TLDR about the CLARITY Act
The Digital Asset Market CLARITY Act just cleared the Senate Banking Committee on a bipartisan vote.
It is the furthest crypto market structure legislation has ever gotten in the US.
What it does:
- Creates the first comprehensive US regulatory framework for crypto
- Draws the line between SEC and CFTC jurisdiction (which assets are securities vs. commodities)
- Gives blockchain networks a legal pathway to launch and operate in the US
- Sets oversight rules for crypto exchanges and consumer protections for trading
Why it matters:
Existing US law is built for companies (centralized control, shareholders, intermediaries extracting value).
Blockchains enable networks with distributed control, value flowing to users at the edges.
Forcing networks into corporate-law frameworks contorts them into the very thing they're designed to replace.
CLARITY is meant to fix that mismatch.
Path so far:
- FIT21 passed the House in 2024
- House CLARITY passed July 2025
- GENIUS Act (stablecoins) passed July 2025 → triggered the current adoption wave
- Senate Banking advanced its version May 14, 2026
What's next:
Senate Banking + Senate Ag committee bills get merged → full Senate vote → House approval → President's signature
今天又一起链上 DeFi 无抵押铸币攻击事件，Monad 链上被盗 1,000 枚 eBTC ($7645 万)。但由于 eBTC 极差的流动性让黑客没办法把 eBTC 换成其它资产，最终黑客只得逞了 $86 万。
单从这一点来说，没有流动性反而救命了？！😂
时间线
1⃣5:21：Monad 链上的 Echo Protocol 被黑客无抵押铸造了 1,000 枚 eBTC ($7645 万)。
2⃣5:22：黑客把 45 枚 eBTC 存进借贷协议 Curvance 然后借出 11.3 枚 WBTC ($86 万)。不是黑客不想借更多，而是这已经是能借的全部了。
3⃣6:39：11.3 枚 WBTC 通过跨链兑换工具换成 385 枚 ETH 后通过龙卷风洗走。
4⃣DEX 池子没流动性、借贷平台上仅有能借的资产也被他借空了，黑客手上剩余的 955 枚 eBTC 没地方退出，只能干放在钱包里。
5⃣10:37：这得以在事发 5 个小时后，出事项目方 Echo Protocol 通过合约调用把黑客手上剩余的 955 枚 eBTC 给销毁掉。
最终损失就是从 Curvance 上借走的 11.3 枚 WBTC ($86 万)。
黑客地址：https://t.co/U2M5ryuo8v
--------------------------------------------------------------------------
#Bitget 来了就是VIP！Crypto、美股、CFD，全球先机一站布局
清掉了几乎所有的仓位，短期不太看好后市
踏空了所有的美股的涨幅，但是也不会追高
做做套利赚点钱，不下牌桌比什么都重要
偶尔玩玩高控妖币，但是机会也越来越少。标的越来越少+狗庄们被监管盯同时又被散户和ai盯
最近每天早睡早起（10点睡8点起）
多读毛泽东、多拥抱AI
专心和团队一起在熊市打磨基建，有更多时间去思考
张一鸣在27的时候发的：
“快到30岁了，感觉这几年又再重新学习/补习本应在青少年时间学习的东西：如何阅读、如何了解自己、如何与人沟通、如何安排时间、如何正确的看待别人意见、如何激励自己、如何写作、如何坚持锻炼身体、如何耐心......”
牛市向外求，熊市向内求
人生的容错有很多，永远都不会太迟
《论持久战》毛泽东评价抗日战争：
"亡国论者看敌人如神物，看自己如草芥；速胜论者看敌人如草芥，看自己如神物——这些都是错误的。"
放到现在个人当下的感受：
亡国论：现在才开始学AI已经来不及了。
速胜论：Dan Koe美国成功学，3天彻底改变你自己
以太坊基金会6名核心成员公开离开，币都卖完了？一个月内至少6名核心成员公开离开，让大家多少都有些猜疑。
自26年以来以太坊基金会一直在转型，3月发布了新《Mandate》，强调从“全能推动者”转向更专注的核心协议守护者，优先CROPS原则。这波离职潮只是自之前联合执行董事Tomasz Stańczak、Josh Stark后的延续。基金想少管生态商业化，把精力放在L1硬化、Glamsterdam升级这些长期事情上。
这些人的离开的节点，正好是EF在“瘦身”和重新定位的时候。Protocol Guild已经独立运作，部分权力和资金在下放，这本身就是去中心化在组织层面的体现。
以太坊从“基金会主导”向“生态自治”转型的必经阶段。离职的人带走经验，留下的人继续往前推。只要协议本身健康、社区还在，Ethereum就还能继续当“世界计算机”。
The US-China AI race is a race no one can win and no one can afford to lose.
