🎬 Every week I share a short (fun!) product video teardown and some hot takes on behavior change. Pro tip: The video only takes 〜2 minutes at 2x speed. And it will be more memorable. ;) ⏩
You’ve probably heard the trolley problem: A train is barreling toward five people on the tracks. There’s a switch. If you flip it, you divert the train and save the five people, but you kill one. Do you flip it?
There’s a variation I find more interesting, and it comes from David Pizarro, a moral philosopher at Cornell. This version is called “Should You Kill the Fat Man.” You’re standing on a bridge next to a large man. If you push him over the side, his body stops the trolley and five people live. He dies. Do you push?
Pizarro isn’t really interested in your answer. What he wants to study is how you and I judge the people who answer differently.
The person who agonizes earns more trust
We tend to judge consequentialists more harshly. Someone who says “I ran the calculation; pushing the man saves more lives, so I pushed” reads as cold and less trustworthy. We’re more favorable toward someone who says “I have morals, I will never take a life,” even when that choice costs five other lives. The person with the bright-line rule reads as more trustworthy, even though they let more people die.
But Pizarro takes it further: the way you push matters, too. Someone who acts quickly and throws their head back in laughter reads differently than someone who deliberates, agonizes, maybe sobs, has no idea what to do, and then finally pushes. We trust the second person more because the visible struggle communicates something about intent that the quick push doesn’t.
Why we see Anthropic as the nicer company
One theory for why Anthropic registers as more trustworthy than OpenAI, despite both having roughly the same models, is that Anthropic exposes how hard the problem is. They think out loud. They agonize publicly. They published a study on the effects of AI and what it might mean for humanity, essentially signaling: we are with you, we are thinking deeply about this. OpenAI doesn’t project the same deliberation. The impression many people have is that they move fast and confidently without broadcasting the weight of what they’re building.
Going back to Pizarro, it’s quite reasonable that we’re judging both companies on the same action but perceiving them differently based on how much deliberation each appears to show. Anthropic looks like the person who thinks more deeply, shows their inner conflict, then finally makes the decision to push. OpenAI looks like the one who pushed and laughed. The models are roughly the same, but our perception of each company is not.
This works on your pricing page and your spouse
If you’re about to do something that makes you nervous, whether it’s a hard conversation with a partner or a difficult call with your team, show that you’re nervous. Be transparent about your thinking and how you worked through it. The worst thing you can do is deliver the decision quickly and let people conclude you didn’t think about it.
If you’re a company raising prices, tell your customers that you deliberated, that you considered how this would affect them, and that it wasn’t the easy choice. Without that context, people fill in their own explanation, and that explanation is usually “profit-hungry company.”
And of course you should know, I re-recorded this substack video multiple times. I thought for a long time about how to tell the story in a way that would actually resonate. I re-read the research and tried to figure out how to simplify the message to pull out the interesting bits. :)
If you’re working on a trust problem in your product or company, I’d love to hear about it: kristen@irrationallabs.com.
🎬 The post above is just a preview. Only takes 〜2 minutes at 2x speed. ⏩
Have a friend who would enjoy these teardowns? Click the button below to refer them (& earn some great rewards).👇
