勃勃 CO

看都不用看，无脑买GOOG就好了， META招人失误AI赛跑倒数第一，TSLA天天吹牛卖不出去即将倒闭。

My hottest and most wrong (year to date) take is also probably the only semi stock that Zephyr and I currently disagree on. Since December I’ve been bullish QCOM and he’s been bearish. Of course, he has been hugely right and I have not been - but I really think that’s going to change over the next 12 months. I also like owning it because it’s a pretty popular funding short that sims a potential beneficiary of any breakthrough in edge (or breakthrough in memory efficiency like if someone can get PIM to work). So you could imagine a scenario where any significant progress for on-device inference leads to moderating expectations* for data center growth, the popular longs take a hit and QCOM squeezes hard both on improving outlook and a squeeze. It’s also just cheap af and eventually the circumstances causing its underperformance will abate and people will still buy phones and eventually it’ll have its day. In the meantime, I see it kind of like a positive expected value hedge against a shock to estimates that could end up being too optimistic if just, like, 20% of inference is done on device. (*I really do mean moderating, I don’t think edge AI / on-device inference means we don’t use data centers anymore obviously. We’re not going to use the AI on our mobile phones to design drugs or run complex systems obviously, but I could see some revisions when some of the demand of “hey ChatGPT, ” isn’t running through the cloud)

目前市场的情绪面都在AI基础设施上，像英伟达，AMD，美光，闪迪，CLS, LITE这些卖铲子的都在赚钱，并且享受市场溢价。 其他板块，比如金融科技板块就遭遇冷落，Sofi财报良好，却因为没有上调指引遭遇抛售。我今天自己小幅度加仓了Sofi和Hood。 我不喜欢在股票高速上涨的时候去追右侧，所以我没有LITE, STX，SNDK。英伟达美光CLS我都有，但都建仓较早，所以我拿的也比较稳。 相比去追已经错过的火车，我更愿意把钱放在被市场冷落、但基本面扎实、还在高速成长的好公司上。