This week I went through onboarding at my new position. Most of it was training videos — senior leaders speaking to a camera, knowing the audience would be people they’d never meet and couldn’t read. They modeled real humility anyway.

I want to tell you that wasn’t surprising. It was.

The cost of that surprise is what I want to write about.

Public discourse in 2026 has a bundling problem.

A bundle is a stack of separate claims wrapped and defended as one package. The propositions inside don’t logically depend on each other; they’re held together socially, by the people who defend them in lockstep and the opponents who attack them in lockstep. Sign on to any one and you’re presumed to have accepted the rest. Reject any one and you’re presumed to have rejected all of them.

Take energy and climate, since I’ve been chewing on it. The bundle, roughly:

  1. Carbon emissions are causing measurable warming.
  2. Specific climate sensitivity numbers are precisely known.
  3. Specific tipping points will occur at specific temperatures.
  4. Battery-electric vehicles are the right answer for every transportation use case.
  5. Wind and solar can carry the grid without nuclear.
  6. Direct thermal pollution from data centers and cooling towers doesn’t matter.
  7. Atmospheric CO2 is permanent on human timescales.
  8. Western lifestyle changes are required regardless of what the technology does.

These are eight different claims. The evidence base for #1 is overwhelming. The evidence base for #4 collapses the moment you describe a Montana ranch or a ten-day trip into the backcountry. #6 is held by absence — the environmental establishment hasn’t measured it and the EPA’s data center framework doesn’t address it, so “doesn’t matter” becomes the default by way of nobody looking. The rest sit somewhere in between, with confidence intervals that the public framing mostly hides.

But you’re not allowed to sort them. You either accept the bundle and stay in the tribe, or you reject any one piece and get assigned to the other tribe — where a different bundle is waiting, just as welded shut.

I think this is most of what’s wrong with the conversation. Not that we disagree. That we’ve stopped being permitted to agree partially.

Once you’ve accepted a bundle, it functions like a lens. It shapes what questions count as serious and which evidence registers as real. The cost is bigger than the wrong answers it locks in: it’s the better ones it stops you from seeing.

Untangling a bundle requires a specific move: I was right about A, wrong about B, and genuinely uncertain about C.

That’s a humility move. It costs the speaker the comfort of total certainty and the simplicity of always knowing which side they’re on. It also requires having seen someone do it — at scale, in public, under pressure — at some point in your formative life.

Most people now have not.

The C-suite version of the last decade, in tech and politics both, has selected hard against humility. Dogmatism works because followers jump on cue with whoever holds power. Acknowledged error breaks the spell. The feedback loops that used to provide some corrective — boards, press, internal dissent — have softened into PR. The leaders who would have modeled “I was wrong about that” mostly got filtered out of the seats where anyone would have observed them.

Here’s the part that locks the bundle in place: most people are incapable of expressing that which they have not observed. The muscle is largely learned by witnessing it in someone with authority over you. If you’ve never seen it, you don’t develop it. You may not even recognize it as a thing that exists.

So the bundle stays welded. The propositions don’t belong together; they stay together because nobody in the public eye has the humility infrastructure required to take them apart.

The cost of that, in my own life, has been a slow recalibration of what I expect from rooms full of leaders. The expectation has been low for a while.

Which is why this week mattered more than it should have. A 50-year-old company headquartered in middle America, selecting for the kind of grounded leadership that a culture optimized for performance would have filtered out years ago. Senior people saying, on camera, to people they’d never meet, the things you only say after you’ve thought about your job long enough to be uncertain about parts of it.

That’s not proof that the broader culture is recovering. It’s proof that pockets remain.

Pockets are how this gets rebuilt, if it gets rebuilt. Institutions can’t restore what they’ve collectively forgotten how to do. People who remember have to keep doing it visibly — in onboarding rooms, in team meetings, in essays, in the conversations they have with their kids — and trust that the example outlasts the cycle that punished it.

Two things follow. They’re the same thing said twice.

Refuse the bundle. When someone hands you one and tells you which tribe you join by accepting it, take it apart. Hold the propositions that are well-supported. Be honest about the ones that aren’t. Decline the package deal even when accepting it whole would be easier.

Model the humility wherever you have authority. Your kids see you do it. Your team sees you do it. The room you’re in for the next forty minutes sees you do it. As far as I can tell, that’s the only mechanism that preserves the muscle while the public culture starves it.

Neither of these is satisfying at the scale of the problem. The problem is much bigger than anything one of us can model in a Tuesday meeting.

But the scale of the problem is exactly the kind of thing where individual modeling is the only thing that actually compounds. The lens you model becomes the lens they look through.

You can’t give people the capacity to express what they’ve never observed.

You can be one of the people they observe.