Howard

Prev. investment @SyncracyCapital

[F Train Freddy @FtrainFreddy]

When the over eager intern tries to ask the CEO for a coffee chat on day 1 of internship season

[Yami @YamiPNL]

🚨 BREAKING: LILY PHILLIPS ACCUSES MICHAEL SAYLOR OF STEALING HER IDEA OF FUCKING 100'000 MEN IN A DAY

[Herman Jin @ShanghaoJin]

凡所有相，皆是虚妄。若见诸相非相，则见如来

作为一个AI死多头，我知道在所有人狂欢的时候，但凡发出一点悲观的言论都必被讥讽、嘲笑

但是，我很想写此文，其实是复盘与反思：不但是为自己这两年半写X、做交易做的；也是也是对这个疯狂市场做的

很幸运，作为闲暇的爱好，写X结识了很多新朋友。感谢互相欣赏与启发，在过去2年里，我从24年4月初清仓BTC换成AVGO NVDA，大选前AAOI PLTR TSLA INTC，关税战后INTC GOOG，25年底又加仓INTC AAOI，到今年春节逃跑、再满仓AMD NOK（最近又操作了ORCL LITE）。虽然，有买彩票SUP归零，MRVL MDB也有赚有亏，但是我一直坚信AI对token的需求会让半导体every dog has its time，获得了不差得收益。期间不少小伙伴们也都赚到不少于10倍收益

然而，此刻要清醒的意识到，收益并不是我们选股牛，眼光独到，而是市场牛！我复盘了我自己今年3月最后一周，周末前全仓买入。赚钱并不是因为当时买对了AMD，而是当时call 对了市场资金结构

所以，在这个时点上更需要静心凝神，好好反思下市场接下去会如何

喝醉酒的人是不会知道自己喝醉的；沉醉美梦中的人也不会意识到黄粱一梦，这过程可以很久，我们也可以继续疯，但是必须要清醒地在更高阶意识中，认识到随时有被敲醒的风险

One of the four beasts saying, COME AND SEE – Revelation 6:1

不同于华尔街很多人，我是“坚信”AI的

30年来第一次，AI模型用并行计算替代串联计算，把半导体的进步直接变成了生产力（之前科技进步只是作为平台，让用科技的软件、以及用软件的人创造价值）

AI的结果是token需求指数级升高vs 而半导体供应只能线性增加。加上整个半导体老登们在过去几年黄金时间，没有因 “信” 而扩产，今年需求真正开始爆发后全产业链缺货。造就了这轮大牛市

不过，大牛市的命脉是市场price in了:

模型收入加速度，不远得未来可以justify Hyperscalers 的Capex

要知道任何动摇命脉预期的都会让市场陷入恐慌

1. 铜墙铁壁：horseman之一就是微观、公司层面真的无懈可击

从看模型架构、从架构看需求，微观看半导体每一个赛道，每一个公司都近乎完美。每一个，都越看的细致，越了解深入，越是有自信---- 这就是典型低PE泡沫

Rule of Thumb：当看了几十、上百家公司，每家需求、盈利都真实solid，你就必须质疑这个市场有没有问题

低PE泡沫一样是一个泡沫，只是他不同于2000年。高PE是个预期泡沫，市场会不停设定更高的收益预期，一旦核心公司预期破灭，整体预期立即调整，多米诺骨牌一样一片片倒下。这就好像去年的IGV，市场忽然不给EV/EBITDA估值了

但是低PE泡沫没有这个问题，盈利增长会给你足够信心在稍微下跌之后不断加仓冲进去。毕竟谁能拒绝个位数PE的存储，每年涨价涨盈利，毛利率80%。这个过程中，小跌小买，大跌大买。只要不被切中“命门”市场会非常坚挺。持续性正反馈，好像美梦一样让人不愿醒来，也不相信是做梦（买过p2p的知道我说的什么感觉吧）

不对称反馈：明显的体验是在过去1年中，市场对于宏观负面风险的反馈明显弱于正面利好。此时，并非没有“野蛮人”，而是公司盈利能力的城池固若金汤。不管伊朗危机、油价、通胀、利率、Fed统统能拒之门外。反之，一旦命门被破，所有野蛮人都会杀回来，过去没跌的够的一次性补完

