Jeonlees

认认真真撸毛

981Following

1.5Kfollowers

Feed

Axis Robotics has finally released the tokenomics for $AXIS. I think this is worth a serious look, especially for friends who have been doing tasks and tracking on the Hub, as well as those who recently participated in the Kaito event. After all the hard work, the main concerns are: how many tokens will we get? When will they unlock? And how are the team and investors' stakes arranged? I reviewed the latest Tokenomics and community sale rules announced by @axisrobotics. There are some interesting numbers and some easily overlooked details. 1. How are the 1 billion $AXIS tokens distributed? The total supply is fixed at 1 billion tokens with no inflation mechanism. The initial circulation at TGE is 165 million tokens, accounting for 16.5%, planned to be fully unlocked by the 35th month after TGE. The specific allocation is as follows: Ecosystem: 51.08%, about 510.83 million tokens Community Contributor Rewards: 10%, 100 million tokens Core Team: 15.2%, 152 million tokens Private A: 9.8%, 98 million tokens Private B: 6.67%, about 66.67 million tokens Liquidity: 6%, 60 million tokens Sonar Community Sale: 1%, 10 million tokens Kaito Reward Pool: 0.25%, 2.5 million tokens According to the official classification, the Network takes 61.08%, and the team and investors together hold 31.67%. At first glance, over 60% is allocated to the network, which is indeed a high proportion. But I think this should not be simply understood as all 60% going to the community. Because only 10% is clearly allocated as an airdrop to data contributors, while the 51.08% Ecosystem portion covers partners, research, ecosystem growth, treasury, and other uses. How these funds are used and when they enter the market is, in my opinion, more worth tracking long-term than just looking at the allocation ratio. 2. Team and private sale have zero unlock at TGE, but don't overlook subsequent releases Many projects have attractive tokenomics, but at TGE, all stakes release at once, and ordinary participants have no idea who is selling. Axis at least has a clearer timeline this time. Initial circulation at TGE is 16.5%, including: Ecosystem unlock 8% Contributor rewards unlock 2% Liquidity unlock 6% Sonar community sale unlock 0.25% Kaito rewards unlock 0.25% Core team and two rounds of private sales have zero unlock at TGE, locked for 12 months, then linearly released over 24 months. I personally approve of this arrangement, at least early on we won't face large-scale unlocks from the team and private sales. But don't just focus on this one advantage. T+3 and T+6 are two time points that need special attention. According to the overall release plan, contributor rewards will have 30 million and 50 million tokens entering unlock schedules respectively. At T+12, the team and private sales will also start monthly releases. Therefore, the early market needs to focus more on the actual use of ecosystem funds and the release of community rewards, while later considering the unlocking pressure from the team and investors. Of course, unlocking does not necessarily mean selling, but it's best to note these time points in advance. 3. Community contributors get 10%, which I care about the most Axis has reserved 100 million tokens for Contributor Rewards, accounting for 10% of the total, divided into two Seasons. Season 1: 40 million tokens, 4%. Mainly corresponds to Simulation data contributions before TGE. 50% of this unlocks at TGE, i.e., 20 million tokens; the remaining parts release 10 million tokens each at T+3 and T+6. Season 2: 60 million tokens, 6%. Focus will shift from Simulation to Sim-to-real, connecting simulated data with real robot scenarios. Specific participation conditions and reward rules are yet to be announced. Here you can see Axis's approach. They need a large number of users to contribute robot training trajectories via browser remote control early on, and later verify whether this data can truly help robots complete real-world tasks. This is also why I have been paying close attention to Axis. Robot training data is not as easy to obtain at scale as text or images, especially data involving actions, operation paths, and failure corrections, which have high collection costs. Axis aims to scale this through a distributed contributor network, then use quality verification and on-chain records to make the data traceable. If this model can continue to work, token incentives may truly connect with real data demand. But one thing to remind everyone: Axis Points currently are only factors that may be referenced for contribution records and reward eligibility, with no official fixed exchange rate. Having many points does not equal having locked $AXIS tokens, so don't confuse this. 4. Sonar community round, unlock rules have also been adjusted Friends who participated in the Sonar