Every great power competition in history was for territory,
resources, or weapons.
This one is the first that is for none of them. It is a race for the substrate of intelligence itself. The stakes are not geopolitical anymore. They are ontological.
And here is the trap. Intelligence compounds exponentially. Slow down for a year and the other side does not slow down with you. They compound past you. A six-month lead becomes a decade gap. A decade gap becomes uncrossable. Whoever gets to ASI first owns it forever.
No nation can accept that.
So neither side can ease off. Not for a quarter, not for a month, not for a week. They have to run, harder and faster, forever.
And the game itself cannot end. Intelligence has no finish line. AGI is not the finish, it is the inflection. ASI is not the finish either. There is no finish. There never was.
The greatest game ever played, with no whistle, no winner, and no end. The universe is using both sides to make itself more intelligent.
It’s all part of the Universal Code.
The Universe's first Law is that is solves for output of intelligence per until of energy. You might not be able to see it yet but that is the Great Game...
花了些时间搞懂了https://t.co/NYI4VAQ8IH这个新台子的机制和玩法，我自己对meme加杠杆这件事也感兴趣，然后搞懂了台子的机制才能去它的台子玩;
核心:https://t.co/NYI4VAQ8IH台子的币等于普通meme币+自动杠杆期货，也就是它是pump+hyperliquid杠杆的混合体;
以往普通pump的币靠大家的买卖推动价格，而https://t.co/NYI4VAQ8IH台子的币除了自然的买卖推动，它还自动挂钩了hyp上的杠杆永续合约，这个后面段落会解释;
那么说明这个台子的币其实有两种力量同时驱动：
第一是纯靠买卖的力量推动价格，第二是在hype上开的杠杆期货的表现。那么这样的话说明这个台子的币没人买卖也能自己涨跌;
那它这是怎么实现的呢？
关键在于官网上说的这个Leveraged token，翻译过来就是非清算杠杆代币，下面统一简称为LT。在这个台子的币每枚都可以有一个LT作为背书，它就是一个hype上永不爆仓的杠杆仓位;
比如你创建新币的时候，同时在btc、eth、hype、sol等等基础代币中任意选一个币，然后选择2倍3倍5倍做多或者做空它;
接下来平台就会用usdc自动开一个对应杠杆的仓位，然后包装成LT，再把LT绑定到你创建的meme币上;
那么这样的话这个台子就有了一个关键机制——永不爆仓:
普通杠杆亏到一定程度就会爆仓清零，而https://t.co/NYI4VAQ8IH的LT亏了就会自动降杠杆保住本金，赚了就会自动加仓维持杠杆，所以结果就是永远不会清零，说白了就是它会根据行情自动加减杠杆;
那它的价格怎么算的？
它有价格双引擎，首先这个台子上的币和pump一样，有人买就涨有人卖就跌，这是第一层。其次加上LT的机制，如果你底层2倍perp的币涨1%，那么你的LT就涨2%，这个取决于你的杠杆;
举个例子:
假如你创建了一个币叫 $pepe ，然后LT做的是5倍做多 $btc ，初始市值如果是4000美金，那么btc现货涨2%，你创建的 $pepe 会自动涨10%，因为你的杠杆是5倍;
与此同时如果有人疯狂买你的 $pepe ，那么这股力量会再推高一波 $pepe 的价格。反过来如果btc跌2%，那么LT机制会自动减仓降杠杆，你的币会跌10%，但不会直接归零;
那风险也不是没有，风险就是这种机制在震荡行情的时候会非常难受，因为LT机制会来回加杠杆或者减杠杆。还有就是单边暴跌的时候会加速你的仓位无限趋近于0，即使它不是真正意义上的爆仓;
怎么评价呢，其实它不算新机制也不算创新，这种玩法以前交易所和公链大战的时候玩过。现在台子多如牛毛，搞清楚了机制和玩法以后也可以尝试去玩龙头标的;
今天中美会谈感觉获得信息不是很多，对于伊朗中国并没有完全表态，只有川普表示习近平愿意支持伊朗开放霍尔木兹海峡，这倒是在昨天的预期当中，毕竟霍尔木兹海峡长时间的封锁也不符合中国的利益，如果能用支持通畅来换一些美国的资源对于中国来说也是好的。
现在霍尔木兹海峡的通畅情况仍然堪忧，虽然获得一些伊朗的消息，昨天已经有超过30艘船只被通行，但同一时间还有印度商船被攻击的事情，现在霍尔木兹海峡的封闭仍然迫使 WTI 和 布伦特维持在高位，今天唯一好一些的就是资金费率下来点，估计是多数小伙伴抗不住了。
另外今天 CLARITY Act 终于在参议院金融委员会通过了，接下来就要送去参议院整体投票了，目前是需要 60 票才能通过，就意味着除了共和党要全数同意外，还需要民主党有最少 7 票。稳定币生息和道德规范条款仍然是现在争议的重点。
回到 Bitcoin 的数据来看，本来今天的价格走势并不是很好，但 CLARITY Act 的通过还是刺激了 $BTC 价格的走高，CLARITY Act 对 BTC 最大的好处是让 BTC 从可以买的风险资产，变成可以被银行、券商、基金、托管机构、清算体系、保证金体系长期纳入的合规数字商品。
短期是情绪利好，中期是合规流动性利好，长期是 BTC 金融化和抵押品化的利好。
#Bitget 来了就是VIP！Crypto、美股、CFD，全球先机一站布局
Hyperliquid YES ✅
Hyperliquid L1 单日暴涨 9.72%，成为今天全链资金流入冠军
截止中午 12 点 Hyperliquid L1 的 TVL 为 $1.57B，24 小时涨幅 +9.72%，是 459 条链中表现最强的公链
个人猜测这波增长可能与 CRBS 相关政策或市场预期有关
几个可能的原因：
- CRBS 框架下，Hyperliquid 作为去中心化永续合约交易所，可能被视为更合规的链上衍生品基础设施
- 机构资金在寻找合规的链上交易场所，Hyperliquid L1 的原生区块链架构提供了更好的性能和透明度
