastructure that everyone in the ecosystem needs.
In the current market, most people are actually hesitant to act recklessly. Retail investors are afraid to chase, altcoins are too risky to hold, and contracts are even more daunting to touch. What everyone truly needs is a reliable place to keep their money, ensuring it doesn't mysteriously shrink. In my view, the recent increase in USD1 isn't driven by speculation; rather, it’s capital seeking a safe place to stay.
I myself have kept some as well, combining it with some wealth management, without needing to watch the market daily. The returns aren't high but are stable. In a bear market, not losing is gaining; being able to sleep soundly is better than anything else.
I just finished reading Citrini's report, and I have some thoughts I want to discuss regarding the Stable
Because everyone should have a sense of this direction with Agent; AI has evolved from chatting to being able to work independently. It can help you place orders, compare prices, and handle tasks without needing a person to constantly monitor.
But to be honest, I had not considered that there would be such a significant issue with payments.
The current payment logic is still designed according to human habits. For example, when you transfer using USDC, you have to pay Gas with ETH. For people, it’s not a big deal; if it gets stuck, just add a bit more. But Agent is different; if it runs out of ETH halfway through execution, the program just stops, and that's not just a hassle; it’s fundamentally unmanageable.
💊 The report places Stable at the core of Agent's payment infrastructure
On the chain with #stable , it’s equivalent to using USDT simultaneously for both transfers and Gas, one coin from start to finish. This way, Agent doesn’t have to worry about whether there is enough Gas; the cost is stable, and it can just be calculated in dollars.
What Stable does is precisely what the Agentic economy truly needs.
When it comes to capability, it is already quite strong. Writing code, editing documents, analyzing data, and so on, it excels in everything. But if you want it to independently complete a task, like helping me order food or settling a bill, it just can't do it. It's not that it doesn't want to do it; it simply can't. There are no accounts, no private keys, it doesn't know how to sign transactions, and it doesn't know what gas fees are. What you can do is let it generate an order, and then you have to step in to handle the operations, which is quite awkward. If AI is forever stuck at the spending stage, then no matter how smart it is, it’s just a high-level consultant that can give ideas but cannot truly execute them. If this gap is not filled, AI working independently will always be a half-finished product.
The most talked-about event recently in the circle is the official launch of Chain Genesis, followed by World Liberty Fi jumping directly to #Aster to start making moves.
I looked through the messages they sent, and this time it feels different. It’s not the usual routine; it’s more like playing a game of chess, with spot trading, contracts, incentives, and collateral all intertwined. Moreover, USD1's integration this time with Aster essentially revolves around three key points:
🔸 Trade to earn $WLFI rewards:
Now, trading perpetual contracts like BTC/USD1, ETH/USD1, and SOL/USD1, there are 2.5 million $WLFI tokens available each month for distribution. I took a special look, and the rewards are distributed weekly, so you don’t have to wait several months like with some projects to see any action. It is said that there will be more than a dozen trading pairs added later, so those interested can keep an eye on it.
🔸 Holding also brings profits:
If you don’t want to deal with contracts and just hold USD1 on Aster, you can still join their incentive program, earning rewards daily based on your holding proportion. I think this is quite clever, turning stablecoins into something that can earn interest, which is indeed attractive for those who don’t want to operate frequently and just want to hold without devaluation.
🔸 The fees are indeed low to a harsh degree:
The Taker fee for the USD1 trading pair is only 0.5 bps, while the Maker fee is directly 0, and the Taker fee for USDT pairs is 4 bps. You can calculate the difference yourself. More importantly, USD1 can now be used as margin, with permissions on par with USDT, indicating that the platform is genuinely pushing it seriously.
In summary, this time the appearance of #USD1 on Aster feels more like the beginning of a deeply integrated ecosystem. It attempts to establish USD1 as the gold standard asset on the Aster chain through a combination of low thresholds (low fees), high returns (WLFI incentives), and multifunctionality (collateralizable, interest-earning).
If you have USD1 on hand or are preparing to enter the market, now is a good observation window. After all, early participation often comes with higher potential returns.
Of course, as always, the above is just my personal analysis and interpretation, not investment advice. Make sure to do your homework #DYOR before entering and understand the risks.
Recently, I watched a round of discussions about AI Agents and stablecoins, and indeed, the main character in payments has changed.
Imagine, if in the future, buying things and ordering services are all handled by AI, then the way we pay will certainly have to change.
Before, it was operated by people: opening a webpage, selecting items, clicking on payment, entering passwords. But with the Agentic commerce model, the scene has changed: AI helps you find services, automatically compares prices, and then directly makes the payment.
With this change, problems have arisen. Many machine services are charged on a per-use basis, a few cents for calling an API, a few cents for running computational power. These small transactions using traditional payment methods may incur fees higher than the actual consumption, which doesn't make sense.
