Transcript
The Algorithm Wrote This Episode
Here's something you should know before we start. YouTube suggested this episode to me. Its recommendation engine looked at my channel, decided that an episode about the dangers of quick answers would perform well, and generated an outline. Complete with sections on dopamine hits, notification culture, and — my favorite — "intentional boredom."
So an AI handed me a pre-packaged answer about the dangers of pre-packaged answers. And expected me to read it out loud.
I'm not going to do that. But I'm not throwing it away either, because that outline is the perfect specimen of what this episode is actually about: content that sounds like insight, arrives instantly, and demands nothing from anyone — not from the machine that generated it, and not from the person consuming it.
Today I want to talk about certainty. Who sells it. Why we buy it. And what it actually costs.
Before You Sort Me Into a Drawer
Let me get one thing out of the way, because I can already hear it: "Easy for him to say, another consultant telling us to slow down and distrust the machines."
No. I use AI every single day, and I have no intention of stopping.
I've been fascinated by technology since I was a child. Before I was ten, my friends and I wanted to build a treehouse, and I had what I considered a brilliant idea: we could plan the whole thing on my computer. A Commodore C16 Plus. It failed, of course — we lacked the software, the know-how, the money, and frankly the computing power. But the instinct was already there: here is a machine, and machines should help us solve our problems.
That instinct never left. Today I build with AI. WP Luminary, my WordPress security plugin, exists because a friend found malicious code on one of his websites and nobody could explain how it got there. Our suspicion: someone had used a modern AI model to find the vulnerability and exploit it. My conclusion was not "AI is dangerous, ban it." My conclusion was: if attackers use AI to find weaknesses faster, then site owners should use AI to find them first — and get an explanation in plain language, without needing to hire a specialist.
I also built my own AI assistant. Her name is Ada — after Ada Lovelace, the woman who was the first to recognize that a machine could do more than pure calculation, and who is often considered the first programmer. Ada searches my email, checks my bookkeeping, manages the CRM, adjusts my website. She saves me an enormous amount of time. I don't sell her — what I sell is what I learn from building her.
So no, this is not an episode about how AI is bad and we should all meditate more. Wrong channel.
The Line: Accelerator or Substitute
Here's where the line actually runs for me, and it's sharper than most people want it to be.
AI as an accelerator is fine for everything that doesn't require deep thought and where a mistake costs you nothing. The weather. A route from A to B. If the AI sends me the wrong way, I correct it mid-drive. If I forgot to mention I'm on a bicycle instead of in a car, nothing breaks. These are small, recurring tasks that eat my time and that a machine handles faster than I do. Delegate them. Ruthlessly.
AI as a substitute for thinking is where it goes wrong. And I have a simple test for where that line is. One question:
Would I put my name under it?
If an AI produces a text for me, a strategy, a script, a presentation — anything — I have to have engaged with it deeply enough to say: yes, I'll sign that. I stand behind it. I would have done it similarly myself. That engagement is the thinking. The moment I put my name under something I haven't examined, I've outsourced not the work but the responsibility. And responsibility is the one thing you cannot delegate to a machine.
Because here's the practical reality: sooner or later, someone will ask you about that content. And if you can't answer — if you're standing there, name on the document, with no idea what's in it — that looks exactly as bad as it is. At that point the AI didn't accelerate your thinking. It replaced it. And that, for me, is an absolute no-go.
Notice what this test does. It doesn't slow anything down artificially. It just refuses to let speed skip the one step that matters.
The Solution Business
Now let's talk about why this matters beyond your inbox. Because the addiction to quick answers isn't a personal weakness. It's an entire business model.
Time is the scarcest resource in almost every company I've ever worked with. And that's exactly why thinking is so hard to sell. Consultants don't sell thinking — they sell solutions*. The customer has a problem, ergo the provider must deliver a solution. Sounds logical. It's also, in most cases, nonsense.
No external party can walk in with a ready-made solution to your internal problem. What they walk in with is a standard solution they've sold ten times before, repainted in your corporate colors. What they *should* be selling is the commitment to *find* one: first understand what the actual problem is, then solve it. But that's nearly impossible to sell today — customers don't recognize it as a product, so providers don't offer it as one.
I once had a client who said all the right things. Yes, we want to understand the real problem first. Yes, take the time. It turned out to be lip service. What they actually wanted was a solution — delivered fast, and delivered from *inside one department*. The silo thinking was so strong that when I did what the problem obviously required — talking to the other departments, because the problem crossed all of them — I was told to stop. Don't talk to the other areas. Find the solution in your own silo.
You cannot do that. Asking me to solve a cross-functional problem without leaving my department is like asking me to turn water into wine without leaving my house. If you want wine, someone has to go outside, tend the vines, harvest the grapes. If I stay in the house, all I can do is open the tap. You'll get water. You'll get it fast. And you'll wonder why it doesn't taste like wine.