2. 去监管与宏观流动性宽裕：

整个金融体系自从2008年金融海啸之后经历了一轮长期的监管与去杠杆，为了弥补系统性流动性损失美联储对应展开了一轮持续的QE。银行当时平均杠杆x30~x40，后来降低到了x15~20。去年美联储降低了eSLR要求，实质性让银行理论可扩4-5万亿美元资产负债表
其实，4月后银行间流动性泛滥，也加剧了风险资产狂欢，低PE下杠杆入市

流动性是一切之王，流动性泛滥征服一切。如果说疫情期间120b/月是细水长流，现在就是开闸放水。对应过往平衡的QE没有转向QT情况下，直接放任银行去监管，制造大量流动性

但是流动性又如多巴胺，不是快乐分子，而是期待分子

放任之后，现在通胀黏着，Fed其实没有短期降息余地。甚至在危机一旦发生之后，都很难去QE，只有降息手段的Fed面对可能的危机是一个纸老虎。而短期流动性的另一面是，目前半导体、特别是存储中的杠杆已经太高了

3. 群雄并起：He went forth conquering, and to conquer

可能因为TSMC带头不相信AI，半导体行业整体在过去2年没有应对AI需求大规模扩产

后果是，当整个产业面临硅基洪流需求，忽然傻眼。过去半导体产业链主要是为了消费电子设计、定价的。而现在，不但订单忽然拉满，而且需要不停迭代架构

NVDA就从GPU fabless，进步到rack方案解决，再往token factory迈进。每个过程需要的技术、供货商都不同。所有供货商都没有经过如此量产设计。这就好像把一台拖拉机强行拉上高铁，让他以200mile速度奔驰，每个零件都会嗡嗡作响。半导体每个细分行业都缺货，缺产能，提高不够产能

与之伴生的一个问题是：原来卖方市场只有TSMC一个“守门员”控货源、控margin、控价格。原仔汤搓的产业有肉吃，需求方对价格产能有预期，良性发展。今天忽然五亿探长“雷诺”控制不住了，每个“小混混“都能来涨价、收保护费。真正的结果不是半导体都牛叉，而是失控：失控会影响每GW需要投入的成本核算，影响到命脉的预期。会给capex与模型收入会师蒙上阴影

4. 钢丝起舞

AI模型收入预期太满

回顾一下，市场坚信定价Anthropic + OpenAI + Gemini的收入会快速增加让Capex合理。这条命脉的信仰是不能有一点动摇的。哪怕略微的动摇都会导致二级市场剧烈波动

那么，过去GPT 5.0也有模型不及预期，市场质疑scaling law为什么没有让半导体大跌？

首先当时半导体没有像今天这样全面开花，也没有如此高的杠杆；其次，当时hyperscaler的现金流冗余较多，只要他们表示相信，拿钱来capex支持是很容易度过难关的。毕竟一旦他们有钱、继续capex，可见2年NVDA收益确定性高，没有人能顶风做空NVDA的

后来也发生过GPT被Gemini追平，并且导致Orcl持有其大量backlog订单而被做空。不过大总统的好兄弟老Larry敢对wall street说F you，用股权融资也要继续加注。很快， Opus横空出世让人意识到AGI时代已经到了

然而，今年capex已经到了770b，明年要1tril，华尔街是不会看Hyperscaler总营收多少的（道理就跟当时ORCL一样：OAI Anthropic可都是拿了他们投资款的），他们必须看到Anthropic 与OAI的增速持续，这个才是维系整个链条自转下去的核心

与此同时，这个游戏中大厂一直担任AI的监护人角色。但是他们的自由现金流已经变负了。“父母老了，力尽了“，后面的路要靠自己。而且市场满满预期下，对这条路一根筋的执着，不会给“主线”一丝一毫的容错

