Confidence and competence

Hi, I’ve been thinking about confidence and competence, and how these two keep getting braided together in modern work until you can’t tell what’s carrying the load and what’s just making noise.
I’m not referring to it in a theoretical way. You’ve seen it. I’ve seen it. The sprint review where the person with the smoothest story gets the green light, while the person with the data gets the parking lot. The hiring panel that “feels great” about the charismatic candidate and barely reads the portfolio. The incident review where the calm voice gets credit for the fix someone else actually designed at 03:00.

You don’t need a conspiracy to explain any of this. Humans are social learners. We use signals to coordinate, through tone, posture, fluency, certainty.
Signals are useful. They’re supposed to carry information: I’ve tested this; I know the failure modes; here’s the plan; you can trust me. When the signal matches reality, groups move fast and safely.
Where it breaks is simple: we’ve started treating the signal as the substance. And once you start doing that, you’re not leading, but amplifying noise.In comms terms: the channel is busy, but the message is thin. People spend cycles guessing what’s true, triangulating from style instead of content. That wastes time at best and steers you into a ditch at worst.

Take the Tacoma Narrows Bridge collapse in 1940. Engineers were confident - confident enough to sign off on a sleek, flexible suspension design that looked brilliant on paper. But they missed the resonance effects of wind. The bridge didn’t politely disagree; it twisted itself apart and collapsed spectacularly. Confidence spoke loudly. Reality spoke last.
Or look at Theranos. A company built almost entirely on confidence - charisma, stagecraft, billion-dollar storytelling. Investors, media, even experienced medical professionals bought the signal. The substance never existed. For years, projection outcompeted proof, until the physics of blood testing refused to play along. Confidence can build hype, but it can’t rewrite reality.
You see the same pattern in software all the time. A team builds a flashy demo that dazzles stakeholders - animations smooth, metrics green, the room buzzing. But behind the scenes, tests are thin, edge cases ignored, monitoring an afterthought. The signal is bright, the substance thin. And two releases later, everyone’s firefighting bugs that were baked in from day one. Confidence shipped. Competence didn’t.

Let’s name terms.
Competence is the earned, repeatable ability that survives contact with reality. It shows up in working systems, not in slide decks. Confidence is the outward signal that says, you can trust me. Both matter. But the healthy sequence is hard to cheat: competence → respect → authentic confidence. One builds mastery and its peers see it over time; confidence grows from evidence. Flip that order and you don’t get a shortcut, but a mask instead.
If you’ve worked in tech, policy, medicine - anywhere with sharp edges - you’ve watched this inversion up close.
Advice culture says, “Just be confident.”
Social media rewards the assertive thread over the careful analysis. Leadership trainings over-index on “executive presence” because it’s fast to teach and quick to spot. Meanwhile, the core loop that actually keeps systems safe - do the work, prove the work, earn the trust - quietly erodes.

Here’s the information-theory version.
A good signal compresses reality enough to move a group without losing the parts that matter. A bad signal compresses the wrong things - keeping the vibe, dropping the variance, smoothing out the unknowns. You get very clean stories… that aren’t true enough to steer by.
We are living through this at societal scale: take AI - slick images of “events” that never happened; confident threads that outcompete peer review for attention. Even if you can tell the signal from noise today, drift is real. Over time, noisy environments raise error rates for everyone. Decision quality drops long before anyone notices the trend line.

I’m not making an argument against performance. Stagecraft has value when it’s bound to substance. A team needs a clear voice when the plan is good. The problem isn’t that people speak well; it’s when speaking well substitutes for knowing well. That’s the pivot where cultures start rewarding projection over proof. Once that happens, two predictable failures show up:
Selection failure. We pick the wrong people for high‑leverage decisions because they look like leaders.
Feedback failure. We can no longer tell, quickly and cleanly, which bets were good vs. which stories were good.
Both decay trust.