community round can focus on this part. The sale price was $0.10, corresponding to an FDV of $100 million, with total sales accounting for 1% of supply, i.e., 10 million $AXIS tokens. In September, Axis optimized the unlock rules for the community round. TGE unlock ratio increased from the original 10% to 25%. There is also a Community Bonus, distributing an additional approximately 5% to 25% token rewards based on participation, fully unlocked at TGE. The remaining 75% of the original allocation is locked for 6 months, then linearly released over 6 months. I think this adjustment is a good change for those participating in the community sale, at least increasing early available stakes. However, the reward calculation method and final individual allocation still depend on the official final announcement. For ordinary users like us participating in Hub and Kaito, I think there are three things worth focusing on at this stage. First, check your contribution records, account bindings, and reward eligibility to avoid previous efforts being affected by minor issues. Second, pay attention to the final distribution standards of Season 1 and the participation rules of Season 2. After all, the total community contributor rewards are 100 million tokens, and how they are distributed later is very important for participants. Third, after the TGE date is officially confirmed, recalculate the actual unlock amounts and circulating market value at each stage. Especially T+3, T+6, and T+12 nodes, it's best to add them to your watchlist in advance. Now at least we know how big the reward pool is and roughly how the stakes will be released. As for how much one can finally get and how much the market will value it after listing, we still have to wait for actual implementation. I still have expectations for Axis's robot data direction, after all, it addresses a real and difficult-to-scale problem. I have been participating in tasks before to see how far this project can go. But having an imaginative project direction does not mean the token price will necessarily follow that imagination. Especially between community contribution, data commercialization, and token value, there are still many links to be connected. Next, it depends on whether Axis can turn its accumulated data advantage into long-term demand. Also, a reminder: the official documents have stated that relevant numbers before TGE may still be adjusted, so please refer to the final announcement. The above is my personal summary and opinion, not investment advice, DYOR.

A project sets aside 1,000 whitelist spots—who should they ultimately be allocated to? Is it the whales holding hundreds of thousands of dollars in their wallets, the creators who wrote dozens of tweets for the project, or the quiet users who have been using the product for half a year? This question seems simple, but many Web3 projects still haven't solved it well. The most common approach used to be Galxe tasks, community lotteries, KOL collaborations, plus some wallet interaction thresholds. The result is familiar: studios participate in bulk, creators desperately boost exposure, and users genuinely interested in the product might not even get a spot. Recently, while studying @NucleusCodes Season 3, I found that what it aims to address is actually the chaotic user allocation market in Web3. Project teams spend money to buy traffic but find it hard to judge how much that traffic is really worth. This is also why I think Nucleus is worth researching. ▍01 Let's look at a real-world case Nucleus recently designed two leaderboards for HoodMiners, providing a total of 1,031 GTD whitelist spots. Among them, 871 are on the Reputation Leaderboard, and 160 on the Contribution Leaderboard. This allocation ratio is interesting. About 84.5% of spots go to historical reputation, 15.5% to current contributions. In other words, in this specific event, the project clearly prefers to reward users for their past on-chain behavior rather than just who posted the most recently. Reputation mainly analyzes wallet transaction history, NFT holdings, token holdings, covering ecosystems like EVM, Bitcoin, Solana, etc. Contribution evaluates original content during the event, project relevance, genuine interaction, and dissemination effect. These two systems correspond to completely different values. One measures assets and participation records a person already had before the event started, and the other measures the actual promotional value a person created for this event—finally, they can be measured separately. ▍02 Why separate these two? Suppose a new NFT project is preparing for launch. It may need collectors with purchasing power and content creators who can explain the project clearly and bring new users. But these two groups are often not the same. Some whales rarely tweet but have long participated in various