（你们可以回想NV rack液冷出问题，交换机CPO良率问题，这些小巫见大巫、假以时日一定能解决的问题，市场尚且没有任何容错）

命门在Anthropic：大过，大者过也。栋桡，本末弱也

供应延迟、乃至技术卡壳的问题是腠理之疾；任何模型端的风吹草动才是心腹大患

现在御三家模型的统一问题：算力不够，所以失智

我在自己大量使用模型编程、交易程序部署，多个AWS服务互动过程中发现，Opus4.8实际能力已经远远不如中国模型kimi了。虽然GPT还勉强可堪一用，但是也在逐步降智。市场在用固有软件思维考虑AI, 系统开发成本高、使用成本很低。试用好用将来也就可以稳定保证系统质量。但是AI是一个工厂，模型输出需要成本，饭店吃饭的人多了，厨房做不过来，质量自然变差了

第二个市场固有思维：指数的token需求是没有波动的，也不因质量改变。事实上，虽然我因为模型变蠢多用了很多token。但是我质疑token质量不断下降，token消耗增加的怪异剪刀差可以持久。现在很多公司确实对token使用有KPI考核，很多惯性订阅使用者也会加速使用，但如果算力瓶颈无法解决，我非常担心Anthropic增长曲线会变缓。而且算力问题非一蹴而就，也不是一个新模型能解决的

但是当市场开始意识到这个问题的时候，投资人会质问：

1. 跑不过蒸馏你的，模型花那么多钱训练干嘛？

2. 模型是不是商品化的model commoditize ？

当承载整个AI时代的模型“梦”也将受到质疑，整个盈收逻辑将来能否自我循环跑通将被质疑，所有今天拿了PE估值的半导体都同样会被质疑--- 你是不是周期性的？？

有人会说，失智既然因为算力不够就应该加大投入半导体解决问题

没错！但是钱呢？

如果刨掉OAI 与Anthropic拿到的融资款反哺云厂，他们能有多少净现金流可以继续支持这个投入？这个账对华尔街来说非常好算。当然不排除他们像当时orcl，哪怕融股权也要继续投入，熬过这关。到那时，可能市场的反应就会远远超过去年，假设AMZN Meta如果面临CDS 飙升、负现金流，他能否像老larry一样孤注一掷是不确定的

另外有人会说，OAI收入增加能不能弥补投资人担心。我认为预期如此之满的状况下，因为“比烂”而被替代是很难让人安心的（两军对战关羽被阵斩了，你说还有个张飞？没用的）

同时，我们也要意识到。这个Low PE的泡沫又是非常坚韧的，在没有直击到命门伤到命脉预期之前，几乎无敌。不过，一旦点中死穴就会轰然倒塌

我就业以来也曾看过好几次次纸碎金迷，潮起潮落，没有一次全民疯狂会ends well；

也没有一次梦在其中的人，会想明白这个赛道“怎么输”。最近一次币圈小伙伴熟悉的Defi summer。Uniswap很短时间狂卷几十bil USD，当时觉得取代部分传统金融只是时间问题、多少问题，根本不会想、也想不明白“怎么会输”。然而，如果喝醉酒的人想明白怎么会输，他就不会输了

不管Anthropic模型怎么样，面临什么问题，都不会动摇AI是人类第三次工业革命。AI过程不会受任何危机影响，碾着狂欢者的尸体也会继续前行，直到他替代几亿脑力劳动者

最后评论Low PE泡沫：不要说做空，哪怕踏空，痛苦指数都非常高。回头看看2007年做空subprime的哪些基金，有几个活到了最后狂欢？

牛市踏空都会变成笑柄的环境，我接下去会需要很多Aura力量与专注的力量，去一边看空一边做多。请原谅会减少在推特上消耗自己修为论道了，赠言：

可以喝酒，但是不能喝醉
可以讲故事，但是不能真信
可以跳上party桌子跳舞，但是双眼必须紧紧盯着DJ，他放着音乐跑了，你也得跟着跑

[The Wall Street Journal @WSJ]

Hyperliquid, a decentralized crypto platform, is open 24 hours a day, seven days a week. The exchange has emerged this year as a go-to spot for Wall Street’s weekend warriors. https://t.co/3dUhMXjVao

[Keisan.hl @Keisan_Crypto]

Been trading perps more actively lately after mostly trading stocks the last few months and I must say it is really damn cool to have both crypto and stock perps all in one portfolio margin account

Stocks liquidity still has work to do but this is no doubt the best product suite. Adding in spot stocks as another layer for PM will be the finishing touch