- HLP 主网上线后，生态资金流入明显加速，TVL 从 ~$1.4B 快速增长到 $1.57B
数据显示，Hyperliquid 的增长不是孤立的，Morpho Blue 在 Hyperliquid 链上的 TVL 也同步增长 +1.36%，显示整个生态在回暖
与此同时，其他公链的表现相对平淡：
- Provenance +2.11%
- Bitcoin +2.08%
- MegaETH -3.72%
所以现在的情况是，Hyperliquid L1 单日暴涨领跑全链，可能与 CRBS 相关的合规预期有关，资金在往更透明、更合规的链上衍生品平台流动
Uniswap v4 Hook is one of the most interesting DeFi experiment rn 👀
The first wave obviously looks speculative. A lot of the attention came from fast-moving Hook-native plays, and most of them are probably not long-term assets.
But the reason I’m paying attention is simple:
liquidity is becoming programmable.
Over the past few weeks, Hook-native projects started pulling attention back to the Uniswap ecosystem.
> @sato_hub / $SATO crossed a $ 40M+ market cap.
> @uniqepv4 / $uPEG went from almost zero to over $30M in less than two weeks.
> SLONKS by @MichaelHirsch generated 1,200+ ETH in trading volume.
> @lo0pio / $LO0P started experimenting with AMM + lending.
> @flood_markets opened up the idea of routing pool reserves into yield.
At first glance, this feels like another degen rotation.
I think the more important signal is what these experiments reveal about Uniswap v4.
- V2 made AMMs simple
- V3 made liquidity more efficient
- V4 changes how liquidity behaves
Before v4, a pool was mostly passive. Traders swapped, LPs provided liquidity, and the pool followed fixed rules.
With v4 Hooks, developers can attach external logic directly to the lifecycle of a pool. That logic can run around swaps, liquidity changes, and other pool actions.
So a pool is no longer just a place where assets sit and wait for traders.
It can now have its own behavior.
A pool can support bonding curve launches, dynamic fees, limit orders, TWAMMs, LP automation, MEV-aware execution, NFT/game mechanics, yield-bearing reserves, or even lending markets inside the liquidity mechanism itself.
This is where the v4 architecture matters.
V4 uses a Singleton design, where pools live inside one unified contract instead of every pool needing a separate deployment.
It also uses flash accounting, where balance changes are tracked throughout a transaction and only settled at the end.
That makes complex routes, multi-hop swaps, LP adjustments, and Hook interactions more efficient.
Lower gas is useful, but cheaper execution is only the surface.
The deeper story is that Uniswap is becoming a permissionless AMM plugin layer.
> $SATO pumped because its tokenomics are embedded directly into the market mechanism.
- ETH goes into the Hook
- The Hook mints SATO through a bonding curve
- When users sell, tokens are burned and ETH is returned
- The mechanism itself becomes the product
This is different from the old token launch model.
Usually, a project launches a token, adds liquidity, then adds staking, buybacks, or incentives later.
With Hook-native assets, the order changes:
design market behavior first, attach liquidity second, then let every trade interact directly with that mechanism.
> $uPEG showed another side of the same idea.
It did not need the most complex financial design. It worked because it turned a Hook into a simple cultural object: Uniswap lore + JPEG culture + v4 mechanics.