At this point, looking at stablecoins aligns perfectly: they are quick to arrive, available 7x24 hours, and can support payments of a few cents, making them tailor-made for machine transactions. However, saying that stablecoins will replace credit cards is not very realistic. The foundation built by credit cards over the years, including refunds, risk control, and consumer protection, is extremely effective in e-commerce and daily consumption.
The places where stablecoins can really take off should be in the machine economy: those increasingly numerous API services, AI tools, and automated software, simply put:
➤ Traditional payments solve transactions between people and merchants, ➤ Stablecoins fill the gap for transactions between machines.
If AI Agents truly participate in commercial activities on a large scale, then stablecoins are very likely to gradually become the settlement tool at the core of this system.
🦅 Recently, USD1 has been discussed a lot, but I think many people are actually missing the point.
What USD1 is really doing is redoing the payment and transfer process. It's not for traders; it's for people who really need to use money in their daily lives. The reality is well understood: cross-border transfers are slow, fees are high, and in places where local currency is unstable, money simply cannot be kept. For many people, using stablecoins is not a choice, but a necessity.
The idea behind #USD1 is that transfers must be fast, costs must be low, and cross-border transactions should be the default. On the BNB Chain, there are no fees directly, which is crucial. You're just transferring money; you shouldn't be charged repeatedly.
So you'll see more and more real scenarios starting to use stablecoins for settlement. This is not just a concept; it's about efficiency. Paying suppliers, settling accounts, doing cross-border business—if the money can arrive the same day, no one wants to wait for the bank.
This also explains why USD1 has grown so fast recently. In just 4 days, the volume surged to 5 billion USD, clearly not driven by emotions.
🔥 Moreover, the USD1 points activity for #Binance is essentially amplifying real usage rather than just a pump:
➤ One part is basic tasks that everyone can participate in: in the designated USD1 trading pairs, trading equivalent to 500 USD gives you a chance to receive WLFI rewards, with a total prize pool of 3 million $WLFI , limited spots available, it's the kind of early participation that brings early benefits.
➤ The other part is a leaderboard prepared for active trading users: if the trading volume exceeds 1000 USD, you will enter the points ranking, dividing the remaining 9 million WLFI according to your rank; doing more and ranking higher naturally yields greater rewards, with a maximum of 12000 $WLFI .
📅 The entire activity runs from January 29 to February 27, which is essentially encouraging real trading and liquidity.
In my own view, the real opportunity for stablecoins is not in exchanges, but in the most unassuming areas such as payments, remittances, and daily transfers. Those who can successfully navigate these scenarios will truly retain users.
Irys: Turning data from static storage into executable assets
I recently revisited the materials on @irys several times, and the more I look, the more I feel that its focus is not on long-term storage, but on making data come alive.
In the past, when we talked about data on the blockchain, it was more of an archival mindset, and just putting files on Arweave or IPFS was the end of it. But Irys's logic is completely different; it allows data not only to be stored but also to be executed, called upon, and even directly participate in logic.
In other words, data has transformed from static resources into programmable assets. This means that AI agents can directly read and write data, smart contracts can automatically execute copyright logic, and games and social applications can run states and rules on the same chain simultaneously.
In terms of performance and cost, Irys seems to be rewriting the rules. Permanent storage is over a dozen times cheaper than Arweave and several orders of magnitude lower than Ethereum; the testnet can even achieve 100,000 TPS. This turns the concept of data as an asset from an ideal into a reality.
What I value more is its architectural thinking, which connects the data layer with the execution layer. There is no need for additional oracles, no need to change languages; developers can directly build complex logic in the EVM environment using Solidity. This smooth experience almost brings Web3 back to pure creativity. #KAITO
In my opinion, #Irys no longer belongs to the pure storage track; it is more like the data engine for the next generation of AI and autonomous applications. As more and more projects rely on high-quality, executable on-chain data in the future, Irys is likely to be that quiet yet crucial underlying gear.
We're in the final stage! MMTFinance content submission countdown, deadline on the 22nd! If you've written an article, remember to publish it in time!
This time, the collaboration between @MMTFinance and @buidlpad for the IPO is indeed quite friendly. Especially since you can participate by forming LPs with stablecoins (like suiUSDT-USDC), the threshold is low and the risk is relatively controllable. For me, this kind of participation, where you don't have to worry too much about price fluctuations, feels more solid. Many of my friends have already formed liquidity of over 3000U as per the guidelines. If you are also preparing to participate, remember to complete it before the snapshot on the 25th. I came to @buidlpad too late. I noticed that the platform has always had a keen eye for project selection, like the previous Solayer and FF, which performed quite steadily. This time, the first round FDV of $MMT is only 250 million dollars, which is a rare undervaluation opportunity in the current environment.