NFT trades; some creators have limited funds but are very familiar with a certain ecosystem and can continuously produce valuable research. If you only look at assets, the project might miss excellent promoters. If you only look at social heat, rewards might be taken by traffic farms. Nucleus's dual leaderboard mechanism at least provides project teams with a more detailed filtering method. It also has an easily overlooked advantage: different projects can choose different users based on their goals. NFT projects might value collection history more, DeFi projects might focus on real on-chain usage records, and early-stage products might need contributors willing to provide ongoing feedback. This means the value of the same user in different projects shouldn't be exactly the same. ▍03 But I think Nucleus's real challenge is just beginning Putting wallet behavior and social data together doesn't mean you can accurately measure a person's value. On-chain data can prove what a wallet has done but may not prove the true intent behind those actions. A wallet with large transaction volume might be a seasoned trader or just repeatedly washing volume. An account with high interaction might truly have influence or might have a mature mutual-like network. A more realistic problem is that if reputation increasingly affects whitelist and airdrop eligibility, users will start actively studying the scoring rules. When everyone changes on-chain behavior to get high scores, the data originally used to identify real users will gradually be influenced by the incentive mechanism. To me, this is the hardest problem for reputation systems: **When a metric can be directly exchanged for benefits, to what extent can it still reflect true value?** Nucleus currently raises the cost of score manipulation by limiting daily scoring tweet counts, lowering scores for low-quality AI content, and reviewing interaction authenticity. These measures are indeed necessary, but whether they will be effective long-term remains to be seen. ▍04 I look forward to Nucleus making reputation something that can be continuously accumulated There is a very wasteful phenomenon in Web3 now. A person accumulates two years of transaction history on a chain, participates in governance, and helps solve community problems. When moving to a new project, they often have to redo tasks and prove themselves again. If Nucleus can continuously recognize user behavior across different ecosystems and convert these historical records into filtering criteria usable by different projects, its value will exceed that of a single event leaderboard. For users, past contributions could influence future participation eligibility. For project teams, it could reduce repeated customer acquisition and ineffective incentive costs. However, one judgment must remain: a high reputation score does not necessarily mean the user will stay long-term in the project. What truly determines how far Nucleus can go is whether project teams can achieve better retention, genuine usage, and long-term participation from these filtered users. This is more important than how many people are on the leaderboard or how many tweets they posted. I will continue to follow Season 3, but what I really want to observe is, after the event rewards end, how many users who qualified through Nucleus actually stay. If this data can be produced, Nucleus @NucleusCodes's reputation mechanism will truly have proven its value. For personal research and observation only, not investment advice, DYOR.

Everyone keeps saying @axisrobotics is about to have its TGE, but I see there are 3 phases of activities on Kaito. The activities won't end until at least November 17. But those who need to bind their information should hurry up. If you haven't done the tasks yet, quickly work overtime to get them done. For those logging in with Google, the binding might not show up at first. Just refresh and wait, it will be fine. For other issues, I see everyone in the Discord group is actively solving them. Everyone should check early. Looking forward to surprises from Axis!!!! There have been quite a few good news from Axis recently. >>>>>> On October 8, Axis secured a noteworthy automotive manufacturing partnership: Lotus Cars and Geely. I think this progress is much more important than Hub adding hundreds of thousands of new trajectories, because Axis is finally starting to push the data capabilities accumulated in the past into real industrial scenarios. Let's talk about what the cooperation specifically involves. Sorting parts on the automotive production line looks simple but is actually very troublesome. Hundreds of parts, different shapes, sizes, and gripping positions. Traditional solutions often require writing rules and debugging actions one by one. When switching to a new batch of parts, you have to start over. Axis's