等等，什么？ “区块链护城河：3/10 分”？ 这也太荒谬了吧。连 Santi 都不认为区块链“没有护城河”。 以太坊已经保持了 10 年不间断的主导地位，期间有数百个竞争者试图抢占市场，其中许多项目融资总额超过 100 亿美元。 但经过整整十年的围攻，每一次挑战者试图把它打下来，以太坊都守住了王座。 如果这还不能证明以太坊有护城河，那我真不知道什么叫护城河了。 Wait what? "Blockchain moats: 3/10"? This is getting silly. Even Santi did not think blockchains "don't have moats." Ethereum has had 10 years of uninterrupted dominance, with hundreds of challengers who've raised $10B+ in funding to try to take market share. After a decade's worth of contenders trying to take it down, Ethereum has managed to defend its throne every single time. If that doesn't tell you Ethereum has a moat, I don't know what moats are.
HaseebManaging partner, dragonfly_xyz
观点原文
等等，什么？
“区块链护城河：3/10 分”？
这也太荒谬了吧。连 Santi 都不认为区块链“没有护城河”。
以太坊已经保持了 10 年不间断的主导地位，期间有数百个竞争者试图抢占市场，其中许多项目融资总额超过 100 亿美元。
但经过整整十年的围攻，每一次挑战者试图把它打下来，以太坊都守住了王座。
如果这还不能证明以太坊有护城河，那我真不知道什么叫护城河了。
Wait what?
"Blockchain moats: 3/10"?
This is getting silly. Even Santi did not think blockchains "don't have moats."
Ethereum has had 10 years of uninterrupted dominance, with hundreds of challengers who've raised $10B+ in funding to try to take market share. After a decade's worth of contenders trying to take it down, Ethereum has managed to defend its throne every single time.
If that doesn't tell you Ethereum has a moat, I don't know what moats are.
Ethereum
相关观点
- IMO，我觉得是时候淘汰“vibe coding（凭感觉写代码）”这个说法了。到现在为止，这件事已经完全不是基于“感觉”的了。它涉及的是： * 管理智能体（agents） * 把它们引导到正确方向 * 给出准确反馈 * 帮它们摆脱卡点 我们更应该把这件事称为：“驾驶（piloting）一个 LLM”。 当然，确实有一些“飞行员”只是按下自动巡航按钮，完成一次很轻松的飞行；但也有一些飞行员，能从同一台机器里榨取出多得多的能力——使用子代理、技能模块、规划系统、在同一代码库中协调多个智能体、合并冲突、同时监管多个项目，等等。 当你说 “我是 vibe coding 写的” 时，这句话暗示：所有的主动性（agency）都在 LLM 身上，等于是抹掉了你自己的作者身份。 这个词本身的诞生，就是为了给“不严肃”找借口，用来合理化对技艺本身的忽视：“我知道这样写代码有点不正经，但要不我们干脆就接受它吧？” 但这和一个飞行员使用复杂仪表成功降落飞机，本质上没有任何不同。那不叫 “vibe flying（凭感觉飞行）”。 未来，几乎所有东西都会被称为“vibe coded”，但程序员的主动性并没有消失，而是上升到了更高的抽象层级。 用 Karpathy 的说法：我们已经从汇编编程，走到高级语言编程，再到现在的自然语言编程。但“作者性（authorship）”这个概念，不会消失。所以我认为，我们应该说的是：“我是驾驶 Claude 构建了这个。” 这，才是软件工程未来真正的样子。而我们用来描述它的语言，也必须随之进化。 IMO it's time to retire the term "vibe coding." At this point it is not vibes-based at all. Managing agents, steering them in the right direction, giving accurate feedback, helping them get unstuck. We should call it "piloting" an LLM. Yes, there are pilots who just press the cruise control button and handle an easy flight, but there are pilots who can get so much more out of the same machine—subagents, skills, planning, coordinating multiple agents in the same codebase, merging conflicts, overseeing multiple projects simultaneously, etc. When you say "I vibe coded this," it implies that all of the agency was in the LLM—it erases your own authorship. The term itself was created to give license to un-seriousness, to justify an ignorance of craft. "I know it's naughty to program that way, but if we just embrace it?" But that's no more true than it is for a pilot who used their complex instrumentation to land a plane. That's not "vibe flying." Almost everything will now be "vibe coded," but the agency of the programmer simply rises to a higher level of abstraction. In Karpathy's terms, we've moved from programming in assembly, to programming in higher-level languages, to now programming in natural language. But authorship as a concept will survive. That's why I think we should say "I piloted Claude to build this." This is what the future of software engineering looks like. Our language for it needs to adapt.Haseeb · 2026/01/13