Hyperliquid

[blood @bloodweb3]

Men only have two moods:

[Lala @cryptolala]

Last time Saylor sold

[Jared Zoneraich @imjaredz]

at a startup event in sf

im the only one drinking

the bartender thanked me for giving him something to do

[Keisan.hl @Keisan_Crypto]

$PURR raised $21.5M cash from the ATM this week. With that they purchased another $10.9M $HYPE and have an additional $9.6M of cash ready to go

$10M in the clip. $20M+/week on top of the ETFs

Hyperliquid

[Degen CPA @DrewVento]

BREAKING: Victor Wembanyama has joined Anthropic.

[Howard @howardx30]

Embrace and be the change

[Dovey "Rug the fiat" Wan (hiring) @DoveyWan]

https://t.co/vatfJDm3hl

[Robert Scoble @Scobleizer]

Lately I've been doing the opposite of encouraging people to move to San Francisco.

Yes, @paulg is right. Being in San Francisco does give many advantages that he lays out well.

But I meet so many who feel FOMO for not being here. Families won't let them move. Or they don't yet have a business that works here. It's EXPENSIVE.

I spent time this week with one guy building a robot that removes snow and ice from sidewalks. He wants to move to SF to be around other people who work on robotics. Being a founder in the midwest is lonely, he told me.

I told him that the whole advantage of where he lives is he is close to the customers. The government officials who could actually buy such a robot.

Being close to customers is a huge advantage. One that many entrepreneurs don't appreciate enough.

I told another guy "come to San Francisco to get up to date for a couple of weeks then take what you learned back to your local community."

I've seen the advantages AI automation brings to businesses, whether big, like CocaCola, or small, like a dentist in a small town.

But most businesses haven't yet automated around the world. Those businesses don't spend time in San Francisco. And people outside of SF are the ones who will help them automate.

In other words, there is a HUGE opportunity outside of San Francisco/Silicon Valley and one that isn't hyped up enough.

[Deedy @deedydas]

The vibes in SF feel pretty frenetic right now. The divide in outcomes is the worst I've ever seen.

Over the last 5yrs, a group of ~10k people - employees at Anthropic, OpenAI, xAI, Nvidia, Meta TBD, founders - have hit retirement wealth of well above $20M (back of the envelope AI estimation).

Everyone outside that group feels like they can work their well-paying (but <$500k) job for their whole life and never get there.

Worse yet, layoffs are in full swing. Many software engineers feel like their life's skill is no longer useful. The day to day role of most jobs has changed overnight with AI.

As a result,
1. The corporate ladder looks like the wrong building to climb.
Everyone's trying to align with a new set of career "paths": should I be a founder? Is it too late to join Anthropic / OpenAI? should I get into AI? what company stock will 10x next? People are demanding higher salaries and switching jobs more and more.

2. There’s a deep malaise about work (and its future).
Why even work at all for “peanuts”? Will my job even exist in a few years? Many feel helpless. You hear the “permanent underclass” conversation a lot, esp from young people. It's hard to focus on doing good work when you think "man, if I joined Anthropic 2yrs ago, I could retire"

3. The mid to late middle managers feel paralyzed.
Many have families and don't feel like they have the energy or network to just "start a company". They don't particularly have any AI skills. They see the writing on the wall: middle management is being hollowed out in many companies.

4. The rich aren’t particularly happy either.
No one is shedding tears for them (and rightfully so). But those who have "made it" experience a profound lack of purpose too. Some have gone from <$150k to >$50M in a few years with no ramp. It flips your life plans upside down. For some, comparison is the thief of joy. For some, they escape to NYC to "live life". For others still, they start companies "just cuz", often to win status points. They never imagined that by age 30, they'd be set. I once asked a post-economic founder friend why they didn't just sell the co and they said "and do what? right now, everyone wants to talk to me. if i sell, I will only have money."

I understand that many reading this scoff at the champagne problems of the valley. Society is warped in this tech bubble. What is often well-off anywhere else in the world is bang average here.

Unlike many other places, tenure, intelligence and hard work can be loosely correlated with outcomes in the Bay. Living through a societally transformative gold rush in that environment can be paralyzing. "Am I in the right place? Should I move? Is there time still left? Am I gonna make it?" It psychologically torments many who have moved here in search of "success".

