Resilient people are made, not born. There isn't a gene or a personality trait for resiliency. It's a skill you can practice and strengthen, just as you would a muscle.
That's the Mayo Clinic. Word for word. And it's everywhere.
It is. And Hannah sent us something this week that pokes at exactly that. She says they talk a lot about resilience with Ezra, the whole parenting idea of the moment, let the kid hit some frustration, don't swoop in and rescue him from every difficulty, give him a chance to learn he can handle things. She hopes that's landing. But what she actually wants to talk about is resilience in adults.
Right, because Ezra's a baby. He's got decades of this ahead of him.
Once we're already fully formed people, how flexible is it? Is it a muscle that can be strengthened, or is it mostly personality and circumstance? And then her second thread, which I think is the sharper one. How much of resilience is really just available bandwidth? When life is going well, a stressful situation is manageable. When you're maxed out, no reserve left, suddenly every small thing sends you into a tailspin.
And then the practical half.
What do you do at the limit? Can you build capacity in the middle of a stressful period, or is the answer sometimes just that you have to reduce the load before resilience can come back? And how do you actually claw your way out once you're already in the hole? Because waiting for life to get easier is not a plan.
No, it isn't. And the research here is interesting, because the muscle metaphor is not wrong so much as it's doing rhetorical work the evidence doesn't quite support.
So where do we start?
Start with what resilience actually is, because the definition has moved. The cleanest version is Troy and colleagues in the Annual Review of Psychology a few years back. They separate factors, processes, and outcomes. Resilience isn't a thing you have. It's an outcome, faring better than expected given adversity, and it sits on a continuum. There's no line where you cross over into being a resilient person.
Which already breaks the metaphor. A muscle is a thing you have. This is a result.
A result produced by processes. That's the key shift. And Bonanno's three axioms from 2024 make it concrete. First, resilience is the most common outcome after potentially traumatic events. Roughly two-thirds of people exposed to major trauma show a stable resilient trajectory. Second, there are no key traits. His line is "resilience is not a type." Third, resilient outcomes happen through flexible self-regulation.
Two-thirds. That's higher than most people would guess.
Much higher. The cultural picture is that trauma shatters people, and for a minority it does. But the modal response is stability.
I want to sit on that for a second, because it reframes the whole conversation. If two-thirds of people are fine after major trauma, then resilience isn't a rare achievement. It's the default. The exceptional case is the one-third who struggle.
And it flips the self-help framing on its head. The self-help framing treats resilience as something you have to acquire, like a rare mineral. The data says it's what most people do by default, and the interesting question is what interferes with it.
So if two-thirds bounce back and there's no trait that predicts who, you'd think the field would be good at identifying resilient people in advance.
It's terrible at it. Bonanno calls it the resilience paradox. The correlates of resilient outcomes are, in his words, generally so modest that it is not possible to accurately identify who will be resilient.
So you can't pick them out beforehand.
Not reliably. Which is a real problem for the self-help version of this. If you can't predict who bounces back from measurable characteristics, then "just build resilience" is less actionable than it sounds. It's advice pointed at something nobody can locate.
And the intervention evidence backs that up.
The most important recent paper is a Nature Reviews Psychology review from May, Schäfer, Stoffers-Winterling and Lieb. They went through the intervention literature and the conclusion is blunt. Current resilience interventions, focused on training individual psychosocial factors, yield only small-to-medium favourable effects. And they argue the field has to move from training abstract individual factors to targeting mechanisms during stressor exposure.
So the thing Mayo says you can strengthen like a muscle, the actual effect size of trying to strengthen it is small to medium.
Small to medium. That's not nothing. But it's a long way from "practice and strengthen, just as you would a muscle."
Let me make sure I understand the size of that gap. If I sign up for a resilience program, what am I actually buying?
A modest nudge. Think of it like the difference between a vitamin and an antibiotic. A vitamin is real, it does something, but if you have an infection, it's not the intervention you need. Resilience training is closer to the vitamin end of the spectrum. It helps at the margins.
And the margins are not where most people are struggling.
The margins are not where most people are struggling. Which is where Hannah's bandwidth question comes in, and I think that's where the biology actually gets interesting.
It does. There's a real framework for this, the energetic model of allostatic load. Bobba-Alves, Juster and Picard, 2022. The claim is that living organisms have a limited capacity to consume energy. Chronic stress overconsumes that energy, and it competes directly with growth, maintenance, and repair.
Say the competition part again, because that's the whole thing.
The transition to maladaptive states arises when the added energetic cost of stress competes with longevity-promoting growth, maintenance, and repair. It's not that stress makes you weak. It's that stress is spending the same budget your body uses to fix itself.
So you're not running out of willpower. You're running out of energy that was already allocated to repair.
Right. And Radley and Herman in Biological Psychiatry in 2023 put the same idea in one line. Transition to pathology occurs only after the adaptive capacity of an organism is exhausted.
"Only after." That's a strong claim.