- Web3 终于死了吗？ 本周我们将讨论 Farcaster 的“转向（pivot）” 及其对 Web3 社交意味着什么； 同时也会参与 L1 估值之争； 还会聊到那篇走红的 《我在加密行业浪费了 8 年》 的帖子（作者还是我们都曾支持过的人！）。 当然，还有 Tarun 那篇如今颇具争议的 ADL 论文。 以下是关于YouTube视频《Web3 Just Died… Crypto Returned to Its Roots (Money) - The Chopping Block》的详细观点整理： 视频主题： Web3 已死？加密货币回归本源（金钱）。 1. Farcaster 的战略调整与 Web3 社交的未来 Farcaster 的转变： 去中心化社交协议 Farcaster 宣布将战略重心从纯社交产品转向钱包产品。尽管融资超过 1.5 亿美元，且开发了 4 年多，但其日活跃用户数一直停滞在 2.5-3 万左右，难以突破加密圈内人群。 “Web3 社交已死”的讨论： Tarun 的观点： 认为这标志着以 A16Z 定义的“Web3”（所有权经济、NFT、拥有社交数据等）的终结。他认为真正的“加密社交网络”其实是像 Polymarket 和 Kalshi 这样的预测市场，人们在这里通过交易来表达观点和获取信息，这是一种全新的媒体形式。 Haseeb 的观点： 对 Web3 社交一直持悲观态度。他认为加密货币的核心超能力在于金钱和金融。如果产品没有很强的金融属性，就很难与 Web2 巨头（如 Twitter、Instagram）竞争。用户嘴上说想要去中心化和隐私，但实际行为表明他们更看重便利性和网络效应。 Robert 的观点： 并不认为这是“死刑判决”，只是还没出现真正的赢家。现有的 Web3 社交产品多是对 Web2 的拙劣模仿，没有提供足够的新价值。 Tom 的观点： 很难通过微小的改进来取代现有的巨头。未来的社交形式可能完全不同，例如预测市场这种结合了金融属性的新形态。 2. Layer 1 (L1) 区块链的估值辩论 背景： 业界关于 L1（如以太坊、Solana）估值是否过高存在争议。一方认为目前的市盈率（P/E）极高（如以太坊 380 倍，Solana 50 倍），估值不合理；另一方（如 Haseeb）认为 L1 类似于早期的互联网平台，市场是前瞻性的，看重未来的增长潜力。 Robert 的观点： 认为 L1 略微高估。原因是竞争永远不会停止，总会有新的 L1 出现（如 Stripe 或其他巨头可能发链），导致现有份额被稀释。他甚至表示如果由他决定，会将所有 L1 的市值砍掉 50%。 Tarun 的观点： 认为长期来看，L1 的承载能力是有限的，流动性会集中在少数几条链上（强者恒强）。但他同时也指出，L1 的估值往往受韩国和土耳其等散户交易者的情绪影响较大。 Haseeb 的观点： 坚持认为随着稳定币和更多金融资产上链（可能占 M2 的 15%），L1 作为底层金融基础设施的价值将呈指数级增长。目前的低收入是由于费用不断降低以换取规模扩张。 3. “我在加密领域浪费了 8 年”——对行业现状的反思 背景： 一篇题为《我在加密领域浪费了 8 年》的文章在科技圈疯传，作者 Ken Chan（前 AO 创始人）表达了对加密行业充斥着赌博、投机和零和游戏，而未能实现去中心化金融理想的失望。 Robert 的观点： 加密行业确实有两种：“糖水加密”（Sugar Water Crypto，指各种积分、代币游戏、甚至庞氏骗局）和“构建真实事物”（Build Real Things，如更高效的金融市场）。在这行容易倦怠，但这并不代表行业没有价值，只是需要选择正确的方向。 Tarun 的观点： 对文章作者表示同情，因为他们选择了做去中心化交易所（Perp DEX），这本身就是赌场业务，如果不想做赌场就不该选这个方向。此外，这也是典型的“起了大早，赶了晚集”的创业者挫败感。 Haseeb 的观点： 承认行业早期确实有很多赌场成分，但这并不意味着整个行业没有价值。对于那些感到幻灭的人，离开也许是好事，让真正想建设的人留下来。 4. 自动去杠杆 (ADL) 机制的研究 Tarun 的新论文： Tarun 发布了一篇关于改善加密衍生品交易所中“自动去杠杆”（ADL）机制的论文。 核心发现： 现有的 ADL 算法假设坏账是孤立事件，但在极端行情下（如 10 月 10 日），ADL 事件会引发连锁反应。 改进建议： 通过数学优化，可以设计出更好的策略，在极端情况下可能为行业节省数亿美元的损失。这体现了加密行业在微观结构层面的持续创新。 总结： 本期节目的核心基调是“回归金融”。嘉宾们普遍认为，试图把 Web2 的社交模式简单搬到链上的尝试（Web3 Social）大多失败了，或者也是死路一条。加密货币真正的杀手级应用和护城河依然是金融（Finance）和金钱（Money）。无论是预测市场的兴起，还是对 L1 价值的讨论，最终都指向了这一点。 Is Web3 finally dead? This week we discuss the Farcaster "pivot" and what it means for web3 social, we weigh in on the L1 valuation debate, as well as the viral "I wasted 8 years in crypto" post (from an author we all backed!). And ofc, Tarun's now controversial ADL paperHaseeb · 2025/12/12