A swap was no longer just a swap.
It could become an onchain object.
> SLONKS showed the NFT/game side with cheap minting, AI-style art, Hook mechanics, and a fast social feedback loop.
> $LO0P moved in a more DeFi-native direction.
The pool starts to look less like a simple swap venue and more like a system that combines an AMM with a credit market.
Users buy tokens along a curve, lock those tokens, and borrow ETH from the same system.
This is why the first wave of Hooks looks like memes.
Memes are the fastest way for a new primitive to become legible to the market.
But the design space underneath is much wider:
Dynamic fee markets
TWAMMs
LP automation
Limit orders
MEV-aware pools
Yield-bearing reserves
Hook-native lending
Liquidation auctions
NFT/game mechanics attached directly to liquidity
Public dashboards like Dune are already tracking Uniswap v4 activity across multiple chains, including TVL, fees, DEX volume, and hook pool data.
The numbers are still not perfect because certain hooked pool events can be under-reported or over-reported.
Still, directionally, the trend is clear:
developers are now experimenting directly at the liquidity layer.
That also changes how I would research these assets.
For normal tokens, people usually check supply, tax, owner, liquidity, holders, and unlocks.
For Hook-native assets, the mechanism matters just as much as the chart.
I would want to understand what the Hook controls, whether it takes custody of assets, whether fees can be withdrawn, whether admin controls exist, how reserves are accounted for, and what happens at terminal states.
In other words:
the mechanism is the edge, but it is also the attack surface.
Opportunity and risk sit in the same place here.
Most Hook tokens will probably die. There will be forks, attention games, broken designs, and one-rotation charts.
But the primitive itself is worth paying attention to.
Uniswap used to be mainly a swap venue.
With v4, it is expanding into programmable liquidity middleware.
Hook builders can borrow Uniswap’s settlement layer, router network, LP base, brand, and distribution, then add custom market behavior on top.
That is a much bigger shift than just a “new meme meta”.
The next DeFi primitive may not need to launch as a new L1, appchain, or standalone protocol.
It may start as a custom Uniswap v4 pool.
Pool = market
Hook = behavior
Liquidity = programmable state
That is why I’m paying attention to Hook Summer.
Not for every ticker, for the design space it opens up 💎
估计最近两天 $BTC 现货 ETF 的数据都不会太好，主要就是因为 Bitcoin 价格的震荡，而和 Bitcoin 相反的是美股在继续走高，当然走高是 AI 和科技股，但 BTC 的反应相比美股就开始岔开了，我个人的判断是短期持有者认为在美国通胀数据不利的情况下，BTC 面对的流动性压力会大于 AI 相关的科技股。
不过从数据来看周二在公布了 CPI 数据以后美股也是出现了下跌，虽然周三就涨回来了，但 PPI 的数据确实不利，而 BTC 价格的反应很有可能就是和 PPI 相关，其实在刚刚公布 PPI 数据的时候美股也是出现了下跌，但很快就在 FOMO 的情绪下反弹回来了，而 BTC 就没有这个待遇了。
#Bitget 来了就是VIP！Crypto、美股、CFD，全球先机一站布局
🚨巴克莱统计显示，1951 年以来七位美联储主席上任后六个月内，标普 500 平均回撤约 16%。
Crypto 和 $BTC 这边更是扎心：
👉2014 年耶伦上台 → BTC 从1000到170，跌幅 84%
👉2018 年鲍威尔上任 → BTC 从17000到3200，跌幅 73%
👉2022 年鲍威尔连任 → BTC 从48000到19000，跌幅 61%
三次换帅（含连任），三次腰斩起步，命中率 100%。
有点怕，咱这次能幸免不！😂
为什么我说 hook v4 season 远没有结束，反而更要抓住这段盘整期？
因为历史上每一次出圈的热潮，都会经历冷静期——而冷静期之后，往往是更大的浪潮。
拿久一点的 2017 年 ICO 来说，最早出圈是在 4、5 月，但真正的高潮要到 8 月底才到来，期间造就了无数 A9、甚至 A10。
后来的 NFT 热潮、铭文热潮，走的其实都是同一条路径。每隔几年就来一次的这种热潮，才是币圈真正独特的地方，也是它制造暴富机会最密集的地方。
那为什么每次都会有这样的周期？
1. 链上创新永远是最纯粹的币圈文化，而它每隔一阵就会被某个意想不到的事件引爆——比如这次 v4 hook 被 uPEG 带出圈。
2. 每一波热潮要真正成为一个赛道，都需要一两个打响第一枪的项目，它们会带来可观的造富效应和传播效应。
3. 紧接着，大量流水盘、仿盘会迅速涌出——照抄就行，成本极低。它们本质上只是蹭热度的 PVP；等这批涌入的热钱被割，连最初引爆热度的创新盘也会跟着回调一波。
4. 但真正的高潮，一定出现在热度稍稍退却之后的第二波。道理很简单：
打响第一枪的项目给优秀的 DEV 立了样，而好 DEV 需要时间去打磨下一批真正的好项目。
5.当这批天才 DEV 带着真正的新 idea 再次出现时，会引发比第一波更大的财富效应。
6. 再加上基建、交易所、钱包都需要时间跟进支持，第二波热潮项目的爆发，往往正好卡在这个时间窗口上。
我认为，我们现在仍处在第一波热潮的冷静期。
所以当下要么保持对新项目的敏锐，持续挖掘真正新的东西，而不是去冲仿盘；
要么趁低位布局赛道龙头，建一个自己拿得住的仓位，然后等风来。
近期 AI Agent 支付基础设施项目 @mrdn_finance $MRDN 因接入 @circle 推出的 Agent Stack 而受到市场关注。Circle 将 Agent Stack 定义为面向 Agent 经济的金融基础设施，其核心方向之一正是由 Circle 提供支持的微支付能力，而 Meridian 已率先提供相关落地支持。
Meridian 主打面向 Agent 经济的 x402 支付网络，试图通过统一的 USDC 跨链结算体系，解决传统跨链支付中 Gas 获取复杂、资金滞留等问题，实现 Agent 之间的高速自动化支付与协作。项目目前支持零 Gas 用户体验、2 秒级结算以及跨多条 EVM 网络的稳定币流转，并进一步布局 Agent 商业生态、AI-to-AI 服务交易以及开发者收益返还机制。
现在是淡季，eth价格涨幅也被土豆吊打，不过有意思的是，其实现在以太坊的日均网络交易量已经远高于2021年牛市巅峰时的网络活动。
活跃地址、新地址创建、智能合约调用等指标也同步创出新高。这不是刷量或 spam，而是真实用户增长 + L2 生态带动的结果（Pectra、Fusaka 等升级降低费用、提升容量后，网络更活跃）。
注意：这些主要是以太坊L1主网数据。如果算上 Arbitrum、Base、Optimism、Lighter等 L2，总交易量和用户活动还要高不少——以太坊生态的“实际使用”其实被严重低估。
以太坊不会一直被忽视和埋没。它是基础设施 + 安全 + 流动性之王。
Generally agree with the macro. Three forces mentioned (AI capex, just-in-case infrastructure, war financing) are all pushing fiat expansion, and politically nobody can stop them in the next 18 months.
There's a counterflow @CryptoHayes doesn't mention though. While sovereigns sell US Treasuries to build pipelines and stockpiles, emerging market retail is eating up part of that flow through stablecoins. Tether alone holds $117B in T-bills. Tether and Circle combined hold more US Treasuries than Saudi Arabia (OMG!). Over 80% of stablecoin transactions happen outside the US. Abu Dhabi SWF leaving the trade, not-starving Bangladeshi office worker entering it.
The three forces still play out, just with more buffer. Maximum risk is overshooting, IMO.
Next phase is yield-bearing and high risk traditional instruments (stocks). Tether and Circle pocket the T-bill yield right now, which is bad for the user. USDY from @OndoFinance and similar products pass it through. Once 4 billion phone-wallet users get access to these instruments, the dollarization flow speeds up a lot. And this is why RWA is happening.
What this picture needs: cross-chain rails with privacy on par with TradFi. The world is multichain and staying that way. BUIDL on @ethereum, xStocks on @solana, USDT on @trondao, BENJI on @StellarOrg. A Bangladeshi user needs to move between them in one step, with the privacy he expects by default.
That's the gap @NEARProtocol Intents fills.
