Pinned Tweet
Quality Equities portfolio at the end of October 2026.
Here is what I own:
- About 40% is in the big platforms: $AMZN Amazon, $META Meta, $GOOG $GOOGL Alphabet, and $MSFT Microsoft. They run the places where people shop, advertise, work, search, and rent computing power.
- About 25% in chips and chipmaking equipment: $NVDA Nvidia, $ASML ASML, and $KLAC KLA. Nearly every advanced AI chip depends on what these 3 make.
- About 19% in payments: $MA Mastercard and $V Visa. They take a small cut of trillions of dollars in spending and need very little capital to grow.
- The rest is $SPGI S&P Global, $UBER Uber, and $BN Brookfield.
Payments now makes up a bigger share, and Alphabet a smaller one based on how the market has decided to rate these companies.
The goal stays the same. Own a small number of the highest quality businesses in the world.
Howard Marks has a forecasting test in his latest memo that I keep coming back to:
"Tell me what you think the e.p.s. for a given AI company will be in five years. Will you bet that the eventual answer will be within +/- 50% of your figure?"
I doubt many of mine would pass.
1
2
665
His answer isn't paralysis. It's "calibrated humility regarding strategy selection, position sizing, and risk-taking" and "a willingness to act in the presence of uncertainty."
He also calls AI exposure an "unusually powerful 'fault line' that runs through portfolios."
1
185
Most of BJ's increase is one item. BJ's SG&A includes a ~$240M interchange settlement (net of legal fees), partly offset by ~$95M of costs it tied to the proceeds.
Net ~$145M, 46% of Q3 operating income (my calc). Adjusted EPS of $1.73 equals GAAP: it wasn't adjusted out.
1
1
256
TransDigm's capex was 2.6% of sales (my calc); acquisitions took $4.4B, about 16 times that capex (my calc).
Both are small: 0.28% and 0.22% of the $4.42B filing (my calc). $MA Mastercard is still the top line at 21.9%.
1
210
A 10% drop needs an 11.1% gain to get back to even. A 50% drop needs 100%.
I built this short clip to show why: the gain is measured from a smaller base.
Formula: 1 / (1 - drop) - 1 (my calc).
Why do you think losses hurt more than gains help?
1
469
13F caveats: long US-listed positions only, as of Sep 30. The filing is due 45 days after quarter end (filed Nov 16), so later trades aren't shown. Share counts only: no prices paid, no non-US holdings.
1
213

