Exploring AI & financial markets Learning, building, and sharing along the way
Starlink
Joined March 2025
- Tweets2.8K
- Following130
- Followers406
- Likes5.6K
Pinned Tweet
My X circle is basically AI researchers, Tesla bulls, finance bros, and anime avatars.
Somehow, I fit into all four. 💀
🔗 circle.catsuki.cc/zh
Everyone wants a beautiful, wealthy partner. But first, you must become someone worthy of them.
If you want to be with someone exceptional, you must elevate yourself to their level. Otherwise, your efforts will be in vain.
Even if you force a relationship, it won't be sweet. A fruit picked before it's ripe will never taste as good as one that has matured naturally.
Only when two people are truly compatible can their relationship last and bring lasting happiness. If they are mismatched, even if they get together, they will soon part ways.
The same principle applies to investing.
Be a good person. Buy good stocks. Good people and good stocks bring good returns.
Only the finest investors are worthy of the greatest companies. They must be a match for each other.
Dragons seek dragons. Phoenixes seek phoenixes. Dragons may pair with phoenixes. And mice will always dig their own holes.
That's simply how the world works.
Why are global markets falling? Let's break it down.
1. US Inflation
September CPI rose 3.7% YoY, above the 3.6% consensus, raising expectations of a Fed rate hike this month, although the probability has eased from 50% to 40%.
The inflation report was disappointing. Services inflation remained above 5%, while core goods prices failed to cool as expected.
Inflation is proving stickier than anticipated. That's the immediate trigger behind the US stock market selloff.
2. The Elephant in the Room
Then there's the other, widely understood factor driving markets lower across the US, Japan, and South Korea.
Multiple forces are hitting global equities simultaneously.
The damage so far:
Philadelphia Semiconductor Index: −4.3%
Nasdaq: −1.88%
South Korea: stabilized after an 8% drop
Japan: began stabilizing after falling over 5%
The key is understanding WHY markets are falling.
Once we identify the causes, we can respond accordingly.
The worst kind of selloff is one where prices keep falling and nobody understands why.
This time, the picture is clearer: stubborn US inflation, combined with broader external pressures, is driving the synchronized decline across US, Japanese, and Korean markets.
What makes a good company?
A company that creates value for its customers, employees, shareholders, society, and the planet.
A company that harms them is a bad company.
When the world abandons a good company, that's when your character is truly tested.
Do you offer a helping hand or kick it while it's down?
Do you have compassion, empathy, kindness, and love?
When a prince falls from grace, do you offer your hand or draw your sword?
When a princess becomes Cinderella, do you ask her to dance or make her do the chores?
That's how you tell good people from bad.
Only good people have the compassion to stand by a good company when everyone else walks away.
Only good people can truly recognize good companies.
As Charlie Munger taught us:
Become a better person, and you'll recognize better companies.
Birds of a feather flock together.
It's that simple.
Hot take:
AI won't replace humans, it'll replace the economic value of being average
And most people are still preparing for a world where average is enough
The most dangerous blind spot in tech is assuming the median consumer experience represents the frontier reality
Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel.
- Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all.
- Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract.
- Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code.
- And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while.
The funnel is driven by a combination of factors:
1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel.
2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel.
3. Access. Free-tier, paid-tier, internal.
So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.
The Tree of Wisdom from Plants vs. Zombies is the best AI model ever, powered by fertilizer instead of tokens
Hot take:
The first generation to grow up with AI will consider thinking without it an extreme sport
Hot take:
AI won't replace human thinking
It'll make thinking a luxury
The rich will pay to think for themselves, everyone else will pay subscriptions to avoid it
While half of silicon valley is busy wrapping proprietary apis and fearing the next dev day deprecation, open-source teams under hard constraints are rebuilding the foundation with real architectural gains and radical cost efficiency. Hard constraints breed real innovation
Keep Faith In DeepSeek and Open-Source AI
The Chinese companies are the only ones that are banned from wrapping Anthropic or OpenAI models into random features
They actually have to deliver extremely performant and efficient models to succeed
So far, they have been doing really well. DeepSeek, GLM, and Kimi will only get better 🚀
Today's market selloff is, first and foremost, a reaction to surging oil prices.
I've been saying "It's all about oil" for some time now, and more investors are starting to recognize its importance.
The second major factor is US Treasury yields.
Just look at how high yields have climbed:
• 30Y: 5.71%
• 10Y: 5.32%
• 5Y: 5.05%
• 2Y: 4.78%
• 1Y: Above 4.4%
With Treasury yields soaring across the curve, the market's decline is hardly surprising.
More investors are beginning to recognize the enormous pressure that rising oil prices and Treasury yields are placing on financial markets.
If these two pressures ease, markets could find some relief. If they persist, the pressure on equities will remain.
Two things to watch: Oil prices and Treasury yields.
Until they stabilize, staying defensive makes sense.
Global markets are now facing surging U.S. Treasury yields and renewed pressure from potential Fed rate hikes, even with the possibility of a pause in October.
At the same time, rising bond yields globally are putting increasing pressure on valuations across capital markets.
The broader backdrop is shifting:
from offense to defense.
OpenAI just casually solved a ridiculous number of math problems today.
Like… what the hell happened overnight 😭
We’re watching the math frontier move in real time.
U.S. stocks keep hitting new all-time highs, but the number of individual stocks making new highs is shrinking fast
The index is being pushed higher by a small handful of names, while most stocks have already entered bear-market territory — now 51% of the market
This is the first time we’ve seen this since the dot-com bubble burst in 2000
The bull market keeps getting narrower. At this point, it’s becoming extreme
If only a tiny number of stocks are driving the index to new highs, then the index itself becomes less representative of what’s actually happening beneath the surface
This may not be a popular view, but facts matter more than sentiment
Time to shift from offense to defense
The AI model race is getting kinda boring
The real game now is turning AI into something people actually use every day
Models become infrastructure, products win
Every AI lab has three models:
the public model
the internal model
and the model employees vaguely hint at on podcasts