proposed solution this time is to first replicate the real production line's simulation environment, then use Axis Suite to generate tasks and train robot policies, and finally deploy them to the physical production line. The goal is to have one policy cover 500–800 types of automotive parts. Note, this is still the goal of the cooperation, not an achievement already realized. But if later verified successfully, it will be very valuable for industrial robots. Because what many robot companies really struggle with is no longer just whether the model is smart enough, but where the training data comes from, how to reduce costs, and whether it can still be used when switching scenarios. This exactly corresponds to the Compounding Data Engine that Axis has been working on. In the past few months, I have been following its data growth. Hub has already accumulated about 6.83 million operation trajectories and 220,000 users, and the supported robots have expanded from Franka to OpenArm, ARX, and more. But I believe 6.83 million trajectories alone do not prove Axis's success. What really matters is how much these trajectories improve the model. Axis's publicly released Sim Dataset V1 contains 207 operation tasks and over 50,000 human demonstrations. In LIBERO-Plus testing, using this data to continuously pretrain π0.5, the model success rate increased from 83.9% to 88.8%. This result at least proves one thing: through community-collected and filtered simulation data, there is indeed a chance to improve robot model performance. And Axis's data logic has a second half. Users remotely operate robotic arms in the browser, trajectories are verified and signed on the Base chain, then enter the training process. After model training, based on failure cases and weak tasks, the next round of needed data is decided. Collect, train, correct, then collect again. If this cycle can keep running, Axis won't have to start from scratch accumulating data every time it encounters a new scenario. The richer the data, the more efficient subsequent training might be. Of course, there is still a long way from LIBERO-Plus test results to real automotive production lines. Can the success rate in simulation transfer to physical robots? With hundreds of parts, how accurate and stable can the gripping be? These questions are what Lotus / Geely really need to answer in this cooperation. >>>>>> Now about Kaito Axis's Epoch 2 continues until October 18, with a maximum of 10 submissions per round, and the best 6 count towards the score. Scoring uses Mindshare × Referral Multiplier; invited users must actually complete tasks and sign on-chain; simple registration does not increase valid trajectory contributions. The entire event allocates 0.25% of $AXIS: top 100 share 0.10%, ranks 101–500 share 0.10%, KAITO stakers share 0.05%, settled by final ranking across all Epochs. Participants should also note the requirement to bind Discord and X before October 10. What I most want to continue tracking is actually the real results of this automotive factory cooperation. Axis already has millions of trajectories and has produced datasets that can improve model success rates. If it can deliver results on real production lines next, its data engine will have truly crossed the important step from research validation to commercial application. @axisrobotics DYOR.

Zeru @zerufinance's Ditto trading competition is underway, from October 1 to 29, the top 600 will receive rewards, with a total prize pool of $50,000. However, this $50,000 is not guaranteed to be fully distributed; it depends on the final trading volume. The entire participants' cumulative open position volume must reach $50 million for the prize pool to be fully unlocked. The $12,000 marked for the first place also follows the unlocking ratio: • 100% unlocked: receive $12,000 • 40% unlocked: receive $4,800 Registration is free, but you must first register for the competition on Zaps, then trade through your own Ditto link. Trades before registration do not count, and perpetual contract closures do not increase scoring volume. If you are already trading, you can take a look; if you want to increase volume for the prize, you need to consider fees and losses first. // The rules clearly state that wash trading will not receive rewards. This relates to Zeru's core business. Its zScore analyzes wallets' past transactions, lending, LP, and staking records, giving a behavior score from 0 to 1000, which protocols use to screen users, allocate rewards, or assess credit. But the competition ranking uses the current Campaign Zaps; old wallet zScores cannot be directly used as competition points. According to project disclosures, zScore already covers over 320 million wallets. The coverage is quite extensive, but for users, it still depends on what benefits this score brings. // Let's talk about airdrops