墨染web3

千呼万唤终于出来的 Backpack 股票 K 线图终于出来了。

对于绝大多数小伙伴们来说，看K线图是要比看折线图更加符合币圈特点。

而且也可以使用各种各样的指标，这样的话，我的那些家伙事儿都可以上了😂

之前搞个各种各样的指标系统，有了k线就都有用场了。 https://t.co/Q2M7kcOFRU https://t.co/pR9BsVaGu4

之前都说欧美的项目良心，国人项目拉垮。

不过最近这一大段时间，欧美项目的口碑似乎也不是很好，尤其是那些拖了不知道多少年还没发币的项目。

比如，MetaMask、Base、OpenSea、Polymarket。

.@dgrid_ai I previously made a video about this project.

Back then, I thought it was another concept of decentralized AI. But when I checked the data, I found out that they first earned money and only then prepared to issue tokens.

Seed round: $5 million. In the first half of 2026, they generated $23 million in revenue, with over 15,000 paid users.

🧵👇
1⃣ First, let’s talk about how that $23 million came about
Genesis membership: $1,580 in the first year, then $200 to renew from the second year onward.
15,000 people × $1,580— the numbers are basically pretty close.
Of course, the membership fees are real cash too; the payment address is a public Safe on BNB Chain.
In this sector, that’s relatively rare.

2⃣ So what exactly are you buying for $1,580
A monthly $300 model-call quota, one OpenClaw host machine, and a membership NFT on BSC.
The key is the NFT.
According to the official economic model, this batch of NFTs is tied to 25% of the total ten-year emissions of DGAI.
In a sense, this is more like a product presale with a token-related entitlement.

3⃣ Product side
AI Gateway is an OpenAI-compatible unified interface that routes to 200+ models—Claude, GPT, Gemini, DeepSeek are all included.
AI Arena lets users blind-test responses from two anonymous models; these human labels then feed back as training data for the router.
Plus DClaw local deployment, Model Marketplace, and Ask Dori. 50,000 DAU and 500,000 MAU.

The logic is like this:
Arena generates quality data → PoQ turns quality into verifiable signals → Gateway selects models based on those signals.

4⃣ My take
The reason DGrid proposed that AI infrastructure must be rebuilt—what they said last year—is even more true today than it was back then.
After AI moves from demos into agents and real workflows, cost, verifiability, and trust become the three most important things.

But I want to remind you:
Right now, the revenue structure and TGE are strongly linked—most of what members buy is, in fact, tied to token expectations.

One thing to watch is whether, after token issuance, the Gateway’s usage volume can actually support this whole narrative.

Project website: https://t.co/khLqZUSfC6

This post is only my personal view and does not constitute investment advice.

#DGrid #AI #Web3

我喜欢用 @Backpack 还有一个原因就是，我感觉我距离项目方距离不算远。

有啥问题可以直接联系到项目方，不需要通过一层层的渠道往上走。

这种应该是项目方那边的配置足够的扁平，所以用户的问题可以直接传到项目方的耳朵里。

.@42space 的 seed 玩法确实能赚到钱。

我这个投了 5u 赚了 14.9u，花了 50 张门票，这 50 张门票可以通过嘴撸的方式零撸。

赔率其实是很不错的，只是投入的金额太少了，最多只能投 5u。

我个人的建议是适当增加一下投入的资金量，这样赚的时候能更多一些，哈哈。

如果还没有参与的小伙伴们可以考虑参与一下：https://t.co/QDFC7uvx2N

钱包被盗之后，应该如何做？

先判断属于哪一种被盗，经常有人私聊我说钱包被盗了。

先说我的结论：
绝大多数情况下，那个地址就别再用了，直接换新号重开。把时间花在查原因上，比在旧地址上纠结有用得多。

但在重开之前，有几步顺序不能乱。

1⃣第一步，先分清是被动还是主动。
被动是指你自己没做错什么。
比如钱放在某个 DeFi 池子里，项目方被黑；或者早年指纹浏览器出漏洞，信息泄露被黑客抓走。
这种先去搜一下新闻，看看是不是哪个项目、哪个团队当天出事了，而你刚好参与过，基本就能对上。

这一类影响其实不大。
被盗的是项目方的池子，你自己的钱包是安全的，还能继续用。
至少冤有头债有主，知道问题出在哪，不用天天焦虑自己会不会再被盗一次。

主动的才麻烦。
私钥、助记词泄露，或者点了钓鱼链接、授权了带毒的合约。
这种要是不搞清楚原因，换个新地址照样会被盗。

主动这一类还能再分两种：
一种是很早之前助记词就在某个不经意的瞬间被拿走了，这种最难察觉；
另一种是你刚点了某个网站、签了某个授权，一般一小时之内就出事，这种你自己心里通常是有数的，比如点了推特上某个仿冒账号。