Ironically, a frequent side effect of this torment is to spin up the very products making everyone rich in hopes that you too can vibecode your path to economic enlightenment.

[ChatGPT @ChatGPTapp]

A preview for Pro users: a new personal finance experience in ChatGPT.

Pro users in the U.S. can securely connect financial accounts, see where their money is going, and ask questions based on the information they choose to connect.

Your full financial picture, now in ChatGPT.

[hari raghavan @haridigresses]

In case you're wondering, this is the stage of the market we're at.

[Howard @howardx30]

Long Lake has a sick business at scale

[Dovey "Rug The CNY" Wan🪐 @DoveyWanCN]

作为Amex重度用户, Long Lake 收购 Amex GBT 这个例子非常值得看. 最近看看到了好几个相对应的例子去呼应 “15年前投fintech的一个learning就是就是作为“搅局者” 的startup最大的upside不是去服务incumbents (譬如银行, 保险公司, 借贷公司) 而是要成为为这些服务本身”的例子

corporate travel 这种线下 touch points 极多、供应商网络复杂、已经有稳定现金流但运营极重的传统服务行业，可以通过 AI 重写 workflow、压低人工密度和运营摩擦，去做组织效率重组和 P/E expansion. 肯定会有很类似Long Lake这种AI first PE 去做这种资本+技术双线运作的操盘 (Long lake已经收购了30多家类似的传统行业了)

很多传统行业已经有客户、现金流、牌照、供应商关系、线下服务网络, AI无法不是凭空创造这些线下资产, 所以买下一个低效但真实存在的系统，在用AI去降本增效，往往比从零造一家公司更快、更现实

这也是为什么我越来越觉得，AI 对非科技/科研行业最大的价值捕获并非AI Native的机会, 这个和互联网时代可以创造很多新的需求不同. 而是发生在谁拥有真实世界的交付网络，谁就更有机会把 AI 的效率红利吃到

[Avi @AviFelman]

The next mega bubbles that will help you escape the underclass:

1. Robotics (producers & picks and shovels — what goes into production?)
2. Space (space n aliens n stuff)
3. Energy (Uranium, battery companies, etc)
4. World financialization (stablecoins)

After that it’s over, you either made it or you didn’t

[Dovey "Rug The CNY" Wan🪐 @DoveyWanCN]

翻到了10年前当时在看 Cerebras Series B 的聊天记录，such a full circle/ 2017年6月，历史正在悄然发生分野。Attention is All You Need arXiv 上架，Transformer 架构当时工业界的重视程度远不如后来其对第四次工业革命的影响来的深远, 所以我对Cerebras的未来也并未觉得明朗

回看当时的判断，充满了时代的局限：那时我还在怀疑 Nvidia 是否具备支撑大规模数据中心的基因，Google依旧是那个自给自足的Google (G家TPU到现在也不会大规模对外), 觉得Cerebras 的“晶圆级引擎”是一个巨大的hype. 时代就像一个巨大的消音器，三维的我们永远听不见四维的声音

2017-2026的十年，是AI 时代完美诠释了海明威笔下的那种改变：“Slowly but surely, then suddenly all at once.” 再往前四年, 2014-2023的10年, 是加密数字货币的黄金10年

几乎每个指数型增长的行业的前 7 年都是漫长的伏笔. 后 3 年则是摧枯拉朽的爆发。过去15年作为前沿科技的观察者, 最迷人的时刻莫过于此, 曾亲手触摸过历史的草稿, 近距离观察在历史中迷失彷徨但依旧坚持的founder, 还有各种因为一个极其随机的选择, 与不可能想象的财富擦肩而过, 或者后知后觉获得了不可想象的巨大财富的人们

第四次工业革命的烟花, 必然一样灿然

[Elad Gil @eladgil]

People at major AI labs (using internal models) 3-4 months ahead of startup silicon valley engineers

SV founders/eng 3-6 months ahead of NY

NY founders/eng 6-12 months ahead of rest of world

Most people have no idea how fast AI shifting as 1-2 years behind SOTA

"The future is here, just not equally distributed" - Robert Heinlein