@photonjaeger Well said. Although I feel that my experience is accumulating much faster than how it can be distilled. 😂 - just kidding, need a bit humor for a day like this.
@ShanghaoJin Exactly! CEOs are very much concerned by the risk of disruption from startups or faster AI adoption from competitors. The risk of not being relevant due to inadequate AI adoption is much bigger than the increased cost of AI. It’s a race to survival!
Herman's analysis was very insightful. If you look at what's happening in the earnings, those direct indications of AI adoption are only getting stronger. The big negative news (e.g. GOOG's search, INTC's PC miss) are actually collateral damages of strong AI growth.
If there is any lesson I should learn from this incident, it is that macro always prevails
我其实已经看到了严重的流动性问题,但太有点侥幸AI earnings能不能带板块幸免于难。当时尤其被得到Anthropic ARR重新加速、HF & fundamental持仓轻的信息给encourage了。并认为Ant为了上市前会给一些“甜头”,而接下去ER与信息大都走反。在macro liquidity严重紧缺的时候任何多头容错都很低,乃非对称性风险。我手上的tail risk CDS hedge都远远不够cover loss
Looking forward,市场会如何?显然macro的impact只是缓解没有消失。乐观方面是:IR vol冲高慢慢回落,Credit spread开始改善。但同时,在Ant上市前,股市内部的流动性其实开始变紧——投资人卖出CSP semi准备买Ant。流动性差的时候市场更容易质疑 Y28 forecast和FCF。这下半周,eq IV开始上升,我怀疑一些“hunting elephant”已经开始要得手
期间我退出了很多Goog和一些其他股票。但是还没有deploy这一部分saving capital,主要因为那部分钱更多以长期saving为主,而且在观察deleverage情况
AI泡沫是不是破灭了?我甚至不觉得是泡沫。我们与2000年有完全本质的区别——最前端labs是有盈利的
我发现近期金融市场流动性 VS 科技公司流动性冰火两重天——某些意义上是科技公司i.e. Ant融资,吸走了流动性导致股市流动性变差。我前几天甚至有些担心,垃圾小型neocloud暴雷会导致CDS整体变宽——private loan fund赎回——Fire sale中出现芯片折价——再引发semi整体残值重估。但是一个CSP的好朋友告诉我:现在leather jack的卡就是硬通货,会被立即抢走,而且CSP很轻松能调用各种外部workload。所以实体不像00年,不缺钱、能融到钱(MSFT 25b长债)导致行业 vs 市场偏差,导致算力需求曲线不但供不应求,且forward contango
金融市场表现得像今天这样是一种流动性不足下的错杀,其原因:1. 市场没有做好准备追捧科技重资产模式;2. AI相关融资没有波及行业外公司——仅仅由AI/semi承受了全部冲击。市场流动性紧而不破已经一个月了,维持这种流动性严重磨损伤害实体,不过同时很难持久。市场要么短期被全面扩散,把包括AAPL、GS全部砸下去;要么流动性改善带动空头平仓 & 极低的AI仓位回补(上周前4天HF已经开始净买入了)。我认为下周后者概率更大
@ShanghaoJin https://t.co/7laMqs6n87 正好最新一期的Invest Like the Best也说到这个理论。When the cost of intelligence races down, you'd want to own the complements to intelligence.
Anthropics shared that their engineers' productivity has gone up by 45x compared to 2024. Why don't we use that sample to infer what might happen in other places?
@spcohn@SiliconValleyPX Indeed. I've seen much better outcomes from many teams in much bigger companies. His CTO may be telling the truth but that's a sample of one.
@spcohn@SiliconValleyPX Indeed. I've seen much better outcomes from many teams in much bigger companies. His CTO may be telling the truth but that's a sample of one.
@spcohn@SiliconValleyPX Indeed. I've seen much better outcomes from many teams in much bigger companies. His CTO may be telling the truth but that's a sample of one.
I am shocked, absolutely shocked, that Anthropic ARR has likely gone vertical in October after they cut token prices.
Except the exact same thing happened for the entire AI market, which is why GPU rental prices have surged while token prices declined.
$100b plus in year-end ARR looks possible again.
Reading ANT's S-1, the single biggest takeaway for me is the power of the scaling law. It's THE engine driving AI growth. Human intuition struggles with exponential curves, which explains the constant market doubt and disbelief like what we are going through right now.