@Fabian_v4 @ContrarianCurse Not surprising given bullish stance on OAI
English
Frank
819 posts
Frank
@Lordvultures
Former capital allocator. Account size is retail’s biggest edge.
Katılım Kasım 2011
1.2K Takip Edilen169 Takipçiler
$6965 zinggggg
I can’t believe I didn’t used to invest in Japan. There is so many good cos
English
Frank retweetledi
@dennisihong This is all assuming no nominal growth. If you account for real growth from productivity enhancements / new economic acitivites (e.g. new scientific discoveries) + monetary easing (driven by the deflation aspect you mentioned), the math falls apart very quickly.
English
Some thoughts captured in a recent discussion with investor friends:
The long-term issue with AI is that as Chinese labs close the capability gap, mature use cases will increasingly overwhelm frontier ones and mature use cases don’t generate much economic value.
Consider translation, the first job meaningfully automated by generative AI. A good translator once cost $100k/year and produced around 1 million tokens of work. Today, an AI model earns roughly $0.40 for the same output, heading toward $0.01. Demand has grown 5–10x, but nowhere near enough to compensate. This pattern has historical parallels. The telephone switchboard operator and the stock chart clerk were also displaced by technology that captured close to 0% of their former wage.
The same dynamic is playing out now with junior analysts, bookkeepers, and coders. These are “frictional” jobs and they process and transform information but don’t create compounding value. Once many models can do the work, pricing converges toward the cost of electricity.
Coding looks like the next translation. Models currently consume a billion tokens to produce a million tokens of usable code with a ratio that’s improving fast. Demand for code will rise, but code, like translation, isn’t something you produce endlessly for its own sake.
This puts a hard ceiling on the TAM.
Global salaries are $45 trillion, roughly two-thirds white-collar. Much of that is irreplaceable: managers, salespeople, R&D, and no small number of roles that exist purely for relationship or political reasons. Realistically, maybe 20% of the $30 trillion white-collar pool gets automated by 2030, and much of what does get automated is worth barely more than the compute cost to replace it.
That makes it very hard to justify $1.3 trillion in AI capex this year, let alone 30% annual growth from here. If this view is right, you can be bearish on SaaS multiples and AI infrastructure simultaneously.
The automotive industry destroyed the horse but didn’t ultimately create much shareholder value. AI may follow the same arc.
English
@chrisbarber Wow words out of @insane_analyst that I can actually understand - need definitely need more of this!!
English
@FarrerWealth SK Hynix
Hyundai Motor
Samsung Life
Short: Ecopro
Caveat: not an expert
English
Anyone who knows the Korean market well - what companies are worth studying?
English
@Merridew__ Real estate didn’t trade at mid-teens cap rate back then though
English
@pandawatch88 Didn’t trust Gavin Baker’s micron call last year? Don’t worry Brad Gerstner got you this year with Hynix! 😂
English
Sorry, forgot to add for the "bUt thE PE is still loW!" enjoyorrs, that Moderna peaked at 8x FWD PE, before proceeding to lose 55% of value in 3 months.
True, SK Hynix still FWD 5x, so don't worry (i'm still long).
At 8x PE Moderna had gone up 550% in 1 year.
At 5x SK Hynix is up like 750% in 1 year.
yes yes, can't compare HBM production by 3 guys with vaccine production by 4 guys. Or can we? who knows.
goldenlabubuwatch@pandawatch88
To help us visualize potential pain ahead, let's look at recent examples of massive investment craze leading to bottlenecks, and what happens after. Covid is a good example: -big pocket bluechip pharmas developing and manufacturing vaccines like crazy, antivirals, etc, revenue flushing in, capacities expanding, every pharma service provider hired, equipment suppliers backlog full etc -add to this all the associated hot VC money going into biotech investment in mRNA on other modalities after the next big thing Wuxi Apptec was a big beneficiary of all this spending: -2020 revenue up 28% -2021 up 38% -2022 up 71%, accelerating all the way -growing margins through pricing power: 2022 GP up 75% in 2022 and NI up 100% When did share price peak? In 2021, at 100x PE. Well ahead of their best massive year which was 2022. Well ahead of peak margins. People looked not one but two years ahead, and were correct. What happened? Share price down 50% in 2022, even as profit doubled. Then a further 60% down over the following 18 months, over 2023 and 2024, as profit flatlined (note profit didn't even come down). There were also some specific geopolitical issues on the name too. But big picture still holds. We can only stretch the Covid and Wuxi analogy so far. Don't reply with an essay on why "AI buildup is not like covid", I know. So, as you look at your gains, what may decide their future is not 2026 FY results, nor the nice 2027 bullish hyperscaler capex predictions, but actually the general mood about what's going to happen with 2028 earnings. Because if by 2H26 the market starts to sniff 2028 could be flat vs 2027 (not even contract, AI is not going away, but are today's pricing power and bottlenecks lasting for 2 more years?) we could move down quickly. Or maybe not, who knows. Sleep well my bull.
