Hey Reader - Thomas here.

Good day to my fellow compounders!

In 3-Bullet Sunday, I share with you the three most interesting piece of article, video or podcast I came across for the week on investing and growth philosophies.

Enjoy!

AI Stocks Are Having a Rout.

This week I spent half a day setting up Grok Bot as my AI assistant. The universe is funny this way, because the very next morning Meta announced Muse is here in Singapore, and of course I cleared my calendar again to set that up too.

So I spent the week setting up AI assistants in the same week AI-related stocks got hammered, with the Nasdaq having its worst day since June.

When prices fall like this, the brain treats a falling stock the way it treats a charging lion, and the only cure I know is knowing what you own. If you understand the business, a lower price is just a lower price.

The other thing that helps is a watchlist. If a business I’ve already studied gets cheaper, I already have a rough idea of what it’s worth, so I can decide without panicking.

This coming week I’m looking forward to reports from Shopify, Booking and MercadoLibre. All three sit right where AI shopping agents could help them or hurt them, so I’ll be reading closely.

In the meantime, the market is worried about what this sea change means for Google Search, and Alphabet fell more than 12% in a single day. So Alphabet is the next report I’ll be publishing.

Compound Steadily,
Thomas

P.S. If knowing what you own sounds easier said than done, my book The Lunch Break Investor is a six-step system for analysing and owning great businesses one lunch break at a time: steadycompounding.com/book​

I. Who Wins When AI Does the Shopping?

53% of consumers across 11 markets have used AI to help them shop in the past three months. That’s from McKinsey’s survey of about 21,000 consumers in June 2026, and the part that caught my attention is further down the funnel, where people actually let the AI pay.

In the US, the share of AI users who completed a purchase inside an AI tool went from 10% to 17% since McKinsey’s previous survey. Globally it’s 14%. Still small, but in the top six categories between 21% and 30% of AI users say they would let AI finish the purchase for them, roughly 1.5 times the number who actually have.

When AI shoppers were asked what had the single biggest influence on their most recent purchase:

AI chat or search tools: 29%
Review websites: 12%
Retailer platforms: 11%
Friends or family: 10%
Retailer AI assistants: 10%
Search engines: 9%

Search engines came in behind friends and family, which honestly surprised me.

Stand-alone tools like ChatGPT and Gemini get used about twice as often as retailers’ own assistants, but that lead shrinks at every step toward checkout, and McKinsey’s explanation for who comes out ahead at the very end is the part I’d click through for. There’s also a surprise in grocery and beauty, two categories that never took to e-commerce and are now catching up fast through AI.

​Click here to read​

II. Howard Marks on the AI Boom: Who Keeps the Savings?

Bubbles almost always form around something new, because the imagination can run as far as it likes. Howard Marks goes back to tulip bulbs in 1620 and the South Sea Company in 1720 to make the point with William Green. He says “You’re never going to have a bubble in paper stocks or timber stocks,” since anyone can work out how many houses will get built and how much wood goes into each.

For Marks, the closest match to today is the dot-com bubble, in kind if not in degree. The internet did change the world, and a lot of what people pictured in 1999 came true, e-commerce especially. With AI the picture is hazier. In his words, “I’ve never heard anybody tell me how AI is going to change the world,” and how anyone makes money from it is less clear to him.

Marks recalls Buffett making a similar point about the internet, that more productivity doesn’t automatically mean more profit. So if AI does cut costs, “To whom will the savings accrue?” The AI providers could compete on price until there’s nothing left for them, or the companies using AI could compete for market share and hand all the savings to customers as lower prices. That’s a pretty simple question, and honestly I don’t have an answer to it either.

The clip comes from a 2025 interview, so the gold and Bitcoin prices mentioned are from then. He also walks through the mistakes he’s seen in euphoria after euphoria, including what he calls the “lottery ticket mentality,” and shares his favorite fortune cookie.

​Click here to watch​

III. Meta Just Reversed a Decade of Buybacks

This one is my own Meta Q3 2026 report, written for Steady Compounding Insider Stocks (SCIS) members, so the full piece sits behind the paywall.

The ad business had another stellar quarter. Revenue grew 26% to $64.6 billion, beating the top of Meta’s own guidance. But adjusted EPS grew just 2% and free cash flow turned negative for the first time, because the revenue is being reinvested as soon as it comes in. The next thing it’s funding is Muse, which more than 10 million people set up in its first six weeks and which earns next to nothing yet.

To pay for all this, Meta is raising $97.5 billion of new equity, about 6% of the company. Between 2017 and 2025 it spent $174 billion buying back its own shares, at an average of roughly $290 by my estimate. Now it’s selling shares at $625.

The report goes through how the raise works, why it comes with a new class of shares that carry no votes (there’s a personal reason for that too), and what I’m doing with my own position.

SCIS members: Click here to read​

Not a member yet? Click here to join​

That's all I have for you today!

Till next week.

Thomas Author of the #1 Straits Times bestseller The Lunch Break Investor​

​