Calling The Bottom + Trade Alert
We hit peak fear, drivers make it rain, an update on the AI race, and I take profits on the only profitable thing in 2026.
Welcome to this week’s edition of The Macro Obsession.
The best round-up of current events and trends in finance, tech, and the real economy currently in your inbox!
Issue #38—Week of March 23rd, 2026
Sell, Mortimer, Sell!
Leaving Chat in the Dust (Redux)
A Win for Uber and Uber Drivers (But Mostly for Uber)
TMO Model Portfolio Trade Alert
Sell, Mortimer, Sell!
It’s been a few weeks since “Tech Bears Called The Top” (TMO #33), but they are still piling in. No wonder why, given the current geopolitical situation.
Every day, we grind lower and lower across every asset that isn’t plain old dollars. Paging “Safe Haven No More (Revisited)” (TMO #36), where I argue that gold was too bought up to be a real hedge going into the Iran war this past week.
Lo and behold, the gold bugs were exterminated.
Fear is rampant on Wall Street. Hedge funds are now short at levels unseen since the last bear market (the bottom of the bear market, mind you). These shorts were built up all year, but they have now hit levels where fund managers are going to ask what the exit plan is.
What put delta1 is “properly hedged?” Plenty of strategies may run 10, even 20, or upwards of 30 if they are conservative. What would we call this? Ultraconservative? Alt-right levels of hedging?
Worst news of all, even passive flows have slowed, and active flows have flipped so negative that the total weekly ETF flow was negative. That’s the first time since Liberation Day, and to a far greater degree. When retail is net selling, you know it’s getting bad.
Even The Economist called the top. Here’s last week’s cover.
If you know the data behind this, then you already know what I think about it.2
The alignment of these metrics is textbook market bottom behavior.3 The move went through institutions, and they are at max capacity for short positions and put delta now. It’s reached retail, and they can only sell for so long.
Passive flows are a monster that will not stop at the end of the day. It can slow and pause for a moment, but only for so long. All institutions have to do is start to cover; retail will resume buying into the first sign of positive momentum.
I am more bullish on risk now than I have been in a while.
The fact that gold couldn’t break through $4,400 despite having all the negative momentum it needed was a great sign. We’ll see if that holds.
Leaving Chat in the Dust (Redux)
It’s time to revisit our good old friends OpenAI and ChatGPT, the bot on top, that we learned back in “Leaving Chat in the Dust” (TMO #29) was rapidly losing market share to Google’s GOOGL 0.00%↑ Gemini.4
Instead of clicking on that link to see the story, I recommend reading the whole newsletter—it’s AI content from top to bottom.5
It was a banger that deserves to be seen, so I removed the archive paywall forever.
But I want to update our priors on the “fastest-growing LLM,” because we officially have a new winner. Enter Anthropic’s Claude, which penetrated normie discourse after its falling out with the Department of Defense (Reuters).
When the public found out that you can be anti-Trump, use AI constantly, and feel coherent about your ideology, downloads took off.
Even before that, Claude had already started to take users en masse when it rolled out agentic coding tools. This is most interesting to enterprise clients, who increasingly are turning to Claude over Chat for their first try, defying the trend we saw in “ChatGPT Is Surprisingly Sticky” (TMO #24).
To be clear, ChatGPT still owns the largest share of the market, but it’s losing the interest of first-time customers. It’s not becoming a Band-Aid. That’s very bad for Chat, since that’s OpenAI’s strategy to ultimately win. Google is winning share too, but not nearly like Anthropic is. Claude is killing it.
I’ve avoided talking about The Moltbook (Wikipedia) to most people in my life, largely because they look at me like I’m insane. But the long and short of it is that it’s AI social media, and a ginormous waste of tokens. It caters to agentic AI bots that run on a model based on Claude, called “OpenClaw.” It’s had several other names, but that’s the one we’ve settled on for now.
Anyway, it’s become more popular than Linux and React on GitHub. It’s safe to say that Claude-like agents are currently producing more internet traffic than people. Insane times we live in.
I remain stalwart in my position that regardless of any gains made by AI, the majority of spending being done at present to operate LLMs is a waste. This subsidizing endless slop videos and text pukes done in the name of user acquisition is going to go down as one of the greatest acts of willful capital destruction in history.
Don’t get me wrong, I understand the “you gotta spend money to make money” strategy. But this is egregious.
A Win for Uber and Uber Drivers (But Mostly for Uber)
In some good news for rideshare drivers, of whom I know a few, wages are up! They’re not at pandemic peaks, but they’re up from any time post-2023.
Perhaps all the pandemic-era drivers have moved on?
Partly, it’s a result of platform fees being on the rise. I can’t say that they’re structurally higher, but they’re definitely not going down. I know technology is deflationary, but how much does that really apply to Uber?
The unfortunate part is that they are taking this opportunity to raise platform fees, and those go up way faster than driver pay. The disparity is huge; that’s where the tech part comes in, in the margins, I suppose.
So good news for drivers, but better news for shareholders.
More Below, But ICYMI
TMO Model Portfolio Trade Alert
The TMO Model Portfolio (“The Model”) has had a bloody month alongside most portfolios, and especially those that are not directly hedged. It is now flat year-to-date, up <1%.
To be fair, that’s 5.57% higher than the S&P 500 SPY 0.00%↑ , and 2.61% higher than our benchmark, a 70/30 portfolio of stocks and bonds, $FASGX. But a losing month is still never fun, and never ideal.
Since The Model’s inception in August 2025 (essentially along with this newsletter), it has never had a negative month until now. The combination of ex-US stocks (e.g., Europe, Asia), gold, and materials stocks all crumbling at the same time really hurt.
But losses are part of investing, and these are only paper losses so far, as none of them have been locked in and sold.
Actually, at the last trade alert in late January, we trimmed the precious metals and material stock positions because they had become too unwieldy.
This trade alert isn’t about locking in any losses, don’t worry.
Actually, part of the portfolio, an ETF in our Sector Rotation Strategy, has been winning all year, and especially in the last month. It’s up 32% YTD, and I’m taking profits on it.
For those new to The TMO Model Portfolio, a brief introduction…
The Model is made up of 10 ETFs and has low turnover—I average 2 months between trade alerts, and give an update in the first week of every month’s newsletter. Normally, I try to align trade alerts with portfolio updates, but the timing is right currently.
It invests across asset classes and geographies and aims to satisfy a moderate risk tolerance, which is why it competes against the 70/30 fund $FASGX.
Read more and see the last trade alert here:
Free subscribers, this is where we part. If you’re interested in investing like me and want to see the holdings of the portfolio and my thoughts on them, here’s a discount link that works for life6:
Thank you, and I’ll see you next week.
Paid subscribers, welcome beyond the paywall, and thank you for your support.
Use the link below to track the TMO Model Portfolio and its holdings, updated live:
