Hello, this is Eduardo and I’m an Engineering Leader and Certified Running coach. Leadership Endurance is back after a long pause: a newsletter about leadership, endurance sports, performance and growth, in the confluence of these worlds.
Working in tech in 2026 requires a truly impressive amount of cognitive dissonance.
Here’s the setup. Every morning the news says some version of: AI is moving too fast, the people building it are asking to be regulated and getting shrugged at, trillions of dollars are chasing a return nobody can articulate, and maybe this whole thing destabilizes society. Then you open Slack and there’s a message from your VP about how excited everyone should be about the new agent rollout. Exclamation point.
You’re expected to hold both of those in your head, every day, and act like only one of them is true.
Compare that to the pandemic. It was awful, but at least it was honest. You could say “this is weird and scary and I’m worried about my kids” in a team meeting and nobody put it in your review. There was a shared reality. (There was also a lot of sourdough, but I digress.)
Tech right now is the opposite. Express hesitation and you’re “resistant to change”. Raise an ethical objection and you’re “not innovative enough”. So people learn to shut up and perform enthusiasm instead.
Look: this isn’t a sustainable way to run an organization. You cannot ask people to compartmentalize every doubt and every ethical objection they have, indefinitely, and then act surprised when they burn out.
I’m not writing this from the sidelines
Full disclosure: in my last corporate role I joined an engineering organization that was AI-averse and left one that had embraced it. Right now I’m building a product where agents write most of the code. I like this stuff. A lot of the excitement is earned.
Which is exactly why the dissonance bugs me. Excitement and doubt aren’t binary! The people I trust most in this industry hold both at once. The people I trust least hold exactly one.
Where the dissonance actually comes from
Definitions first: cognitive dissonance is the discomfort of holding two beliefs that don’t fit, or of acting against something you believe. The well-known finding is that people resolve it by bending the belief, not the behavior. So when your paycheck depends on being enthusiastic, guess which one bends. The doubt doesn’t go away. It goes underground, where it’s useless to everyone.
That’s the mechanism behind a lot of the burnout I’m seeing. Sure, the workload is real. But the thing that actually grinds people down is the daily effort of performing a certainty they don’t feel. Anyone who’s run a marathon knows this cost: holding someone else’s pace is far more expensive than covering the same distance at your own. You can fake it for a mile. You cannot fake it for 26.
What the pandemic got right
Three things from 2020 worth stealing back:
- Leaders said the true thing first. “This is hard, I don’t know how long it lasts, here’s what we know today.” Naming reality didn’t cause panic. It made the panic manageable.
- Fear was treated as information, not insubordination. When an engineer said “I can’t focus with my kids at home”, the response was to change the plan, not to question their commitment.
- There was a shared timeline. Everyone knew we were in the middle of something with a before and an after. That’s what makes endurance possible. An emergency with no end date isn’t an emergency, it’s a lifestyle, and nobody sustains a lifestyle of emergency.
None of this requires a pandemic. It requires a leader willing to go first.
Okay, so what do you actually do?
If you manage people in tech today, here’s what I’d do. (Here’s what I try to do, anyway. Results vary.)
Say the true thing first. Something like: “I’m excited about what this lets us build, and I don’t know where it ends. Both are true, and both are allowed in this room.” Try it once. Watch how much air comes back into the team.
Separate adoption from enthusiasm. You can require that people learn the tools. You cannot require that they love them. So make the actual expectation explicit: use it, evaluate it honestly, tell me what breaks. The engineer who reports that the agent produced garbage on the third attempt isn’t resisting change. They’re doing precisely the job you need done, and that report is worth more than the demo that went well.
Make doubt a deliverable. Put it in the retro. Ask “what worried you this sprint?” right next to “what shipped?”. Amy Edmondson’s psychological safety research is pushing thirty years old and the finding hasn’t budged: teams that can voice concerns catch problems earlier and fix them cheaper. The concern you never hear about is the expensive one.
Draw the ethical line where you actually stand. Your people are wondering where it is. If there are things your team won’t build, data you won’t feed into a model, or decisions you won’t hand to an automated system, say so. A boundary people can name is one they can rest against. A boundary they have to guess at is one more thing to carry.
Pace the org like an athlete. Not every week is race week. A training plan that’s all intensity and no recovery doesn’t produce a faster runner, it produces an injured one. If every quarter is “the most important quarter in the company’s history”, congratulations: you’ve removed the recovery block from the plan. The injury is coming.
The endurance frame
The athletes who last aren’t the ones who feel no pain; they’re the ones who can feel it, name it, and keep making good decisions anyway. That’s the skill. Compartmentalizing isn’t endurance. It’s debt, and it compounds.
The industry is going to keep moving fast, and a lot of what’s being built will turn out to be worth it. But the people building it are carrying more than they’re allowed to admit, and the gap between what they feel and what they’re permitted to say is exactly where burnout lives.
You’re allowed to say it’s weird. If you’re the leader, you’re obligated to say it first.