It's a strong claim. The stress response isn't the problem. It's an adaptive defense mechanism. It becomes pathology when the capacity to sustain it runs out. Which is exactly Hannah's intuition. Being maxed out makes small things overwhelming because the reserve that would absorb a small thing is already committed.
And it gets worse with repetition, doesn't it?
Stress sensitization. Giotakos, 2020. Repeated emotional stressors leave traces that render individuals vulnerable to the onset of new episodes, even under the influence of only moderate psychosocial stress. The threshold drops. You don't need a big stressor the fifth time. A moderate one does it.
So the tailspin isn't a character flaw. It's a lowered threshold.
It's a lowered threshold with a mechanism behind it. There's also an inverted-U here. Kawamoto and colleagues, 2025. Moderate stress triggers beneficial adaptation. Chronic or excessive stress leads to allostatic load and pathology. Same stressor, opposite outcome, depending on where you are on the curve.
Which means "what doesn't kill you makes you stronger" is only true in a band.
Only in a band. Past the band it makes you sicker.
Is there a way to feel which side of the band you're on? Like a physical tell?
Sleep is the tell. If you're sleeping badly for weeks, you're past the band. If small things are landing harder than they should, you're past the band. The body usually tells you before the mind admits it.
Now, there's a piece of this I want to push on, because the "resilience as a depletable resource" idea has a famous controversy attached to it.
Ego depletion.
Because that's the laboratory version of "you run out."
And it's contested, so let's not pretend otherwise. The big multilab replication, twelve labs, one thousand seven hundred and seventy-five participants, found an effect of d equals zero point one zero. Small and significant, they called it. Other meta-analyses found the effect basically vanishes once you correct for publication bias.
So the lab version of "willpower runs out" is shaky.
It's shaky. But Baumeister and colleagues pushed back in 2024, arguing the failed replications were due to ineffective manipulations. So it's unresolved. I'd say the honest position is that the specific ego-depletion paradigm is in trouble, while the broader energetic picture from allostatic load is on much firmer ground. Different levels, different evidence.
That's a useful split. The metaphor is contested. The biology of exhaustion isn't.
That's exactly the split. And it matters for Hannah's question, because if resilience were a muscle in the straightforward sense, the answer at the limit would be "train harder." If it's an energy budget, the answer at the limit is "stop spending."
Which brings us to the fork. Can you build capacity mid-stress, or do you have to reduce the load first?
The literature leans hard toward load first. If you're in overload, the body is diverting energy away from repair. Training harder in that state is spending from the account you're trying to refill.
And the intervention trials that do show gains, what do they actually show?
Modest but real. An eight-week guided self-help mindfulness intervention raised Connor-Davidson Resilience Scale scores and lowered burnout in a hundred and eighteen psychiatric nurses. A five-week Enhanced Stress Resilience Training improved burnout and personal accomplishment in ICU nurses, forty-eight of them. A partially clustered trial of a hundred and forty-four mental health nurses improved coping self-efficacy.
All nurse populations.
All nurse populations, which is not an accident. That's a workforce with a known, measurable, chronic load. And notice the sizes. Forty-eight people. A hundred and eighteen. These are small trials with small-to-medium effects, which is exactly what the Nature Reviews review said to expect.
So the ceiling on training your way out is low.
The ceiling on training your way out is low, and the floor on reducing your load is high. That asymmetry is the finding.
Then what actually helps you claw back? Because "reduce the load" is not always available. You can't always quit.
No. And this is where the modifiable-resource studies are useful, because they don't ask you to change your personality. The breast-cancer study, Korecki and colleagues, a hundred and eighty patients. Lifetime stressor severity predicted depressive symptoms, beta of zero point three zero. But sleep quality, social support, optimism, and mastery buffered that relationship.
Buffered how?
They weakened the link between the stressors and the depression. Same stressor count, different outcome, and the difference was carried by things you can actually change. Sleep is not a personality trait. Social support is not a personality trait.
Optimism is a bit of one.
Optimism is the fuzziest of the four, fair. But mastery, the sense that your actions move outcomes, that's trainable in a way that "be more resilient" isn't.
And the cohort study?
Contador and colleagues, 2026. Six hundred and seventy-seven older adults at risk for Alzheimer's, followed through COVID. Sixty-six percent followed a resilient trajectory. The predictors were physical and leisure activity, sleep, social relationships, and emotional support, and the effect was mediated by perceived stress.
Mediated by perceived stress. So the resources work by changing how stressful the situation feels.
Which is the bandwidth story from the other direction. You can add resources to the system, or you can lower the perceived load on it. Same ledger.
And there's the physical activity piece too.
Bonanno and colleagues, 2023. Physical activity may buffer against depression and promote resilience after major life stressors. That one's interesting because it's cheap, it's available when you can't quit your job, and it works on the energy side, not the mindset side.
Now, the practitioner accounts. Because the trials are small and the effect sizes are modest, and yet people do get out.
They do. And the accounts converge in a way the trials don't. The pattern is reduce the load first, then sleep, diet, exercise, then the mental work. One person put it as: the very first thing I did was reduce my workload. Sleep, diet and exercise were next.
