Daniel Kahneman and the feeling of being caught
I like a good reframe. Don’t you?
It used to be that saying “I don’t know” felt like honesty. Why does it now feel like it means something else?
Why do follow up questions feel like predetermined resentments? (a phrase I used on my first date with my wife of now 13 years about having expectations)
I’m Nate Sowder, and this is unquoted, installment 17. Knowing something and understanding it are two different things. Today, we’re looking at Daniel Kahneman.
Daniel Kahneman
Kahneman is known well for his work around cognitive biases and decision-making errors. A lot of recent attention comes from Thinking, Fast and Slow, published in 2011, around the idea that people aren’t nearly as rational as they believe themselves to be.
Go figure, right?
While that’s true, what’s more interesting here is the work he did around how a person gets to an answer… and then whether they can explain it.
The idea that Kahneman was looking at this dilemma before mass social media and well before AI only makes it more relevant now. We’re going to look at how this plays out, and then push it a bit.
But first…
Let me ask you a question
If I ask you a question, you’ll almost certainly have an answer… almost immediately. It will sound coherent. To you, it will sound right, and most of the time, it will be good enough to keep the conversation moving.
How do I know this? I’ve tried it a bunch over the last couple weeks, about some really bizarre topics, and no one said “I don’t know” to the first question. Can you believe that?
NOTE: Someone with no children gave me instant guidance on how to put an infant to sleep… without hesitating.
It’s interesting how good we are at answers. What’s less obvious is what happens if you ask that same person to explain how they got there, or why they believe it. (This is where I’m going to ask you to do a lot of visualizing and relating. It wont be hard.) If you visualize that happening, things start to slow down. Not always, but often enough that it becomes obvious.
The answer shows up quickly, but the explanation takes more work.
Kahneman described this as two different kinds of thinking working together. One that moves fast and produces answers without much effort, and another that has to slow things down, check what’s been said, and figure out whether it makes sense.
Most of the time, this feels like thinking. But it isn’t the same thing.
Systems 1 & 2
The simplest version is this:
Here is why that’s so interesting.
We like to imagine we think our way toward conclusions. Kahneman is basically saying: That’s not usually the case. Most of the time we reach a conclusion, and reasoning how we came to that conclusion comes after the fact… if at all.
Most of the time, our minds do not begin by thinking carefully. Instead, they begin by producing something that feel like an answer. We get this as a judgment, impression… sometimes a guess.
Let’s dive into these:
System 1 is automatic. It’s always on, noticing patterns, completing sentences, reading tone, recognizing faces, making snap judgments, filling in gaps, and constantly trying to make the world feel… normal. It’s efficient as hell… because it has to be. You can’t reason from scratch every time you cross a street, read a room, or hear a sentence. You’d drive yourself crazy.
System 1 is what lets you hear, “2 + 2 = ?” and instantly know the answer. It’s what lets you read a headline and immediately gauge whether it sounds plausible, annoying, stupid, or true. It is also what lets you answer a question confidently (this is how you put a baby to sleep… it works every time because, Instagram) before you’ve really thought about whether you know what you’re talking about.
That last part.
Tldr: System 1 produces answers, then looks at whether that answer is plausible and coherent enough to keep going.
System 2 is different. System 2 is what happens when someone says, “Okay, why does that work?” Notice the mindset shift away from giving a statement to having to explain how you got there, having to answer the next question, work through skepticism, and then decide if that’s still what you think.
Tldr: So if System 1 is the part that says, “I’ve got it,” System 2 is the part that has to prove it.
I think what is really interesting is that Kahneman’s work proves that these two systems are not equal partners. System 1 drives a lot more of life than we want to admit. Most of the time, System 2 just rationalizes what System 1 already came up with.
This is disturbing, because most of us would assume System 2 is the boss, but a lot of the time, it’s more like legal counsel doing damage control.
Stay with me, because this is why follow-up questions are so revealing.
A first answer can come entirely from System 1. Fast, coherent… fine. But a follow-up question often forces this internal handoff. Now the person has to do more than produce an answer. Uh-oh.
This is when “pausing” is not the same thing as “thinking.”
So what?
Low stakes, you know how this goes in interviews. Candidates are nailing the first questions with clear answers and confident delivery. It all sounds great until the follow-up questions. I’m not talking about trick questions, just anything slightly off-script.
I’ve had group conversations recently with hiring managers from top enterprises, and resoundingly, their feedback to each other was, “Wow, you too?” The comment being, “I’ll get a great answer from a candidate, but when I ask a follow-up, nearly every time, the candidate has to go back to AI to look something up.”
That illustrates problems… but it’s not the biggest problem. We have no problem scrutinizing job candidates. That’s the job, right?
There’s another group leaning just as heavily on AI that’s not used to that kind of pressure. As long as the narrative is slick and the presentation is tight, heads nod, and things tend to move forward.
But this is the group setting strategic direction. These folks are making decisions that affect entire teams. Product changes, priorities… layoffs.
This is where organizations should be relying on experience and sound judgment. Being able to depend on people who can think something through, stand behind it, and then understand consequence.
(Read closely, because this is also why personal use of AI across enterprises hasn’t gone well)
…You’ve read everything up to now and you know the punchline. Leaders are using AI, but this group is over-indexed on their organizations using AI as a task engine. Faster decks, strategies, summaries and answers. This is all System 1 thinking.
What it’s missing is context and intent — not intelligence. We’re talking about reasoning. Without the ‘why’, AI is going to produce the most probable result which will end up giving you visible mediocrity.
I’m sure it makes for a great presentation, but when that first question comes in from the Board that’s something like, “Wouldn’t it be better if…?”
“Umm, no further questions… please.”
No further questions… please. was originally published in UX Collective on Medium, where people are continuing the conversation by highlighting and responding to this story.