- 加密世界里一直都有“赌场”。 比特币上第一个真正病毒式传播的应用是 Satoshi Dice（2012）。 以太坊上第一个爆火的智能合约是 King of the Ether Throne（2015），本质上就是一种“烫手山芋式庞氏游戏”。 一旦你拥有了可编程的货币，人们做的第一件事就是赌博、玩一些 愚蠢的游戏。这是人性。 加密世界里从来都有热门赌场。永远都有。从 ICO 赌场，到 DeFi 食物币赌场，到 NFT 赌场，再到现在的 memecoin 赌场。形式会变化，但赌博深深嵌入在人类天性里。甚至早在文字发明之初（公元前3000年的古苏美尔），就有赌博的记录。 赌场闪闪发亮，在社交媒体上获得极大关注（因为财富会在这里瞬间暴涨或蒸发，而人们很容易说服自己“这次就是我的好运时刻”）。但是，如果你只盯着赌场的光芒，你就会错过更大的故事。 让我十年前踏入这个行业的那个故事—— 当时听起来像科幻小说—— 加密货币是一种更优的金融底层结构，它将永远改变货币的本质，也会永久改变个人与政府之间的权力平衡。 ——这个故事，现在真的正在发生。 比特币正在挑战国家机器。各国开始把它放进自己的资产负债表。稳定币正在扭曲货币政策。各国央行慌忙应对。像 Uniswap 和 AAVE 这种无需许可的智能合约，现在比许多独角兽金融科技公司更大、更有价值。世界正在随加密而改变形状。 也许这一切比你预想得更久，但变革性技术的扩散几乎总是比你想象得更慢。我曾以为生成式 AI 会在几年内彻底改变就业市场，但在 ChatGPT 出现三年后，它仍没有体现在 GDP 或就业数据中。 改变世界是很难的。需要时间，也需要工作。 工业革命始于 1781 年蒸汽机的商业化，但直到 1830 年——整整 50 年后——才真正影响生产率、就业和工资。 互联网如大家所知，也花了 20 多年。 你真的以为我们能在短短 5 年内颠覆整个金融体系吗？ 这是全世界监管最严、最敏感的行业。 如果你因为自己做了一个 memecoin L2 之类的项目没有暴富而绝望，请深呼吸一下。这个行业并不欠你什么。 而这些“精神投降”、抱怨与悲观，信不信由你，其实是健康的。 一片森林之所以能保持强韧，是因为它会定期清除枯木。 没有周期性的火灾来清理，腐朽会蔓延，整片森林都会腐烂。现实是残酷的——要继续成长，就必须清理掉那些已经不再贡献价值的部分。 所有这些负面情绪必须以某种方式被排出。 让这些人慢慢离开，空气就会变得清新。 要么他们改变想法、重新关注未来， 要么他们离开，让剩下的人继续把事情做完。 因为—— 这项事业还没有完成。 There have always been casinos in crypto. The first viral application on Bitcoin was Satoshi Dice (2012). The first viral smart contract on Ethereum was King of the Ether Throne (2015), which was basically a hot potato ponzi scheme. Once you have programmable money, the first thing people will do is gamble and play stupid games. It's human nature. There's always been a hot casino in crypto. Always. The ICO casino, the DeFi food coin casinos, the NFT casinos, and now the memecoin casino. It evolves over time, but gambling is deeply ingrained in human nature. Literally as soon as humanity invented writing (ancient Sumeria in 3000 BC), there's records of people gambling. The casino is shiny and gets a lot of attention on social media (it's where fortunes are quickly won and lost, and it's easy to convince yourself that this is your lucky break), but if you only pay attention to the shimmer of the casinos, you're missing the bigger story. The story I originally got into this industry for, that sounded like sci-fi to me ten years ago—that crypto is a better substrate for finance, that it will forever alter the nature of money, and that the balance of power between individuals and governments will be permanently changed—that story is actually happening. Bitcoin is challenging nation states. They are now buying it on their own balance sheets. Stablecoins are bending monetary policy. Central banks are scrambling to respond. Permissionless smart contracts like Uniswap and AAVE are now bigger and more valuable than unicorn fintechs. The world is warping around crypto. Maybe it took longer than you thought, but transformative technologies almost always take longer to diffuse than you think. I thought generative AI would completely transform the job market within a couple years, but three years after ChatGPT, it still doesn't show up in GDP or employment stats. It's hard to change the world. Takes work. The industrial revolution began with the commercialization of the steam engine in 1781, but even the industrial revolution didn't have any impact on productivity, employment, or wage growth until 1830—that's 50 years later. The Internet, as we all know, took 20+ years. Did you think we'd be displacing all of finance, the most regulated and governmentally sensitive industry in the world, in the span of just 5 years? If you're despondent that you were working