The Halo Effect

In 1920, an American psychologist named Edward Thorndike was given a tedious assignment by the Army. He was asked to study how commanding officers rated the men under them. The ratings came back across a list of unrelated traits: intelligence, leadership, physical bearing, loyalty, discipline. Thorndike expected the numbers to vary. They didn’t. An officer judged tall and handsome was also judged intelligent. An officer judged sloppy was also judged disloyal. The brain, Thorndike concluded, was not evaluating traits at all. It was forming a single impression and laundering it through every category on the page.

The video is here:

Watch it before the next product announcement. You will see them differently.

Thorndike called the mechanism the Halo Effect. A century later it’s still the most reliable cognitive shortcut in the literature, and it’s the engine quietly driving the largest capital allocation event in modern financial history.

Consider what we’re being asked to believe. A man worth two hundred billion dollars walks onto a stage and announces that he is building a god. He provides no benchmarks an outside party can verify. His company’s internal data is sealed. The technology he’s selling loses more money the more people use it. And the response from the financial press, from sovereign wealth funds, from pension managers in Sacramento and Oslo, isn’t skepticism. It’s deference. He must know something the rest of us do not. Right? He is, after all, worth two hundred billion dollars.

What the Swedes Figured Out

The largest study ever conducted on the relationship between extreme wealth and cognitive ability was published in Sweden in 2023. It examined the top earners in the country across decades of tax and military testing data. Its finding was unambiguous. Above a certain threshold, additional wealth correlates with capital access, timing, and inherited position far more strongly than with intelligence. The study made almost no impression on the public conversation. The Halo Effect doesn’t run on data. It runs on impression, and the impression is being printed next to the man’s name in every headline.

Belief as Business Model

What Thorndike couldn’t have anticipated is how the bias would eventually become load-bearing. A growth-stage technology company trades at a price-to-earnings multiple many times its actual earnings. The inflated stock isn’t decoration. It’s the company’s working currency, used to acquire competitors, retain executives, and finance operations that cash flow alone can’t support.

The multiple holds only as long as the growth narrative holds. The narrative holds only as long as the founder is perceived as a visionary. Remove the visionary and the multiple collapses. Remove the multiple and the company can no longer function as a going concern. The Halo isn’t a marketing layer on top of the business. In several of the largest companies on the American exchanges, it is the business.

This is the part that’s rarely said aloud, even by people who understand it. The financial architecture of the present technology sector requires a savior figure for structural reasons. Even an investor who’s read Thorndike, who can recite the Swedish data line by line, has every incentive to keep buying. The collapse of the savior narrative is the collapse of the position. Skepticism is professionally expensive. Faith is professionally rational.

Bizarro ‘Optimism’

And then there’s the doctrine. Marc Andreessen’s Techno-Optimist Manifesto is not a financial document. It is a catechism. It declares technologists “Technical Supermen.” It denounces dissenters as Luddites and communists. It argues, in plain text, that slowing the development of artificial intelligence is “a form of murder.” Every ruling class in history has produced a theology that places its own interests at the apex of cosmic significance and frames objection as a moral failing. Andreessen has simply written ours.

In the full episode I walk through what I am calling the Thorndike Test, a simple diagnostic any viewer can run on any claim made by any tech executive, in any keynote, on any earnings call. It takes about ten seconds. It is one of the cheat codes the series exists to surface.

A note from Julian. The Silicon Mirage is produced without sponsors, without ad reads, and without the kind of access that has to be paid for in coverage. If you find the work useful, the paid tier of this newsletter is what keeps it independent. Think of it the way you think of the pledge drive: ten dollars a month, no tote bag, no compromise.

For the comments. The Thorndike Test is a tool for the individual mind. It strips the halo off a single claim. But the Halo Effect does not only distort what we believe. It concentrates power. Nation-state-scale wealth and influence end up lodged in a small number of private hands, shielded by the same cognitive shortcut we just examined. So I want to open the floor. Not to dunk. Not to cheerlead. To think together.

One. In a functioning democracy, how much of a structural problem is it that individuals can accumulate nation-state-scale wealth and influence? Bug, feature, or something we have stopped noticing?

Two. Should billionaires exist? If your answer is “yes, but,” what is on the other side of the “but”?

Three. The Thorndike Test diagnoses one person, one claim, one moment. What would a collective version look like? At the level of the individual, the neighborhood, the ballot, what might actually loosen the Halo’s grip on the commons?

I will be reading. Disagreement welcome. Certainty, less so.