Cognitive Crush, Part 2: The Decisions You Don’t Know You’re Not Making

How AI moves the bottleneck ... and makes everyone stupider (if you're not careful)

Last month I wrote about “Cognitive Crush” - the cumulative weight we clever folk all feel when we use AI to think faster.

The response surprised me. Not because people disagreed, but because of how many said “that’s exactly what’s happening to us right now.”

Lot’s of people.

So let me make it worse.

Because Cognitive Crush doesn’t just exhaust clever people.

It also degrades the quality of our decisions - silently, invisibly, in a way that nobody notices.

Two Brains are better than one

To explain why, I need to borrow from Daniel Kahneman, who you probably know from his book “Thinking, Fast and Slow.”

(And, if you’re anything like me, you love the book but haven’t actually finished reading it.)

Kahneman describes how your brain (and mine!) runs two very different thinking systems.

System 1 is fast, automatic, and always on. It’s the part that reads emotions on a face, completes the phrase “bread and ___,” and gives you a gut feeling about whether a project is on track. It operates without effort. You don’t choose to use it. It just runs.

System 2 is slow, deliberate, and effortful. It’s where you do the hard thinking - weighing trade-offs, questioning assumptions, doing the maths on whether a plan actually adds up. System 2 feels like the “real you.” The careful, rational one.

Here’s the first problem: System 2 is so, so, so lazy.

That’s not an insult - it’s a design feature. Hard thinking burns through real, measurable energy, and our brains have evolved to conserve energy. Kahneman’s research showed your pupils physically dilate when System 2 is working hard. Your body literally strains under the effort. It uses so much energy and it makes you tired. So your brain rations it. System 2 only kicks in when it absolutely has to.

The rest of the time? System 1 runs the show. And much of the time System 2 - the part that’s supposed to check the work - just rubber-stamps whatever System 1 delivers.

This works fine most of the time. System 1 is fast and usually good enough.

Usually... eek!

Here’s the second problem: System 1 is sneaky, confident and makes stuff up.

When System 1 hits a hard question, it doesn’t stop and wait for System 2 to show up. It does something sneaky. It swaps the hard question for an easier one - and answers that instead. Without telling you.

Kahneman calls this “question substitution.” You won’t notice it happening. You’ll just feel like you know the answer, and that the answer is good.

“Is this feature going to deliver the outcome we promised?” is a hard question. “Do I trust the person who proposed it?” is an easy one. System 1 will quietly swap the second for the first and hand you a confident answer that has nothing to do with the actual question.

“Should we cut this from the roadmap?” is hard. “Will cutting it create an uncomfortable conversation that I’d rather avoid?” is easy. Guess which one your brain actually answers.

And here’s the really unsettling part.

System 1 doesn’t just give you a guess. It gives you a feeling of knowing. You walk away from these substituted answers with genuine confidence - because they feel right. You don’t feel like you cut corners. You feel like you made a good decision.

(I’m sure you know this feeling - the same sort of thing happens when a confident-but-thick person tells you something and you believe them because they seem confident.)

Now let me see if I can put this together with Cognitive Crush.

AI tools have just replaced a huge chunk of your leadership team’s routine work with concentrated decision-making.

The easy stuff (that used to be hard and time consuming) is now largely automated.

What’s left?

The hard stuff - the trade-offs, the priorities, the judgment calls.

That’s all System 2 work. And System 2 has a fixed, limited budget.

And you’ve got too much thinking to do ... and the System 2 budget has run out.

Arrhghghghghghgh …

So what happens when that budget runs out?

System 1 takes over.

Not dramatically - nobody suddenly becomes stupid. It’s subtle. Leaders start going with their gut on things that need analysis. They approve plans that feel right without checking the numbers. They substitute easy questions for hard ones - “Does this deck look polished?” instead of “Does this strategy actually hold together?” - and they don’t notice the swap.

(System 2 is supposed to notice, but it’s having a nap remember 💤💤💤)

Cognitive Crush doesn’t make people visibly bad at their jobs.

It makes them invisibly worse.

They still feel confident.
They still make decisions quickly.
From the outside, everything looks fine.

But the thinking gets shallower. The hard questions - the ones that actually matter - quietly stop getting asked. And everyone’s brain is too busy to notice.

It spreads like a virus …

Because when System 2 is depleted across your whole leadership team, you don’t get one bad decision. You get a slow, organisation-wide drift toward shallow thinking. Every leader, every day, burning through their cognitive budget faster than they can replenish it - and backfilling with gut instinct dressed up as judgment.

And nobody raises the alarm, because everyone feels fine.

System 1 is very good at feeling confident.

System 2 is too busy to notice.

---

One of the joys of my life is that I get to work with a lot of very clever people. And if I had to sum up the problem in one sentence, it would be this: they don’t have enough time to think slowly, and carefully.

Gut thinking - System 1 - is absolutely brilliant. It’s fast, it’s experienced, and most of the time it’s right.

But if you don’t complement it with System 2 - the slow, careful, effortful thinking - you end up doing stupid things that look brilliant.

And that’s the most dangerous kind of stupid, because nobody catches it.

---

So what can we do about it?

We need to start treating cognitive energy like the precious, finite resource it is. You treat it like the bottleneck. Because that’s what it is.

If you’ve read my books, you’ll recognise the FOCCCUS Formula here. We’ve done the hard part already - we’ve found the bottleneck. It’s the cognitive capacity of the people making your most consequential decisions.

Now you need to protect it - which is how we optimise it (the O in FOCCCUS).

That starts with rest. Not as a wellness initiative - as an operational necessity. System 2 replenishes when it gets downtime. If your leaders are sprinting from decision to decision all day, you’re running your most valuable resource at 100% utilisation. And we know how that ends.

Then you slow the flow. Space out the decisions. Protect thinking time in the diary. Stop treating a packed calendar as a sign of importance and start treating it as a sign that your bottleneck is overloaded.

Then you offload. Not everything requires the leader’s brain. Delegate the decisions that don’t need System 2 - and be honest about which ones those are. You can delegate to colleagues, consultants, and - yes - AI. Most leaders hold onto far more than they need to.

And then - most importantly - you curate.

Be ruthless about which decisions actually deserve your best thinking and which ones don’t warrant a minute of that precious capacity. Right now, in most organisations, leaders are burning their best thinking on decisions that don’t matter - while the ones that actually matter get whatever’s left over. Which, increasingly, is System 1 on autopilot.

The bottom line …

You can’t solve a brain bottleneck by feeding it faster. You solve it by making space for the hard thinking to actually happen. On the decisions where it matters most.

TLDR: Protect your brainy bottleneck!

I really hope that helps.

Clarke - who loves it when you hit reply and so hello or ask a question!

p.s. I wrapped up with a client a couple of weeks ago. They make real stuff and we got a 23% increase in revenue, over 3 months, and it all came from a bunch of well spaced dedicated thinking time. So, what I’ve written above works … and I have room to take on one more client. Hit reply if you’d like to chat.