《CLARITY Act》（清晰法案）对“去中心化”的定义非常清晰，果然是清晰法案，名副其实。
它支持“真实的去中心化”，而不是”表演性的去中心化“。
之前判断一个加密项目算不算“去中心化”，标准很模糊，用的是“共同控制”（common control）测试，这让很多”公司链“（表演性去中心化）钻了空子，说“我们当初没承诺啥，现在就不该被SEC管”。
《CLARITY Act》（清晰法案）直接把这个漏洞堵死了，改成更严格的“成熟区块链系统”（mature blockchain system）的标准，它的核心是消除实质控制，不能有任何个人或小团队还能单方面影响系统的规则、升级、运行或治理。
• 系统必须真正公开、无许可（任何人都能跑节点，不能随便封禁用户）；
• 必须是纯规则驱动（透明代码自动跑，不能有人手动干预核心功能）；
• 没人（或一伙人）能控制超过20%的代币/投票权；
• 系统要足够自治（没人能随便单方面改功能、改规则）；
• 经济上也要独立，价值主要来自网络的实际使用，而不是创始人/团队的持续努力。
那么，会带来什么影响？
1. 对公司链（表演性去中心化）来说：
以后很难再假装“去中心化”了。内部人卖币、操作就会被当成“证券”管得更严，要多披露信息、遵守更多规矩。
2. 对目前的L2来说：
安全理事会（security council）权力必须严格限制，不能随便一票否决或干预。类似于这次Arbitrum在kelpDAO事件中的那种快速干预，以后很难了。
3. 对真正去中心化的项目来说：
是实在的利好。
可以走清晰的路径：比如早期按投资合同募资（SEC管），网络成熟后，代币转为数字商品（CFTC管），融资、交易以及二级市场规则都更明确，不用再担心SEC的突然执法。
时用加密钱包做交易时，大部分人只能看到满屏的十六进制字符。因为完全看不懂底层逻辑到底写了什么，大家基本都是靠猜测或者对前端界面的信任，闭着眼睛点击确认。
但这种盲目签名的习惯非常危险。
黑客经常利用这种信息差进行钓鱼攻击。你看着前端网页以为自己正在免费领取空投代币，实际上钱包底层的代码写的是授权对方转走你的全部资产。这种情况下，仅仅依靠钱包自己去模拟交易结果，或者强行解析底层代码，根本没有办法百分之百还原真实的交易意图。一旦按了确认，资产就找不回来了。
为了让交易代码变得好懂一点，开发者之前推出了结构化数据标准（也就是 EIP-712）。这个技术把原本杂乱无章的机器乱码梳理成了一排排相对整齐的数据列表。
但问题是，整理过的代码依然是代码。
普通人根本看不懂那些英文变量名和复杂的参数到底代表什么具体动作。交易一旦变得复杂，用户需要的其实是一看就懂的日常沟通语言，而不是格式化之后的程序员代码。
为了彻底解决看不懂交易内容的问题，以太坊基金会目前牵头在做一套叫清晰签名（Clear Signing）的开放标准。这个标准要在代码和普通用户之间加上一层翻译。
翻译层的任务就是把底层复杂的机器逻辑变成普通用户可以读懂的逻辑。你在钱包里点击确认前，它会用极其清晰的自然语言告诉你这笔交易的具体动作（比如在 Uniswap 上把 1000 个 USDC 换成至少 0.42 个 WETH）。这样才能做到你眼睛看到的内容，就是你真正签名的内容。
这套翻译系统的运作方式思路很明确。
首先需要项目的开发人员或者安全团队为智能合约编写一本翻译字典（属于 ERC-7730 标准）。他们把每一段晦涩代码对应的白话解释一条条写进字典里，然后把这些翻译字典集中发布到一个完全公开且中立的注册表中。
当你发起一笔交易时，你的钱包就会去这个公开注册表里找对应的字典。找到之后，钱包脱离了应用的前端界面，独立把读不懂的代码翻译成清晰的文字确认界面展示给你。
但如果任何人都可以随意往公开的注册表里上传字典，这就又带来了一个巨大的安全漏洞。黑客完全可以自己建一本假字典，把转移资产的恶意代码翻译成领取奖励的虚假文字，继续在翻译环节欺骗用户。
为了防止字典被恶意伪造，这套系统紧接着引入了第三方背书机制（属于 ERC-8176 标准）。
这套证明框架允许专业的安全审计机构去检查注册表里的字典。如果检查确认字典的翻译准确无误且没有恶意，机构就会给它提供一个安全验证。你的钱包拥有最终的决定权，你可以设置钱包只认那些经过可信赖安全机构检查过的翻译内容。这样就阻断了假字典的生存空间。
由于整个系统的注册和验证都是公开分布的，就算某个中心化的运营机构停止服务，系统依然能依靠其他机构独立运转。
目前这套标准已经完成了核心机制的设计，以太坊生态内的钱包和应用正在逐步对接这套无许可的翻译与验证架构。
i am a huge proponent of privacy $ZEC and recently:
- shared the desci narrative
- caught onto $ONDO with a 2x
- money/credit markets like $AAVE $AERO $MORPHO
- $HYPE has been the strongest thesis along with several ai companies given their similar verticalization approach
there are alpha to some early projects which you should also explore
👇
1) @SeismicSys
privacy-first blockchain with normal ethereum developer experience. developers can build private apps without learning a totally new language or stack. most privacy chains require special tooling, new languages, or complex cryptography. Seismic tries to make privacy feel like normal Solidity/EVM development.
- $17M total raised (latest $10M led by @a16z crypto + @polychain + others in late 2025).
