Eric Simpson

[Eric Simpson @ProfectusHldgs]

Generational dip buying opportunity on $TSLA here

[skyte @ihateskyte]

Get your dopamine from what you create not what you consume

[Lisan al Gaib @scaling01]

short reminder that we are solving mathematics with cute sub 10T models

I hope you are prepared for 100T models and 1000x more compute spent training these models by 2030

[Just Another Pod Guy @TMTLongShort]

Maybe so many people are struggling to understand what model progress will look like because they don’t have the intellectual horsepower to self-analyze how their own intelligence works and why it’s superior to that of those around them.

Because if they did I think it becomes very obvious why model progress is unlikely to stall and why it radically changes everything it touches.

And yes I am well aware of how conceited that sounds.

[Dr. Julie Gurner @drgurner]

In a world where everyone has limited attention...

Do something that takes time to master.

[Gavin Baker @GavinSBaker]

Love it that people think this was an accident instead of super deliberate.

More effective than actually buying it.

[Boring_Business @BoringBiz_]

Completely disagree with this take

Ken Griffin figured out something that is incredibly simple on paper, but very hard to execute

That is

1. Hire the smartest people you can with a very high bar and focus on talent

2. Give them incredible incentives to pour 100% of their talent and effort into generating returns for you

3. Fire bad people quickly, do everything you can to retain the best people

This is much more difficult for companies to implement than one might think, especially at scale

[atlas @reatlashype]

Jeff Bezos - “If you have a 10% chance of a 100x return, you should take that bet every time."

[Noam Brown @polynoamial]

An internal version of Astra, @OpenAI’s next major model family, solved 10 major open problems in mathematics, quantum complexity, and theoretical computer science.

We believe it will be a major step for scientific reasoning. https://t.co/iP6cyheZ7i

[Just Another Pod Guy @TMTLongShort]

When someone is AGI pilled their entire lifetime calculus shifts.

They are willing to take more risk, pull forward consumption, and are susceptible to introducing volatility in everything they do.

Not a lot of people are AGI pilled so this is not an observable phenomenon at a societal level. Even in tech enclaves like SF only a handful truly believe instead of paying lip service.

That is going to change very rapidly.

And if you care about macro, politics and sociology you better start thinking about what this looks like now because when it happens it’s going to occur virtually overnight.

That Citrini post was just a taste.

Use your imagination. It’s free.

[David Orr @orrdavid]

The podshop model:

Every year recruit a ton of young people from "elite" schools. Put them through a boot camp. Give them some instruction on how to analyze and invest. Most of these people get fired within a few years, which filters the group down to skilled analysts.

After some time, let a skilled analyst run a tiny portfolio with extremely tight risk control. Extreme tight risk controls are meant to idiot proof running a portfolio. Most analysts fail at this step. Either they blow it on risk. Or they had no alpha and variance makes them a hit risk limit pretty fast on a bad run. This filters the group down a lot again.

Now what do you have? A tiny group of professional *analysts* who made it as "portfolio managers" working in tight guard rails. Most of these analysts have a tiny edge. Say it's 2.5%/year alpha on 100% gross leverage. So now you lever them up 6 times and voila you have 15%/year "alpha" before fees.

Besides that, the podshops constantly steal the truly talented PMs from each other.

-

This seems fragile to me. Examples:

* What happens to a ton of 6-8 levered portfolios someday when VIX spikes to 65 overnight on a true tail end event?

* What happens to a bunch of really thin edges per turn of leverage when the bull market finally ends, and markets become negative sum for a decade instead of positive sum?

* What happens if the factor hedges themselves get too crowded?

* What happens if AI turns out to be a better analyst than humans (it already is IMO), and someone figured out how to create a zillion of these super low edge analysts to turn into PMs using the same framework?

[Boring_Business @BoringBiz_]

People often forget that Peter Thiel similarly ran a hedge fund during the dot com bubble era that ended up failing

He then turned around and started one of the most successful VC firms in history (Founders Fund)

I fully expect Leopold to make a similar comeback, whether it’s VC, another hedge fund or a startup

He is incredibly talented and still extremely young. Lot of room to run ahead

[Fed @lord_fed]

This quote from Hedge Fund Market Wizards feels particularly relevant at the moment:

“Limit your size in any position so that fear does not become the prevailing instinct guiding your judgement.”

[Anders @AndersPicks]

There has NEVER been a walk off grand slam down 3 with 2 outs in an 0-2 count in the history of baseball

A kids dream

[BatmanCapital 🦋 @capitalatrisk23]

I am gonna tell my kids this was Charlie Munger and Warren Buffet

[BOSS @thebeautyofsaas]

no one tells you growing up that you don't really have to be clever or gifted to make it in life. you just have to be VERY curious about things that you find interesting. do it long enough, and everything will align to allow you to get all the things you are looking for

[High Yield Harry @HighyieldHarry]

Jokes aside, I hope Leopold gets another bite at the apple.

Obviously much smarter than the average 25-year old, needs actual risk controls and a PM though.

[Leo 的宏观笔记 @Leooo2099]

6天前他给投资人写了一封信："这轮下跌是买入机会。"

6天后他把所有持仓卖光了。

包括Anthropic的股份，也在找人接盘。

写这封信的人是管了240亿美元、上半年赚了439%的AI圈第一多头。

Leopold Aschenbrenner，前OpenAI研究员，写过AI投资领域引用率最高的那篇论文"Situational Awareness"。

2024年被OpenAI解雇后创立同名对冲基金，All in AI基建、电力、算力、存储、矿机。

上半年赚了439%，Stripe创始人、前GitHub CEO、Jane Street全是他的LP。

华尔街最炙手可热的AI基金，没有之一。

然后7月AI暴跌来了，四倍杠杆，三周时间。

SK海力士、Bloom Energy、CoreWeave、闪迪，持仓全是重灾区，CNBC确认，他已经把所有公开市场持仓一次性卖给了一个买家。

基金目前只剩私有资产，Anthropic的股份也在找人接盘。

7月24号那封信里他还在说"Anthropic IPO是下半年最大的催化剂"。6天后Anthropic的股份就成了他急着变现的筹码。

从"这是机会"到"全卖了"，6天。这就是四倍杠杆对一个人判断力的扭曲，6天前你还觉得自己是对的，6天后保证金通知替你做了决定。

他对AI的判断错了吗？没有。他的论文到今天还是AI投资的圣经。他赌对了方向，赌对了赛道，上半年赚了439%。

但方向对了不够，四倍杠杆意味着只需要跌25%就归零。市场不需要证明你错了，只需要在你扛不住的时候晃一下，你就出局了。

上半年赚了439%，三周后被迫清掉所有公开持仓。

方向从来不是问题，杠杆才是。

[threadguy @notthreadguy]

there it is

[SandemanStocks @Sandeman52]

$BE Bloom Energy is one of the best real time indicators of AI data center demand.

If the AI data center buildout had truly stalled, Bloom would have missed expectations, lowered guidance, and talked about customer delays.

Instead, they crushed earnings, raised full year guidance, and cited accelerating demand from hyperscalers and AI customers.

The narrative that AI infrastructure demand has stalled simply is not supported by today’s results.