John Hund

@bennpeifert Coming from emerging markets, I know one fact. Fast or slow, correlation collapse/convergence comes for us all. Good luck!

the amount of lovely outreach from all corners of finance and otherwise has been wonderful. we have so many friends and many people have loved following us and our content and it's just been fantastic.

Good morning my loves, happy Saturday. Sorry I've been quiet, obviously been busy, but thought it'd be nice to give you all the details on the multi-strategy absolute return program that experienced the 28% drawdown this year. (1/n)

@KhoaVuUmn The proof is immediate from observation (Figure III).

@ClueHeywood You think you are unhappy. I WAS happy, off a boat in the Keys, shark fishing, Friday beers. And then it happened. Pure tragedy. The opening notes of Doves Cry from the two piece cover band. Played it. Sadness.

Had the kind of week which made me think, multiple times, that maybe I’d be happier herding sheep alone on some remote island. That said, it’s finally time for some delicious ice cold domestics, yes that’s right it’s time for FRIDAY BEERS.

So I’m in the Florida Keys, and there are many things we all can disagree about. Fed Policy. All of politics. But can we unite slowly just under one flag? Stop fucking covering Prince. You can’t do it.

@michaelschwab13 At least he ended Cal Raleigh’s slump /s

Not good.

@heynavtoor @rolandsmartin Maybe an AI influencer could use a more recent model to do math, and not use AI to write an insanely overly long post about a 2+ year old paper. epochai.substack.com/p/gpt-54-set-a…

🚨SHOCKING: Apple just proved that AI models cannot do math. Not advanced math. Grade school math. The kind a 10-year-old solves. And the way they proved it is devastating. Apple researchers took the most popular math benchmark in AI — GSM8K, a set of grade-school math problems — and made one change. They swapped the numbers. Same problem. Same logic. Same steps. Different numbers. Every model's performance dropped. Every single one. 25 state-of-the-art models tested. But that wasn't the real experiment. The real experiment broke everything. They added one sentence to a math problem. One sentence that is completely irrelevant to the answer. It has nothing to do with the math. A human would read it and ignore it instantly. Here's the actual example from the paper: "Oliver picks 44 kiwis on Friday. Then he picks 58 kiwis on Saturday. On Sunday, he picks double the number of kiwis he did on Friday, but five of them were a bit smaller than average. How many kiwis does Oliver have?" The correct answer is 190. The size of the kiwis has nothing to do with the count. A 10-year-old would ignore "five of them were a bit smaller" because it's obviously irrelevant. It doesn't change how many kiwis there are. But o1-mini, OpenAI's reasoning model, subtracted 5. It got 185. Llama did the same thing. Subtracted 5. Got 185. They didn't reason through the problem. They saw the number 5, saw a sentence that sounded like it mattered, and blindly turned it into a subtraction. The models do not understand what subtraction means. They see a pattern that looks like subtraction and apply it. That is all. Apple tested this across all models. They call the dataset "GSM-NoOp" — as in, the added clause is a no-operation. It does nothing. It changes nothing. The results are catastrophic. Phi-3-mini dropped over 65%. More than half of its "math ability" vanished from one irrelevant sentence. GPT-4o dropped from 94.9% to 63.1%. o1-mini dropped from 94.5% to 66.0%. o1-preview, OpenAI's most advanced reasoning model at the time, dropped from 92.7% to 77.4%. Even giving the models 8 examples of the exact same question beforehand, with the correct solution shown each time, barely helped. The models still fell for the irrelevant clause. This means it's not a prompting problem. It's not a context problem. It's structural. The Apple researchers also found that models convert words into math operations without understanding what those words mean. They see the word "discount" and multiply. They see a number near the word "smaller" and subtract. Regardless of whether it makes any sense. The paper's exact words: "current LLMs are not capable of genuine logical reasoning; instead, they attempt to replicate the reasoning steps observed in their training data." And: "LLMs likely perform a form of probabilistic pattern-matching and searching to find closest seen data during training without proper understanding of concepts." They also tested what happens when you increase the number of steps in a problem. Performance didn't just decrease. The rate of decrease accelerated. Adding two extra clauses to a problem dropped Gemma2-9b from 84.4% to 41.8%. Phi-3.5-mini from 87.6% to 44.8%. The more thinking required, the more the models collapse. A real reasoner would slow down and work through it. These models don't slow down. They pattern-match. And when the pattern becomes complex enough, they crash. This paper was published at ICLR 2025, one of the most prestigious AI conferences in the world. You are using AI to help you make financial decisions. To check legal documents. To solve problems at work. To help your children with homework. And Apple just proved that the AI is not thinking about any of it. It is pattern matching. And the moment something unexpected shows up in your question, it breaks. It does not tell you it broke. It just quietly gives you the wrong answer with full confidence.

@SeanTPendergast Is that Angels logo in crayon? It’s gotta go.

Easily the Mariners. Looks like a sign on the front of a third rate fast food joint that you took the whole Little League team to after their games in like 1981

Here's the somewhat troubling news in the jobs report: The unemployment rate fell to 4.3%, but not for great reasons. There's a big drop (almost -400k) in the labor force. The labor force participation rate also fell. It appears people stopped looking for work in March or perhaps more migrants left the workforce (or both). This data can move around a lot month-to-month, but it's worth keeping an eye on.

@Acyn @MalcolmNance This should be the last time Brianna Lyman appears on CNN.

Lyman: It's hard to take seriously half of the country's dissatisfaction with the president when they were rallying against war with an actual nuclear power. Blow: You think that we shouldn’t have been supporting Ukraine? Lyman: I do not think that was our war. Mockler: So you guys are going to complain in one breath we shouldn't support Ukraine at all. And then you'll be like, wait a minute those same allies refused to come help us?

@herr_naphta @tysonbrody Amazingly, that paper doesn’t cite this much earlier work finding much the same thing (using Craigslist as an IV for decline of circulation), but also extending to the cost of bond financing. sciencedirect.com/science/articl…

@tysonbrody The econometrics on this are very solid as well: academic.oup.com/restud/article…

Everyone hates on Facebook and Google, but Craig Newmark was the man who invented making a billion dollars by killing newspapers.

@astros @MethodistHosp Do better, Stros

W is our favorite letter. #ChaseTheFight x @MethodistHosp

@CliffordAsness I’m just really sad that beta decided to become less forgiving. Beta, who hurt you?

It’s not just cringey (though it is massively so) it’s also chock full of total bullshit.

Hey @malcolmwardlaw, the day we’ve been waiting for has finally come, the present IS the past, Claude for COBOL is here! streetinsider.com/dr/news.php?id…

@Seahawks @Gatorade @TerryCollege Represent. Finance FTW.

Mike Macdonald. Super Bowl Champion. Presented by @Gatorade

I’m honestly missing the Clydesdales and a dude driving a Ford truck right now. Just because you can AI doesn’t mean you should AI.

@BobStelton @Seahawks Still stressed. Cracked another Fremont Lush.

Phenomenal first half for this Hawks defense! Pats defense is no joke. How are we feeling @Seahawks fans?

@macrocephalopod Sure, but how was his cursive handwriting? I’m sure you must have asked about that?

Interviewed a guy for a quant research job today. Couldn’t invert a 3x3 matrix by hand. Couldn’t derive Black Scholes. Couldn’t explain the in-place quicksort algorithm on the whiteboard. Just kept talking about “stochastic gradient descent” and “reinforcement learning” If you can’t do the basics, what’s the point? Obvious no hire.