2⃣看被盗的是原生代币还是非原生代币。
这一步是我觉得最关键的判断。

单纯的授权钓鱼，一般只能转走非原生代币，也就是 ERC20 和 NFT 这些。

但如果发现原生代币也没了，那说明黑客拿到的是这个钱包的最高控制权，等于私钥或助记词已经泄露。
到这一步钱包就是彻底废了，没有抢救的必要。

3⃣取证。
要搞清楚两件事：我是怎么被攻击的，钱去了哪里。

追资金流向我一般用一个可视化的链上分析网站：https://t.co/dPY17Om9ui

支持 ETH、BSC、Polygon 这些，可以查全链。
把地址粘进去会出一个树状图，能看到这个地址的钱是从哪来的，是跨链过来的还是从 Uniswap 出来的，之后又转去了哪。
每一层还能继续点进去，一直往下追。

这个网站我最早其实是撸毛的时候在用，看自己几个地址之间有没有链上交集，判断会不会被判女巫。
拿来查被盗资金也一样好使。

4⃣分析完原因，放平心态。
有些事查到最后是链下的问题。我
之前遇到一个开工作室的朋友，链上翻了非常久，各个方向都找了，什么都没查出来。
最后的结论是他把助记词写在一个小本本上放在抽屉里，某次被别人偷偷抄走了。
这种情况在链上你是完全看不出来的。而且如果他不改这个记录习惯，他在换十个新号也一样会被盗。

也有查出来也没用的。
我自己被盗过一次，是早年玩跑鞋的时候项目方给我生成的地址，后来被证实那一批助记词在生成的时候算法就有问题，早就泄露在网上了。

所以我一直建议资金量大的话上冷钱包 。
中心化交易所也不是绝对安全，被黑和倒闭这两种风险都是存在的。

5⃣日常我们个人能做的其实就两件事：
一是定期清授权。
打开 OKX 钱包的授权管理，它会标出恶意地址、欺诈合约、长期未使用的授权，可以看全链。
我自己钱包里的授权攒了非常多，一般挑 Gas 便宜的时候统一取消一批。