first, which is easier to understand. Two wallets both performed ten interactions; one used normally for a year, the other completed tasks concentrated a few days before the snapshot. If scoring only by interaction count, both might get similar rewards. Studios creating dozens of accounts in bulk could take a larger share. Zeru analyzes wallets' longer-term records together. When projects distribute rewards, they can reference past lending, liquidity, and continuous usage, not just how many tasks were done this round. For old users, past activity can still be useful. Of course, a low score doesn't necessarily mean a bad actor. New users have little history, and normal users might spread funds across multiple wallets. While filtering out bulk accounts, these users should not be blocked together. // Lending is even more complex. Traditional credit reporting carries past lending and repayment records to the next application. On-chain records exist too, but switching protocols doesn't necessarily grant lower rates or less collateral. Zeru aims to organize history to give protocols more basis to judge borrowers. But timely repayments in the past might be because collateral was still there; failure to repay leads to liquidation. In the future, if less collateral is posted, the borrower's default cost changes. Good past performance with full collateral doesn't guarantee reliability with less collateral. So a high score can be a reference, but how much collateral can be reduced or credit extended still requires actual loan and bad debt data support. Currently, a high score found cannot be directly interpreted as qualification for low-collateral borrowing. // Zeru has received strategic funding from Eigen Foundation, with the scoring infrastructure running as EigenCloud AVS. According to the project introduction, the aggregation and commitment of scoring results are proven by decentralized operators, then distributed cross-chain, allowing other protocols to integrate and use. This part is responsible for verifying and transmitting scoring results. Whether the model's judgments are accurate depends on actual outcomes. Currently, I am more optimistic about them first doing well in airdrop filtering and reward distribution. With the same budget, can more be allocated to normal users, less to bulk accounts, and fewer new users mistakenly penalized? These results are easier to verify. Lending has greater potential, but if judged incorrectly, the fund pool bears losses, so it needs gradual validation. If you have old wallets, you can check once on Zaps. Previewing zScore is free; minting the scoring NFT is optional. First see if existing records are recognized, then see which protocols accept this score. There's no need to pay fees in advance to boost scores.

These past two days on X, quite a few people must have seen that 🪑. Some have put the chair emoji in their nicknames, some are showing off their “Big Chair Energy,” and a bunch of people have started applying for the TMF Pass. @TheMutualFun only released 200 Passes this round, so it’s no surprise that there’s suddenly a lot of chatter. The Mutual Fun started by designing something more concrete: If the fund itself is born on-chain, does the fund share have to be just a number? The answer it gives is a Seat. The whole system has only 4,001 Seats, divided into 5 funds. The Seat itself is a pixel avatar, but the really useful stuff is behind it: It corresponds to a fund, a weekly voting right, and an ERC-6551 Briefcase that belongs to the Seat itself. Simply put, each Seat has its own on-chain account. Dividend checks issued by the fund can be stored inside, and more uniquely, when the Seat is transferred, the Briefcase goes along with it. This means two Seats that look similar might not have the same price. One might have just enrolled and have nothing inside; Another might have been running in a well-performing Fund for a while, with assets accumulated in the Briefcase. Buyers are no longer just buying “how rare this NFT is,” but also the account status that this Seat has accumulated. Many NFTs’ so-called Financial Utility is essentially “holding this thing lets you claim some benefits elsewhere.” TMF directly embeds the financial relationship into the NFT itself: Seat is the seat, Briefcase is the account, after enrolling you have fund rights, and you can participate in Proxy Ballots weekly. And it doesn’t fully hand over fund management to the so-called “fund managers.” Holders vote weekly to decide how to allocate money and handle profits. The five Funds start with similar rules but can gradually evolve into five different styles based on holders’ choices. If the five Funds really develop different positions, returns, and community styles, then the pricing logic for Seats won’t just be the usual NFT “rarity + floor price.” You’ll also have to consider which