Load first, in that order.
Load first. Another one: time, time and more time, be patient with yourself. And my favorite, because it's the least clinical: if you can afford it, quit your job and don't do anything until it sounds good, because you're giving your mind time to run a defrag on itself.
It's the best description of recovery I've read. Not a metric. A felt threshold.
But here's the thing I keep circling back to. If the answer is "reduce the load," that's not always available, and there's a version of this discourse that uses resilience to avoid that fact.
The dark side of resilience. Mahdiani and Ungar, cited in the 2026 review. Resilience discourse can be used to justify not fixing broken systems.
Which is the sharpest counterpoint in the whole literature. Because "build your resilience" aimed at someone covering three jobs is not advice. It's a way of not hiring the third person.
And it connects straight back to the paradox. If you can't predict who's resilient, and the interventions are small-to-medium, then telling people to be more resilient is often telling them to absorb a structural problem personally.
So the honest answer to Hannah's question might be: sometimes the answer is reduce the load, and sometimes you can't, and the gap between those two is where people actually live.
That's the gap. And I'd add one more thing, which is that the muscle metaphor isn't harmless in that gap. It implies the deficit is in you. If resilience is a muscle, then not having enough is a training failure.
When it might be a staffing failure.
It might be a staffing failure.
I want to test the metaphor one more time before we move on. Is there any version of it that survives?
A weak version survives. Flexible self-regulation is a skill, and skills improve with practice. What doesn't survive is the implication that it grows reliably with use, without limit, and that the growth is the main lever. Beyond a threshold it doesn't grow with use. It degrades.
Like a muscle that tears instead of builds.
Like a muscle that tears instead of builds, and the tear doesn't announce itself until you try to lift something small.
Okay. That's the research picture. But what does it actually look like when you're in the hole?
Hilbert: It's about four hundred dollars.
Sorry?
Hilbert: The cot. For the dispatch room. We priced one in seventy-nine and it came to about four hundred dollars with the mattress, and the company wouldn't sign it, so we slept in the parking lot instead.
You dispatched.
Hilbert: Regional trucking. Nights. Small room, two radios, a phone, and a board with the routes on it. Mostly you sat there. The sitting was the job.
And the sitting was the stress.
Hilbert: The waiting was. Ninety-five percent nothing, five percent catastrophe, and the catastrophe always came at four in the morning when you'd already been up eighteen hours. That job taught me more about this than anything since.
More than the training programs.
Hilbert: We got one of those eventually. Videos and a workbook. Resilience training, they called it. The only thing that measurably changed was how good people got at saying they were fine.
That's the dark side of resilience with a product number.
Hilbert: The ones who lasted weren't the ones who toughed it out. They were the ones who figured out how to sleep in the car between shifts. Kept food in the desk drawer. Had one specific person they'd call at three in the morning. The ones who tried to push through burned out first, every time.
So the ones who lasted were managing load, not building capacity.
Hilbert: They were managing load. Then the company hired two more dispatchers so nobody covered three routes at once, and the burnout problem went away. Not the videos. The two dispatchers.
That's the whole episode in one hiring decision.
Hilbert: It usually is.
So if the muscle metaphor is doing harm, what replaces it? Because "resilience is an outcome produced by processes" is true and completely unusable as a sentence.
It's unusable. But the 2026 review actually gestures at the replacement. Target mechanisms during stressor exposure rather than training abstract individual factors.
Which in plain language means help people while the thing is happening, not in a workshop beforehand.
Help them while it's happening. Which is a staffing question and a scheduling question and a sleep question, not a seminar question. It's Hilbert's two dispatchers.
And it means the unit of intervention shifts. From the individual to the conditions the individual is standing in.
Which is a much harder sell, because you can't put a system in a workbook.
You can't. And there's a real tension there, because the individual-level stuff isn't worthless. Sleep, social support, mastery, physical activity, those are real levers and they're available to people whose load can't change this month.
Both are true. The personal levers are real and small. The load is structural and large. And the failure mode is pretending the small levers are the large one.
Which is what the metaphor does. It takes the structural lever and relabels it as a personal one.
Which is why the honest answer to Hannah is a fork, not a slogan. If you can reduce the load, do that first. If you can't, the modifiable resources are what you have, and they work, slowly.
The resilience paradox says don't expect to know in advance whether it's working.
Don't expect to know in advance. Expect it to be slow, expect the threshold to be lowered for a while, and expect the recovery to be measured in seasons, not weekends.
One more thing before we close, because I think it's the most useful reframe in the whole literature.
Go on.
The inverted-U. Moderate stress adapts you, excessive stress degrades you. Which means the question isn't "how do I become more resilient." It's "which side of the curve am I on."
That's the question. And you can't answer it from the inside, usually. That's what the person you call at three in the morning is for.
If you found this useful, a review helps other people find the show. Thanks to our producer, Hilbert Flumingtop. This has been My Weird Prompts. Email us at show at my weird prompts dot com.
We'll be back soon.
