Why I Write About AI Chip Shortages and Lead Paint in the Same Newsletter

I get some version of this question often enough that it deserves a real answer instead of a shrug. One week I'm writing about a high-bandwidth memory shortage in AI chips. The next it's a 7-8% stop-loss rule from a stock trader's book published decades ago. The week after that it's lead paint in a Glendale duplex. On the surface, those look like three unrelated newsletters written by three different people. They're not. They're the same idea, applied three times.

The idea is: find the actual constraint, not the one everyone's arguing about

Every one of these topics has a loud, obvious story sitting on top of a quieter, more specific one underneath, and the quieter one is almost always the one that actually matters.

The AI chip shortage headline says Nvidia can't make enough GPUs. The real constraint, at least right now, is a three-company chokepoint in high-bandwidth memory that most coverage never mentions. The AI-adoption headline says the technology isn't good enough yet. The real constraint is usually psychological, algorithm aversion, habit, and trust, not the model's capability. A stock thesis says the company is undervalued. The real constraint on your returns is usually your own reluctance to admit you were wrong and sell at 8% instead of 40%. A property's inspection report says the building is sound. The real constraint might be a rent ordinance you didn't check, or lead paint a standard inspection was never built to catch.

In every case, the obvious story is the one that's easy to write a headline about. The real constraint is boring, specific, and usually sitting one layer beneath it.

What that looks like across the actual essays

William O'Neil's 7-8% rule is a discipline for overriding your own instinct at exactly the moment your instinct is least trustworthy. Ray Dalio's diversification research is the reason I'm not depending on any one position, or any one asset class, to be right in the first place, five uncorrelated pieces instead of one concentrated bet on being smart. Those two ideas sound like they belong to different investors. They're actually the inside view and the outside view of the same problem: control what a single mistake can cost you, and don't let any single bet be large enough to sink the whole structure.

The AI writing follows the identical instinct, aimed at a market instead of a portfolio. Everyone's attention goes to the loudest, most crowded part of the AI story, the chatbot, the model, the capex number. The more interesting position is almost always the unglamorous constraint everyone else skipped past, the memory chip, the adoption psychology, the physical bottleneck. That's not a different skill from stock discipline. It's the same skill, pointed at an industry instead of a stock chart.

And the lead-and-asbestos essay is the same idea in its most literal form. I run a firm built around the principle that you should only own things you can actually inspect. That property was the one time I didn't inspect deeply enough, and the standard checklist looked clean right up until it wasn't. The rent-ordinance essay is the same lesson filed under a different heading, real estate has a regulatory layer most buyers never check until it's already too late to negotiate around it.

Why it's all one thing

Glencore Associates runs five asset classes because no single one of them has to be right about the next two years. AIDiscover exists because most companies trying to adopt AI are fighting the wrong battle, treating it as a technology problem when it's usually an adoption-psychology problem. The newsletter covers whatever this week's version of that same question turns out to be, a chip market, a stock rule, a rental property, because the underlying discipline doesn't change even though the subject matter does: assume the obvious story is incomplete, and go find the actual constraint hiding underneath it.

That's the whole answer to why this looks scattered from the outside. It isn't. It's one habit of mind, applied to a portfolio, an industry, and a property, because that habit doesn't actually care which of the three it's pointed at.

I'm David Bellairian, founder and CEO of Glencore Associates, a Los Angeles private investment and asset management firm, and founder of AIDiscover, an AI-powered marketing and brand discovery agency. This reflects my own way of thinking about risk and constraint, not investment, legal, or professional advice.