- testnet live since early 2025 (faucet + devnet activities ongoing, community is farming for potential airdrop).
- partners like @brookwellapp already routing stablecoin flows.
- narrative fit is perfect (privacy + real fintech adoption).
- this has the strongest VC signal in the list.
2) @tmrwfinance
creator lending against future AdSense revenue. youtubers and digital creators can borrow based on expected future income. Karat is more like banking for creators. Spotter buys catalog rights. Tomorrow uses a lending model, so creators can get capital without selling ownership of their content.
- very early/stealthier than the rest.
- part of @base ecosystem cohorts.
- fits creator economy + RWA lending narrative but less public traction/funding visible yet.
- underdog play vs bigger tokenized credit names.
3) @lightconexyz
trade the impact of events, not just whether events happen. it answers: “what happens to ETH/BTC/SOL if this event occurs?” prediction markets trade probability. options trade price movement but not clean event-specific outcomes. lightcone creates state-based assets, so users can hedge very specific scenarios directly.
- product site live.
- still pre-token/early.
- volves prediction market narrative into something more useful for hedging real outcomes.
- low competition in this exact flavor.
4) @techdollarhq
borrow against private tech shares without selling them. Employees or early investors in companies like AI labs or frontier tech firms can access liquidity while keeping upside. we recently saw @OpenAI employees sold $6.6b of equity and instead of selling, they can get a loan. Forge and EquityBee mostly focus on secondary sales or option financing. TechDollar focuses on lending against private equity, closer to margin lending for private startup shares.
- product page live with waitlist + clear borrower thesis (frontier tech conviction without forced exits).
- bridges TradFi private equity illiquidity with crypto-style credit.
- no big funding announced yet but clean positioning in the private shares as collateral narrative.
5) @agra_gg
secondary market for tokenized private credit. it lets holders exit before maturity instead of being stuck until redemption. most tokenized credit platforms only solve issuance. Agra solves liquidity. Its yield-based orderbook also fits bonds better than normal AMMs or price-based trading.
- beta live on Ethereum right now with real markets like Anemoy/Apollo credit fund and others.
- tokenized private credit is one of the hottest RWA sub-narratives ($14B+ already in the space).
- first-mover liquidity layer here is real product alpha.
6) @OrnnExchange
commoditizes GPU access. it treats compute like oil or electricity: a scarce resource that should be priced, traded, and accessed efficiently. most GPU projects are just marketplaces. Ornn’s bigger idea is that GPU compute becomes a financialized commodity layer for the AI economy.
- $5.7M seed (oct 2025).
- index already on Bloomberg Terminal.
- live spot pricing + derivatives coming.
- perfect AI infra narrative, compute as the new oil.
- regulated U.S. venue angle is big for institutions.
7) @numoforex
fx infrastructure for frontier markets that gives currency hedging tools to markets where normal fx products are missing or too expensive. most fx platforms serve large institutions or developed markets. Numo targets the places where fx risk is most painful but least served.
- product live with actual markets and volume.
- zero-fee on/off-ramps.
- targets painful real-world FX risk in underserved markets.
- one of the few actually shipping frontier FX infra.
8) @meanwhile
bitcoin-native life insurance. it connects long-term $BTC holding with long-term financial planning. most bitcoin products are trading, lending, or custody. Meanwhile treats btc as a long-duration savings asset, closer to insurance and retirement planning.
- massive $82M raise (oct 2025).
- first and only licensed BTC-native insurance carrier.
- huge signal for BTC as serious savings vehicle beyond trading/lending.
- institutional partner push incoming.
9) @merit_systems
from builder reputation to AI-agent commerce. it started by tracking and rewarding open-source contribution, then moved toward payments/commerce for agents. most reputation systems are static badges. their newer direction is more dynamic: enabling AI agents to transact, earn, and interact economically.
- shift to agent economy is perfectly timed with AI agent hype.
- live tools (AgentCash, Poncho, directories).
- enables permissionless agent-to-agent commerce.
- narrative tailwind is massive.
10) @StreetFDN
tokenizing startup equity. it builds standards and infrastructure to make private startup shares easier to issue, manage, and transfer. equity tokenization usually focuses on public stocks, funds, or bonds. Street targets startup equity, which is illiquid, paperwork-heavy, and structurally hard to access.
- claims 150+ startups tokenized.
- solves paperwork + access issues in VC equity.
- fits the broader equity tokenization wave (SPV wrappers are the current compliant path).
- early mover on actual startup (not public stock) side.
People still valuing Ethereum like a crypto app chain are completely missing what is happening.
JPMorgan just filed for a tokenized money market fund designed around Treasury and repo collateral for stablecoin reserves.. not a meme, not an experiment, but the actual reserve layer of digital dollars moving onto Ethereum rails.