English
@pandawatch88 @LowAlphaHighVol You laugh but they actually do this with incredible consistency. I like to think of them as an elite pod set free from tight drawdown limits.
English
@LowAlphaHighVol I respect a man that during Q1 buys a stock that has gone up 20 times by selling one that has gone up 2 times. Gigachad.
English
WT Capital just rotated its entire 13F book.
❌OUT: NVDA, AVGO, TSM, GOOG, AMD, CIEN, CRDO, ALAB, VRT, GEV
✅IN: LITE (20% of 13F filings), COHR, SNDK, MU, GLW, MOD, TER, KEYS, AXTI
GPU compute → optical / memory / cooling / test.
English
@LowAlphaHighVol I’m just sad that WT back office decided to leave this out of their month-end to do list this quarter for an early labour holiday
English
My 13F szn for Asian hedge funds I track begins today with Trivest. AI degens will be happy to note the top buy is $MU while $LITE is the third biggest buy and there's a new position in $COHR. $TER is also new but methinks a bit late to the game here. Fully sold $BABA & $GOOGL
English
@ZeeContrarian1 @surajsippy Him and Howard Marks are literally such a waste of time to listen to… we need more airtime given to people like Druck and Tepper
English
@surajsippy I don’t think Ray Dalio has really been involved in actual trading for the past 10 years.
At this point, he’s mostly become an ambassador for macro investing and economic thinking.
English
We all grew up on Ray Dalio.
To be honest, I even interviewed at Bridgewater a few times, which is a funny story by itself.
And what I’m about to say is just one observation.
The problem with these macro legends is that they sound incredibly smart, they speak incredibly smart, and most of the time they just end up disturbing your investment process.
Because at some point they stop being investors or traders and become commentators.
They’re usually semi-retired, managing reputation more than risk, and every time they open their mouth it somehow becomes one of the greatest contrarian indicators on Earth.
A few classics.
2018:
Dalio warned about late-cycle conditions, tightening liquidity, and structurally lower future returns.
Market response:
S&P up 31% in 2019.
2020:
Dalio warned future returns would likely stay low because rates were near zero and asset prices were elevated.
Market response after the COVID crash:
One of the greatest bull runs in modern history.
2021:
Repeated warnings about bubbles, money printing, and expensive valuations.
Nasdaq response:
Up another 27%.
2022:
Dalio:
“The stock market offers about a 5 to 5.5 percent expected return which is pretty low.”
Market response:
The market bottomed shortly after and the AI rally began.
2024 and 2025:
Again warning that equity risk premiums are too low and long-term returns from these valuation levels will disappoint.
Meanwhile AI stocks continued going vertical.
To be fair to Dalio, he’s usually talking about long-term annualized returns from current valuation levels, not short-term market direction.
But markets can stay expensive for years while continuing to compound higher.
And the last decade taught me one thing.
Listening too much to famous macro people is one of the fastest ways to miss massive moves.
Ray Dalio has correctly predicted 14 of the last 2 bear markets.
English
@origoinvest @kieranwgoodwin He kinda is because oaktree is big in distressed so just pitching his opportunity set
English
And Oaktree is a manger of BCDs
Wall Street Rollup@WallStRollup
Oaktree's Armen Panossian on what the next round of BDC selling could look like: "Those sellers that sold, they sold out of their best asset pool. So the next round, who knows." "What I'm more surprised about is that the banks have not tightened the screws as much on the lending that they provide to the BDCs”
English
@LowAlphaHighVol Any interesting names they like that aren’t all over fintwit yet?
English
April was just another normal month for Asian hedge funds:
CloudAlpha Singularity +13%
Keystone +15%
GH China Century +20%
Grand Alliance Alpha Plus +23%
WT LS +28%
E20 +35%
Anatole +38%
English
@Merridew__ is there an issue with decile bracketing using ex-post return information? If it keeps going up then it would drop in decile ranking making mid deciles look more positive, and if it drops then it would be the top decile, which is self-confirming
English
@pandawatch88 Any views on chery auto? I trusted hillhouse on the ipo and got left with a cheap bag that kept getting cheaper 😂
English
Geely just reported too, flat volumes (this was known), revenue up 15%, profit up 31%.
In this economy.
Nobody cares in the US cause the dodgy ADR is up just like 1.5%. Maybe it's all priced in. Maybe is not.
goldenlabubuwatch@pandawatch88
BYD reports most horrific results in a while (as expected), price up 4%. Also one of the very veery few China consumer names up YTD (alongside the other EVs). Basically since Dec/Jan, after a rough 2H25, everyone positioning ahead for the domestic EV market to recover in a couple of quarters (and exports to continue). Maybe. We'll see. I added because wisdom of the crowds and stuff. Also option value on Trump Xi meeting announcing Chinese OEMs can create jobs in US (very low chance, UAW etc midterms blah blah, but who knows). One would imagine UAW would love to be part of a communist enterprise (until they find out the definitions may differ).
English