Fund it sits in, how that Fund has invested in the past, and how it has performed. Beyond that, there’s a pretty wild M&A system. Two Seats that meet conditions can merge, leaving one higher-level Seat while the other permanently exits. Combined with a Transfer Agent: one Seat corresponds to 401,000 $TMF; Seats can be deposited to exchange for $TMF, and $TMF can be exchanged back for Seats at a fixed amount. So from the start, it’s doing one thing: making Seats increasingly difficult to price simply. Why do I think it’s worth participating now? Seats don’t have a long fund history yet, the five Funds haven’t really diverged, and the Briefcases haven’t accumulated enough asset records. Whether M&A will make some Seats scarce is still unknown. So this time I will seriously apply for @TheMutualFun’s TMF Pass. That chair 🪑 on X has been really hot lately, but if it were just because everyone’s showing off chairs, I wouldn’t bother reading a Prospectus. What really makes me want to sit down for a while is my curiosity: When these 4,001 chairs really start managing money, will the market ultimately treat them as NFTs or as a kind of on-chain fund seat never seen before? @themutualfun @Kingpickle

If you happen to have an idle phone, you can actually try this activity casually. @Acurast turns your phone into a Processor, providing computing power to earn $ACU, and later you can also stack points and staking gameplay. I'll break down the process directly; just follow it basically. 1️⃣ First, see if your phone is suitable for Lite or Core If it's a phone you still use regularly, go with Lite. Android 12+ or iPhone 6s and above, iOS 15+ are all supported. Just download and install Processor Lite, and back up your mnemonic phrase. Android also requires setting up a work profile, then keep the app connected to the internet and charging while running. After setup, I recommend scanning the QR code to bind it to Acurast Hub, which makes device management much easier later. If it's an old Android phone you no longer use, consider Core. Core basically turns the entire phone into a dedicated Processor, with a higher reward cap, but be very careful: it will erase all phone data. The process is to first factory reset, then stop at the phone welcome screen, generate a QR code in Acurast Hub, then quickly tap about 8 times on the welcome screen to bring up the scan entrance. After scanning, the phone will be locked as a dedicated computing device for Acurast. So definitely don’t try Core on your main phone, especially if it contains photos, wallets, Authenticator, or other important data. Both methods start from Acurast Hub: 2️⃣ Prepare a wallet and a small amount of $ACU You can use Talisman, SubWallet, or AirGap wallets. Once the device is running, it will send online heartbeats regularly, about every 30 minutes, so it’s best to keep some $ACU in the account to pay related fees. Here’s a common confusion: Acurast mainnet uses $ACU, while the Canary testnet uses cACU. Don’t treat them as the same coin. 3️⃣ After connecting your phone, how are rewards calculated? It’s not a fixed amount for keeping the phone on all day. Acurast considers the device’s hardware benchmark, online heartbeat, and whether your Processor matches actual deployed computing tasks. So different configurations, online times, and task situations will result in different earnings. That’s why I recommend trying with truly unused old phones. Electricity and network costs are real expenses, and earnings are not fixed. Run it for a while to see your device’s actual performance rather than relying on so-called “daily earnings.” 4️⃣ Besides running the phone, you can also participate in Cloud Rebellion Acurast also has Cloud Rebellion, where you can join tasks via Discord invite codes and accumulate MIST by completing tasks. But be sure to distinguish that MIST is event points, which can involve season rankings, merchandise rewards, etc., and it is not equal to $ACU. 5️⃣ After having $ACU, you can also participate in staking computing power Acurast has a Staked Compute system. People providing computing power with phones can lock $ACU and commit their online computing power; those without phones can delegate their $ACU to providers and share in the earnings. Currently, about 70% of inflation rewards go into this part. If you just want a low-threshold experience, I suggest starting with Lite. If you happen to have an old Android phone you definitely don’t need, then consider Core, since Core will erase data and turn it into a dedicated device. Finally, a reminder: always keep your mnemonic phrase offline, make sure data is backed up before using Core, and try not to use your main phone that stores wallets and crypto assets for this. This is just a personal experience record; specific rewards and rules are subject to the latest official Acurast page. DYOR.