二是装个能识别钓鱼站的工具，OKX 钱包、GoPlus 这类都行，打开有问题的网站会提前弹提示。
只是降低概率，但成本很低。

最后说一句：
黑客其实没有大家想的那么牛。
正面攻破一个保管得当的私钥并不容易，还是多努努力提高安全意识最重要。

钱包被盗后该做什么？时间顺序是核心，多数人第一步就做错了。 https://t.co/oX6rD3I3TF

现在全职撸毛不是很好过，其实倒不是币圈完全没有机会。

只是机会从链上转移到链下了。

这个链下包括：CEX、创作者、AI、嘴撸、所内活动等。

看到这个新闻再结合最近很多小伙伴的支付宝都被限制了额度。

其实就能看出来 U 卡还有无比巨大的空间。

但是 U 卡的问题也很明显，主要有两个吧：

1️⃣U卡大多不长久，一个长期稳定好用的U卡非常难得。

2️⃣U卡资金量仍然有限。一般消费还好，大额的话也不行。

解决这俩问题，U卡的使用量估计要翻好多倍了。

最近市场新热点不多，我刷 @predictdotfun 的时候也是回归了一些传统板块，比如政治、加密这一部分。

整体来说，我个人觉得预测市场差不多再刷一两个月就可以考虑发币的问题了。

很多项目需要 KYC 没问题，但是很多项目压根就不会妥善的处理拿到手的 KYC 数据。

大公司尚且也会出现信息泄漏的情况，小项目更是这样了。

尤其是币圈很多三无项目，上来就要 KYC 的，说不定哪天数据就泄漏了。

.@axisrobotics 的任务我感觉我手残，太难玩了😭

虽然我发过教程如何去操作，但是这么多任务做下来感觉还是有些难度的。

如果要多号的话，简直不敢想会有多么的痛苦。 https://t.co/O8Ml8LyRDf

Plasma U卡用满一个月，分享一下真实扣费、汇率损耗、拒付记录情况。 https://t.co/QbssUqX7dt

无论是币安广场还是 OKX 星球，其实给我的感觉都是圈地自萌。

有点像之前看《三体》里，星球把自己包裹起来，不能进也不能出。

逻辑是这样的，推特是一片大鱼池，写内容还是干嘛都有机会带动增量市场。

但是在 cex 内的社交，纯纯就是自己和自己玩了，这个圈子手牛熊的影响无比巨大，受交易所影响也是无比巨大。

受各位小伙伴们的帮助，我这个 Startale U 卡的候补名单排名已经排到了第 116 位了🤠

感觉拿到这一张卡的几率大大增加。

我感觉这一张卡和 Plasma 的卡比较接近，等拿到这一张卡第一时间出一个使用体验。 https://t.co/4RUEqcsCgl

最近这一段时间，我的重心放在了拿 @42space 的门票上。

拿到门票后可以换提前进入市场的资格。

我感觉这个玩法在早期应该是有的赚的。

唯一一个我觉得不好的点就是投入太少了，5 USDT 压根不够赚的，要是能多投一些就好了。 https://t.co/433lt3MA0m

币安主站内的网页版也接入了预测市场 @predictdotfun 。

看了一下也是 pre 提供的数据，这一点又进一步提高了 predict 的预期。

不过我看了一下，我用的日本节点是没法正常打开这个网站的，必须切换其他的节点。

但是 predict 网站本身是可以正常打开的。

估计是为了合规，交易所卡的更严一些吧。 https://t.co/NJipF5zCun https://t.co/OGb3fjKYB0

Gaining followers and ranking for xhunt are contradictory.

If you write some beginner-level, mass-audience content, it’s the fastest way to gain followers, but when the experts see it they think:
"Damn it—why are you sharing such low-quality beginner stuff?"

If you write some premium content that even experts read and can’t stop praising, the xhunt ranking will be the highest, but when new people see it they think:
"Damn it—what are you rambling about here?"

Twitter third-party creator platforms. I used them one by one over the past few days, and I’ll group them together to talk about Lighthouse, Xinhuo, and Twitter Self-Serve Cafeteria.

1⃣ First, Lighthouse @Lighthouse_2026
What I’ve used so far, it’s still the best one in terms of overall performance—but also the most “intensive” (most demanding).

2⃣ My own judgment is that Lighthouse is currently hitting a bit of a bottleneck:
The rules have been adjusted several times in the front end. Projects that are willing to post large-dollar deals have become fewer, with no new funds coming in. As a result, fewer people are willing to take orders, and liquidity has worsened.

3⃣ For the rules, I wouldn’t say they’re overly harsh—just slightly stricter.
The daily order-acceptance quota is calculated based on the成交量 (deal volume) from the past seven days:
If there are no completed deals for seven days, you can only take 10 orders per day;
If there are 1–4 completed deals, you can take 30 orders per day;
If there are 5 or more completed deals, you can take 50 orders per day.

However, if your goal is to promote your own account, I actually think Lighthouse’s tweet heating is quite suitable: likes, retweets, comments, or both like+comment can be selected, with a minimum starting investment of 2 USD.

Participation link: https://t.co/3MuCWXIfwr

4⃣ The second is Xinhuo @MingLin_Crypto
When you open Twitter, it shows as a personal account—I guess it’s being run by an individual.
But overall the experience is actually a bit better than Lighthouse: one reason is there are more orders, and another is the rewards are slightly higher—you can also see amounts like 0.5 and 0.15.
Task types include sponsored placements (name rights), custom bios, original tweets, and so on.
Participation link: https://t.co/eSVLEvo6jj

5⃣ The third is Twitter Self-Serve Cafeteria.
I only discovered it yesterday, and it highly imitates Lighthouse.

Right now there aren’t many tasks on it:
Publishing and pinning promotional tweets for the task plaza—complete both and get 42 USD;
Sponsored-name tasks: 6 + 15, total 21 USD.
Participation link: https://t.co/v7kPkPBDLv

If you have multiple accounts and can do a matrix, this one is worth doing.

5⃣ These three platforms share a commonality, and this is also the main “trap” I think you need to pay attention to:
The withdrawal threshold is at least 100 USD for all of them.
For the self-serve cafeteria, currently there are only these two tasks—no others. Even if you manage to get the rewards, you still can’t withdraw; you can only do tasks slowly and accumulate slowly.

6⃣ In my view, there are only two advantages to platforms like this:
First, whoever can make money goes where it pays.
Second, they must naturally form a tight-knit group of KOLs or a community coalition from the start, creating synergy.
On the “coalition” aspect, Lighthouse is definitely doing better right now. Although the effect isn’t that obvious yet, I believe this is the direction these platforms will take next.

7⃣ Also, all three have an invitation mechanism.
If you have many accounts, or can bring in some big accounts that are also doing things on Twitter, you can go through invitations to earn points or money—and they’ll also share some with you.

8⃣ Finally, there’s a fourth platform: @tuttihq
I can’t participate in it for now—I’m still on the waitlist.
So I don’t have a very good angle to recommend it to everyone; I also don’t know what’s inside looks like.
There are people promoting it on Twitter, but not many.
You can start by joining the waitlist. If you get in, go ahead. If not, then first do the first three.
Participation link: https://t.co/43yfsEVeNe

If there’s profit, it’s better than having none.

Twitter creator platform roundup: a one-time breakdown of the three platforms—Lighthouse, Xinhuo, and X Self-Help Canteen. https://t.co/N6JiiSpgoM