Now, about stakes. In low-risk domains, this is expensive but survivable. In high-risk domains, it’s lethal. Imagine a confident misread of a CT, or a bridge signed off by someone who loves a microphone or a policy authored by people who think tone is evidence.
The rule is boring and universal: the higher the consequence, the lower the tolerance for noisy signals. Good organizations enforce that; great ones ritualize it.
We used to have rough social guardrails for this. Guilds and apprenticeships tied reputation to verified skill. Peer ridicule - ugly, yes - was also a brake on blatant fakery.
Reputation was a fragile asset, and lying about ability got you frozen out. We’ve rightly rejected cruelty and mindless gatekeeping. But we also lost some precision tools of accountability - the shared expectation that claims are costly, that you must be able to show your work. Today, challenging a confident claim is often treated as hostility.
That taboo is how noise scales.

We don’t need to bring back cruelty to bring back the consequence. We need norms that say: Claims are welcome; evidence is mandatory. Not humiliation but verification. The point isn’t to punish theater; it’s to keep the stage wired to the truth.
If nobody pushes back, the bluff sets the price. And reality is very diligent with invoices. It charges principal and interest, on time, every time.
You can negotiate with people but you cannot negotiate with load, latency, biology, or budgets…

There’s a human layer here that’s easy to miss. The person performing competence lives in quiet fear of an audit. The armor gets heavier. The incentives nudge them to double down. Meanwhile the person with competence but less theate r gets the lesson: speak less, ship more, and watch others get promoted for narrating your work. That is how you bleed out the people who keep the lights on.
So what do we do - you, me, your team.
s an individual, build your signal from your substance. Don’t undersell; don’t cosplay. Speak in specifics - what you tested, what failed, how you know.
Use the sentence that saves teams: “Here’s what I know, here’s what I don’t, here’s how we’ll find out.”
That line projects confidence because it projects truth. When you don’t know, say so, and then shrink the unknown with a plan. Your credibility will compound faster than any performance trick.

As a manager or decision maker, instrument your culture so that reality wins arguments:
In design reviews, require unknowns, risks, and rollback plans up front.
In postmortems, separate story quality from decision quality; grade decisions by what was knowable then, not how clever the post-hoc narrative sounds now.
In promotions, weight outcomes and ownership over airtime. Make artifacts - design docs, dashboards, experiments - the first class citizens of evaluation.
In hiring, run work - sample tests and pairing exercises; treat interviews as noisy signals to be triangulated, not oracles.
You’ll be amazed how fast the theater de-optimizes when it can’t beat evidence.

As a team and a company, normalize measurability and challenge without humiliation:
Adopt the boring mantra: “In God we trust; all others bring data.” (Deming.)
Practice steelman + measure: repeat the other side’s argument at its strongest, then ask, What would we expect to observe if this were true? and go look.
Make disconfirming evidence a KPI. Reward the person who finds the flaw before customers do.
Treat clarity as a service: teach people to write tighter specs, cleaner dashboards, sharper readouts. Performance is welcomed, anchored to proof.

At the societal level, we need to rebuild shared literacy for signals. Teach kids and remind adults that certainty isn’t evidence. Make “show your work” fashionable again. Treat virality like a noisy channel, not a truth meter. And when AI keeps turning the volume up on the wrong signals, respond with redundancy: citations, replication, provenance and time.

All of this rolls up to one blunt point: confidence without competence is a liability. It may win rooms. It will lose reality.
It burns trust, misallocates authority, and taxes the people doing the work. Real confidence is quieter and stronger. It grows out of contact with the real, the measured, the shipped, the repaired. It doesn’t need a costume. It needs evidence.
We don’t need more swagger. We need a higher signal‑to‑truth ratio. We need norms that reward being right later over sounding right now. We need people, especially leaders, who can say, “I might be wrong; here’s how we’ll know.” That sentence is not a weakness. It’s competence speaking plainly out.
So let’s re- tune how we work. Reward what survives the audit. Make style serve substance. Build signals that carry truth instead of drowning it. Then the confident voices we follow will be the ones worth following - and the systems we build will deserve the trust we place in them.

“In God we trust; all others must bring data.” - W. Edwards Deming

These are weird times, and I fully reserve the right to be wrong. But everything I’ve said here is wholehearted. Godspeed and thanks for listening.