Don't Eat the Menu

Once, I was almost killed while walking in downtown Oakland. I'd waited for the light to change so I could cross the street. After the crossing light came on, I started to walk. Just then, a car sped through the intersection, missing me by inches.

I'd done everything “right”: I was paying attention (i.e., not looking at my phone), using the crosswalk, and had waited until the light said it was OK for me to go… and I still almost got hit. What happened?

I was “eating the menu,” a phrase I picked up from Antony De Mello and J. Francis Stroud's book Awareness. It appears in the context of a rhetorical question:

How many people spend their lives not eating food but eating the menu? A menu is only an indication of something that's available. You want to eat the steak, not the words.

Here's how I read this: there's a difference between information about a thing and the thing itself. Information helps us understand the thing and make decisions about it — but it's not the thing. You can't “eat” it.

Often, information about the thing maps closely to the thing — i.e., the menu offers a good representation of the food you'll get. But sometimes, there's a misalignment between information and the thing itself.

For example, the restaurant may have run out of steak that evening and not had time or resources to reprint the menu or the driver may have been impaired, leading him to ignore the red light.

In either case, information about the thing (the menu, the crossing light) misrepresents the state of the world. There's no steak, no matter what the menu says. It's unsafe to cross, no matter what the light says.

The accuracy of information — i.e., the degree to which it represents real-world conditions — depends on it being tightly coupled to real-world conditions. The chef knows if there's steak available that evening and can correct the menu (or not.)

On the other hand, the crossing light doesn't “know” what's happening on this particular street. The light is merely a timed device attempting to regulate traffic. Compared to the menu, the light is uncoupled from the situation it aims to elucidate.

Knowing whether information about a thing is tightly coupled to the thing's actual conditions helps us act more skillfully. For example, if we know the crossing light is “dumb,” we'll look for other data points before risking our lives.

As children, we're taught to look both ways before crossing. But we don't extend that cautious mindset to other domains. Our information-driven culture offers compelling menus — some of which are mostly decoupled from the conditions they describe.

Eat the steak, not the words.

Worth Your Attention

• Mindful set-making. Thoughts on crafting contexts through language, sparked by a question from Daniel Kahneman.

• The Undoing Project. Speaking of, here are my notes on Michael Lewis's biography of Kahneman's friendship with Amos Tversky. Worth reading if you're interested in learning about human behavior. (Pro tip: you should be.)

• The state of AI-driven design. TL;DR: interesting tools but not very smart; more like a high-level programming language than an artificial designer.
• Einstein Designer. How SalesForce is using AI to “make it efficient (cheap) to deliver high-quality (good) products at an industry-leading pace (fast).” Intriguing. (H/t Madonnalisa Chan)
• Fear in design strategy. How to overcome obstacles to designing more strategically. (H/t Laura Yarrow)
• The Metaverse Primer. A guide to understanding an important tech development. (H/t Kevin Kelly)
• Mental models and conceptual models. A high-level overview, with pointers on how to build strong conceptual models.
• A timeline of Google's messaging products. “What a decade and a half of confusing services can tell us about the future.”
• Search Atlas, a tool for seeing and crossing “invisible borders — geographic, linguistic, cultural, political — that circumscribe the information each user sees.”
• This is why we can't have nice things.

The Informed Life with… You!

We're trying something different with Episode 67 of The Informed Life podcast: instead of interviewing a guest, I'm answering questions sent by listers. Specifically, I'm addressing questions sent by Vinish Garg, José Gutierrez, and Elijah Claude. All are about information architecture.

I loved hearing from listeners, and would like to do another Q&A show. It will depend on two things: 1) whether you found episode 67 valuable, and 2) whether I get more questions from listeners. Please let me know if you liked the show or if you have a question you'd like me to answer.

The Informed Life episode 67: Listener questions

Parting Thought

In situations of unpredictability, organizations need to improvise. And to do that, the players on the field need to understand the broader context. At the team level, this is self-evident. But at the broader institutional level, it is more difficult to engineer structures that are both coherent and improvisatory.

— Stanley McChrystal, Tantum Collins, David Silverman, and Chris Fussell, Team of Teams