on a memecoin L2 or whatever and didn't get rich, take a deep fucking breath. This industry doesn't owe you that. And yet, all this spiritual capitulation on the timeline, believe it or not, is healthy. A forest only stays robust when it periodically sheds the deadwood. Without occasional fires to clear it out, decay spreads and the whole forest rots. It's a harsh reality, but the only way to keep growing is to clear out whatever's no longer contributing. All this negativity on the timeline needs to flush out one way or another. Let people keep shuffling out, and the air will clear. Either these people must change their minds and refocus on the future, or they need to leave so the rest of us can do our jobs. Because the job is not yet done.Haseeb · 2025/12/07
- 这两点完全错得离谱。 在进入加密行业之前，我曾在 Airbnb 做反欺诈。因此，我经常和 Uber 的反欺诈团队交流。Uber 之所以放弃中国市场，就是因为遭遇了成批职业化欺诈团伙的疯狂攻击——他们每个月制造成千上万次假订单来薅 Uber 的补贴。Uber 在中国几乎就是被“薅羊毛”薅到倒闭的。 任何提供激励机制的科技公司都必须应对欺诈，并且要预料到一定比例的补贴会被骗子拿走（在加密行业我们称之为“farmers”）。认为这种情况是加密行业独有的想法，实在太荒谬。很多 Web2 创业公司之所以倒闭，就是因为他们没能有效控制激励滥用，而投资人却被告诉这是在“补贴需求”。 Gwart 的第二个观点是：在加密行业，这些激励是“零成本”发放的，而在 Web2 公司里，“风投是在烧真金白银。” 请停下来想一想——当这些“薅羊毛的人”把你给的代币抛掉时，他们最终拿到的是什么？他们卖成了现金。 所以 在 Web2 里： 股权 ⇒ 卖给私募投资者换成现金 ⇒ 用现金补贴用户 ⇒ 用户拿到现金 在加密里： 代币 ⇒ 给用户代币 ⇒ 用户抛售代币 ⇒ 用户拿到现金 这完全是同一件事，你只是跳过了一个中间环节而已。 市场会在不断试错中进化。从失败中学习，区分“好支出”和“坏支出”。过去加密初创公司认为大量砸钱做短期激励是有效的，但现在他们已经意识到这并不起作用，并正在调整策略。在“零工经济”时代也是同样的过程——大量公司倒闭，幸存者学会了让经济模型变得可持续。 Literally dead wrong on both points. I worked in anti-fraud at Airbnb before I got into crypto. So I often shared notes with people who worked in anti-fraud at Uber. Uber gave up their China business because they were beset with waves of professionalized fraud rings, taking thousands of fake rides a month to farm Uber incentives. Uber was quite literally farmed out of business in China. Every tech company that offers incentives has to police fraud, and they expect some amount of waste that gets captured by fraudsters (in crypto we call them "farmers"). The idea that this is unique to crypto is silly. Many, many web2 startups have gone out of business because they failed to put a lid on incentive abuse, while investors were told they were "subsidizing demand." Part 2 of Gwart's claim is that in crypto, these incentives are given out "at 0 cost," while in web2 companies, "VCs are burning real cash." Stop and think for a second. What do farmers get if they are dumping your token incentives? They are selling it for cash. So, in web2: equity => sell for cash to private investors => give cash to user => user gets cash web2: tokens => give user tokens => user dumps tokens => user gets cash It's the same damn thing, you're just skipping a middleman. Markets evolve through trying shit and learning from failure. There is good spend and bad spend. To the extent that crypto startups previously thought spending a lot of money on short-term incentives was effective, they have now learned that it isn't, and they adapt. Same thing was learned in the "gig economy" era where a bunch of companies went bust, and the ones that survived learned how to keep economics sustainable.Haseeb · 2025/12/04