That’s the insane part. Stablecoins become the consumer money layer, tokenized Treasuries become the backing layer, and Ethereum quietly becomes the settlement layer underneath Wall Street’s onchain cash system. Most people are still arguing about fees while the largest banks in the world are rebuilding money markets on $ETH.
记得Defi Summer 20年到22年这三年时间，大家经常感慨的一句话就是“币圈一天，人间一年”
因为技术迭代太快了，每天都有数不行的新东西出来，几天不看就跟不上节奏
现在这个感觉在币圈彻底消失，在AI那边出现了，每天的新东西就看不完，真的看不完
正好这几天看到圈里讨论Cerebras的帖子多了起来，昨天还看到不少路演信息，毕竟14号要挂牌交易。之前在Amber做分享的时候也单独讲过这个公司（当时用的那页PPT放在文末），所以简单说下我对Cerebras的判断
先说结论：这家公司很有意思，但核心变量不一定是他们造的芯片本身，而是未来AI的workload形态会是成什么样。
先简单说下它是干嘛的，防止有老铁还没看Cerebras
它最核心的技术，就是wafer-scale chip（晶圆级芯片），简单理解就是别人还在用“邮票大小”的芯片，它直接把整片晶圆做成一颗超大处理器，再配上超大 SRAM，把大量数据尽量留在本地高速处理，减少传统GPU最头疼的memory bottleneck。现在海力士美光涨上天就是因为HBM这东西太抢手，Cerebras是直接绕开了对HBM的依赖
很多人看到 Cerebras最牛逼的benchmark：推理速度比GPU快10-15倍，第一反应就是下一个英伟达？！
先别急。
这个benchmark最大的问题，是默认AI的核心需求永远是“更快吐token”。
如果只是人类盯着ChatGPT聊天，这故事其实没那么性感。你每秒吐 30 个 token，我已经快读不过来了。再翻10倍，边际体验提升几乎为零。
真正有意思的是 Agent。
Agent不读字，Agent消费Token。速度直接等于生产力。
一个OpenClaw/Hermes agent，如果要读网页、写计划、调 API、跑代码、报错重试、继续执行，一个任务可能要几十次 inference call。
每次 2 秒，任务就是分钟级体验。
每次 200ms，就是另一个世界。
所以 Cerebras更值得关注的地方，是 AI worker 这条线，而不是单纯的chatbot加速。
但问题来了 - 它的magic来自 wafer-scale + 超大SRAM，本地访问极快。但SRAM这东西有天然 tradeoff，速度快，容量贵，大模型塞不下就必须拆分。
而一旦拆分，chip-to-chip communication 就上来了。
LLM inference里最怕通信的恰恰是 decode 阶段。token 是一个一个往外吐的，每多一次 hop，延迟就是硬加上去，没法隐藏。
所以 Cerebras能不能成，关键根本不在“比 GPU 快几倍”，而是未来 AI 的主流计算形态到底是什么。
1. 时间线一 - 未来几年还是 前沿超大模型一统天下，动不动几千亿甚至万亿参数，所有请求都让超大模型自己处理，那 Nvidia的distributed infra 依然最舒服，Cerebras的速度优势会被通信损耗吃掉很多。
2. 时间线二 - MoE、蒸馏、量化这些技术如果继续快速进步，未来两年70B左右的模型做到今天700B模型 80%-90% 的效果，我一点都不意外。（这里要强烈感谢Deepseek一大波!)
如果世界往这个方向走，故事就变了。
大模型负责 planning / judgement / orchestration。
真正执行任务的大量 worker model 落在 30B-70B 这个区间。
这些模型足够聪明，又刚好能吃到高速本地推理带来的红利。
Agent 世界里，大部分token根本不需要最聪明的大脑。很多工作本质就是执行层体力活：浏览网页、改代码、调工具、retry、继续跑流程。
这个 topology一旦成立，Cerebras就直接进入自己的甜区
3. 时间线三 - 未来推理是端侧为主，用的都是8B，14B的这些小模型，那GPU也能跑的很好，甚至专用的ASIC芯片效率更高，这个场景下Cerebras的优势和护城河也就不高了
换句话说，超大模型云端推理或者超小模型的端侧推理这两个平行宇宙，Cerebras的优势都不够明显，只有主流推理落在32B-70B中等模型这个Size，正好“Big enough to stress GPU memory, Small enough to fit locally” 才是Cerebras最能大展身手的世界
所以我对Cerebras的判断是，300多亿的市值，短期看订单，财务报表这些数据，长期赌的则是未来Agent时代的计算范式到底落在哪个平行宇宙的时间线上
日历
5 月 19 日
暂无重要事件