That project ended early, by the way. The people who had originally asked me to solve the problem were fighting the same walls I was. They gave up and resigned — and my contract ended with them. I never found out what solution the company eventually landed on. My guess: a fast one.
Speed Gets Rewarded. Doubt Gets Punished.
The YouTube outline framed all this as a self-discipline problem. Turn off notifications, embrace boredom, meditate. I'm not a fan of "meditate more and talk less" — although, in fairness, for some people talking less would genuinely help. My version is simpler: think first, *then* talk.
But the real issue isn't discipline. It's incentives.
Speed gets rewarded. The fastest bidder wins the contract. The fastest answer gets the praise. And doubt gets punished. If you take too long because you're actually thinking — if you tell a customer "the solution is something we'll have to find together" — you lose. The customer takes the guy who loudly promises a finished solution. How that guy arrived at his solution? Nobody asks. The person who needs a solution doesn't question the person who claims to have one. He expects it to exist, and he expects it to be good — after all, he's paying for it.
And there's a second incentive, a stranger one. We've apparently decided that a hundred-percent answer isn't allowed to exist. Ninety-nine percent — fine. Below ninety — not good enough. But a hundred? Impossible. So when someone actually does the work and delivers something with nothing wrong in it, people go hunting anyway. Some marginal detail with almost nothing to do with the topic gets dragged into the center and picked apart — just so there's something to criticize.
An acquaintance of mine put it beautifully: you take a mosquito, inflate it into an elephant, then stomp it out like a forest fire — so you can stand there as the hero at the end.
That's the environment we operate in. Speed is rewarded, doubt is punished, and thoroughness is treated as a provocation. Is it any wonder people prefer the fast, confident answer?
Confident Ignorance
Which brings me to the most expensive version of all this. I call it confident ignorance — certainty as the product itself.
Years ago, a client asked for my honest assessment on a topic. I gave it. He was impressed enough to ask me to sit in on a meeting the next day with a large consulting firm — not to participate, just to observe and tell him afterwards what I thought. The firm certainly didn't expect another external in the room, but the client insisted.
And in that meeting, they lied so hard the beams were bending.
The pitch was about introducing a particular piece of software. The consulting firm claimed they had acquired the software vendor's entire development team — the whole team from Barcelona, they said, would be flown in, because they now owned it. And given the size of this firm, that wasn't absurd. They *did* buy teams and companies regularly. A normal customer would have had no way to check.
Here's what I want you to understand about that moment: their confidence was so absolute that even *I* started doubting myself. I sat there re-examining my own assessment — the one I'd given the day before — three times over. Did I get it right? Am I the one who's wrong here? That's what polished certainty does. It doesn't just convince the ignorant. It destabilizes the informed.
There was just one problem. Those developers didn't sit in Barcelona. They didn't sit in Spain at all — different city, different country. And I knew that, because I happened to know some of them personally.
I told my client. The firm didn't get the contract, and I suspect they didn't get a foot in that door for a long time afterwards. Because the one thing customers will not forgive is being lied to — even if they can't always detect it themselves.
Now, one nuance, because this is Iconoclast Insights and not a morality play. There's a difference between bending the truth and fabricating it. When NASA wanted to send people to the moon, they couldn't hire people who had already done it — nobody had. They had to find the ones who were confident they could make it happen. That kind of confidence is legitimate, and without it we'd have no progress at all. The same applies in business: no company has *the* perfect ready-made solution. If I promise something I haven't built yet, but I have a solid concept and can genuinely develop it into something the customer is happy with — that's selling a direction, and there's nothing wrong with it. But selling a fabrication as established fact, purely to be *perceived* as certain? That's not confidence. That's a reputation time bomb — and when it goes off, it takes the customer's trust and your business with it.
What You Take With You
I'm not going to end this with a system. No "from now on, only do this, never do that." I can't spend twenty minutes criticizing pre-packaged solutions and then hand you one — you'd be right to switch off.
Instead, three small things, and one question.
First: the name test. Before anything leaves your hands — AI-generated or not — ask whether you'd put your name under it and defend it when someone asks. If not, you're not done thinking.
Second: when someone presents you with a solution, ask them how they got there. Not to catch them — to see whether there's a path or just a performance. The confident ones with nothing behind them get remarkably quiet at that question.
Third: the next time someone says "we need to find the actual problem first," don't treat it as a delay. Treat it as the only honest offer in the room.
And the question to sit with until next time: the last time you were completely certain about something — was that because you had examined it? Or because certainty was simply faster?
Because more often than not, the answer you reached for isn't the problem. The speed at which you reached for it is.
Thanks for listening. Until next time.
