Anthropic's 5% Public Trust Has a Precedent: Paul Newman's Salad Dressing

Anthropic will dedicate 5% of its shares to a company-owned public benefit trust, the company announced last week. If its valuation holds, those shares could be worth billions on day one — and hundreds of billions over time. The AI lab, widely rumored to be approaching a record-setting IPO, has effectively decided to cut the public in on a piece of whatever its technology produces next.

The scale is unprecedented. Run the same math on an older giant and the numbers get loud: had Google set aside 5% of itself before its 2004 IPO, those shares would be worth $184 billion today.

The idea itself, however, is not new. Anthropic was beaten to the punch by a movie star turned salad dressing tycoon. And the comparison exposes risks the AI version carries that the original never did.

What did Paul Newman do first?

By age 57, Paul Newman had starred in almost 50 films, including Cool Hand Luke and Butch Cassidy and the Sundance Kid. He had also taken second place at the 24 Hours of Le Mans auto race. When someone suggested he put his face on bottles of salad dressing, Newman reportedly thought the idea felt like "exploitation."

Then he reframed it. Hollywood had always exploited his looks and boyish, all-American charm. Why not do it for a good cause? In Newman's words: "Why not, really, go to the fullest length, and the silliest length, in exploiting yourself, and turn the proceeds back to the community?"

So he did. Approaching 60, he launched Newman's Own, a salad dressing company that donates 100% of its proceeds to charity. The company grew into a full line of grocery products — its salsa, as any Whole Foods shopper can attest, remains respectable. Since its founding in 1982, Newman's Own has generated over $600 million in charitable donations, according to AP reporting, a massive sum for a company the actor started "on a bit of a lark."

Where does effective altruism fit in?

Newman called his approach "shameless exploitation in pursuit of the common good." Today the same philosophy carries a formal name: effective altruism. As The New York Times explored earlier this year, the mindset is pervasive across frontier AI, and Anthropic sits at its center.

The logic is straightforward. Amass as much wealth and power as possible, then give it away or direct the influence toward humanity's benefit. It sounds nice on paper.

The philosophy has a documented dark side. Prosecutors and commentators argued it drove "crypto king" Sam Bankman-Fried to defraud his way to riches, the better to fund causes of his choosing.

Why is Anthropic's version riskier?

Here the analogy breaks. Even if Anthropic gave away all its shares à la Newman, the drive to maximize their value for humanity could push the company to scrimp on safety — or to build, by its own admission, things that risk destroying the world.

Observers will read the trust through different lenses. Critics may see it as an attempt to shake regulators off the company's back, or to quiet public anger before it reaches the company's SoMA office. Supporters will see a genuine effort to give back.

The author's warning is blunt: as the humans stuck holding the bag if AI goes horribly wrong, the public should not let the gesture obscure the bigger questions of AI safety, creative rights, and societal impacts that still surround Anthropic's work.

Altruism, the argument runs, is not just about money, and positive ends do not always justify the means. There is not much point in setting aside cash to help humans if your technology has inadvertently stamped them out. For investors sizing up a record-setting IPO, that trade-off — generous trust on one side, unresolved risk on the other — is now part of the offering.