TL;DR: AI isn't replacing workers. It's replacing people who never had a core. The ones who thrive have three things: a backbone, the guts to be the client, and a high tolerance for failure. If you're missing any of these, the amplifier will burn you out before it builds you up.
James here, CEO of Mercury Technology Solutions. Hong Kong — July 2026
There's a book I read as a kid. How the Steel Was Tempered. Everyone remembers the title. Nobody remembers the actual point.
The point isn't about steel. It's about what happens when you put something under pressure. Steel gets stronger. Putty gets flattened. Same pressure, different results.
AI is that pressure. And it's not subtle.
The Amplifier, Not the Replacement
Here's what everyone gets wrong about AI: they think it's a replacement technology. Like the loom replacing weavers, or the spreadsheet replacing accountants.
It's not. It's an amplifier.
If you have a backbone — a clear sense of what you want, what you value, where you're going — AI magnifies your reach. You can suddenly do the work of ten people. Not because you're working harder, but because you have clarity. You point, AI shoots. You iterate, AI delivers. The feedback loop is fast, cheap, and relentless.
But if you don't have that backbone? If you're hollow inside, drifting from one trend to the next, hoping something external will save you?
The amplifier doesn't care. It just makes the hollowness louder.
**The AI Paradox:** The same tool that makes the strong unstoppable makes the weak invisible.
Question 1: Do You Have a Backbone?
There's an old Chinese saying: 韓信帶兵，多多益善 — Han Xin leading troops, the more the better.
Han Xin was one of the greatest generals in Chinese history. He could command massive armies because he knew exactly what he wanted. Every unit had a purpose. Every maneuver served a strategy. He wasn't overwhelmed by scale — he was empowered by it.
But give a confused person an army, and you get chaos. Not because the soldiers are bad, but because the commander doesn't know what to command.
AI is your digital army. It can do almost anything you ask. But "almost anything" is the problem. If you don't know what you want, you'll get everything and nothing. You'll generate a hundred variants, iterate eighty times, and end up with a polished version of your own confusion.
The hard truth: AI doesn't think for you. It executes for you. If your thinking is garbage, your AI output will be garbage — just faster and more expensive.
I've seen this in my own company. The team members who thrive with AI are the ones who already had strong opinions. They use AI to stress-test their ideas, not to generate them. The ones who struggle? They're waiting for AI to tell them what to do. And it can't. It can only reflect back what you put in.
Question 2: Can You Be the Client?
Some people are terrible at being the boss. Not because they're bad people, but because they have no confidence in their own judgment.
Give them a team of AI agents — digital employees that never sleep, never complain, never unionize — and they freeze. They can't give orders. They can't set direction. They spend hours "preparing" themselves, doing the work manually first, before they dare to ask the AI for anything.
This is the person who hires a contractor to renovate their apartment, then spends three weekends learning plumbing before they can write the brief. They're not being thorough. They're scared. Scared of being wrong. Scared of looking stupid. Scared of the responsibility that comes with saying "this is what I want."
But here's the thing about AI: the feedback cycle is short, the cost is low, and the AI has no ego.
You can tell it to redo something eighty times. You can reject the first seventy-nine versions. You can change your mind mid-stream. The AI doesn't care. It doesn't get tired. It doesn't resent you. It doesn't bill you for emotional labor.
This is the superpower of the AI era: the ability to iterate at machine speed.
The people who win are the ones who treat AI as a simulator for their own judgment. They throw ideas at it, watch what breaks, refine, repeat. They're not afraid of being wrong because being wrong is cheap. They learn their own preferences by seeing what they reject.
In the old world, this made you a nightmare client. The kind of person who sends a designer back to the drawing board eighty times. The kind of person who doesn't treat service providers like humans.
In the AI world? That's exactly the right approach. Because AI isn't human. It has no feelings. It doesn't need sleep. The cost of iteration approaches zero.
Stop treating AI like a person you might offend. Start treating it like a simulator that exists to sharpen your judgment.
Question 3: Were You Raised on Fear?
I had a classmate in university. Brilliant kid. Top three in his province on the gaokao — China's college entrance exam, arguably the most competitive academic test on the planet. He was also an incredible gamer. The kind of person who could master anything if he applied himself.
But he never went to class. Never took exams. Partied for four years, then faced expulsion.
I went to him with a plan. "Look, we have two textbooks. We cram for two weeks. We take the grad school entrance exam. Back then, grad school was easy to get into — everyone was chasing jobs, not degrees. We get acceptance letters, then we go to every professor, hat in hand, and say: 'We're not going anywhere. We have grad school offers. Give us a break on these failed courses.'"
He refused.
His reasoning? "I did well in high school because I did well in middle school. I did well in middle school because I did well in elementary school. But I've wasted four years in university. How could I possibly get into grad school now?"
He went home. Dropped out. Last I heard, he was working a dead-end job in his hometown.
This is what it means to be raised on fear. From childhood, he was told: one mistake and you're finished. Fail a test in elementary school, and you'll never get into a good middle school. Fail middle school, and you'll never get into a good high school. Fail high school, and your life is over.
This is probabilistic nonsense, of course. Einstein failed exams. So did a lot of people who changed the world. But when you're a child, you don't know probability. You absorb the fear. It becomes a reflex. And even when you grow up and learn the math, the reflex remains.
The person raised on fear sees low fault tolerance everywhere. One failure means total collapse. One wrong move means game over. This isn't rational — it's conditioned. And it kills initiative.
I wasn't raised that way. I don't have a perfectionism complex. Wrong? Do it again. Edison failed 999 times before the lightbulb. That's not a tragedy — that's the process.
But my classmate? He saw one obstacle and concluded his entire life was over. He didn't just give up on grad school. He gave up on himself.
The Real Division
Here's what the AI era is actually revealing: the divide was never about skills. It was about constitution.
Some people are strong at the core. They have judgment, confidence, resilience. Give them AI, and they become superhuman. Not because AI makes them smart, but because AI removes the friction between their ideas and execution.
Others are fragile. They've been conditioned to fear failure, to avoid responsibility, to seek external validation. Give them AI, and they collapse under the weight of possibility. Too many options. Too much freedom. Too much requirement for self-direction.
The AI doesn't create this divide. It reveals it.
Think about it: in the old world, someone with no backbone could still survive. They could follow processes. They could execute checklists. They could be cogs in a machine that someone else designed. The machine provided the structure they lacked.
But AI is replacing the machine. And when the machine goes away, what's left?
Either you have your own structure, or you have nothing.
The Three Tests
So here's my framework. Three questions. Answer them honestly.
1. Do you have a backbone?
Can you articulate what you want, why you want it, and how you'll know when you have it? Or do you need someone — or something — to tell you what to do?
If you can't lead yourself, you can't lead AI. You'll have digital employees and no mission. That's not a team. That's a mess.
2. Can you be the client?
Can you give direction without knowing all the answers? Can you set a target and iterate toward it? Can you tolerate being wrong eighty times before you're right?
If you need to do everything yourself before you can delegate, AI won't help you. You'll be the bottleneck. The person who could have scaled, but chose to crawl.
3. Are you raising your own fault tolerance?
When something goes wrong, is your first thought "this is a disaster" or "this is data"? Do you see setbacks as permanent, or as feedback?
If you were raised on fear, you can unlearn it. But only if you recognize it. Only if you actively practice doing things that might fail. Only if you build the muscle of recovery.
The Bottom Line
AI isn't coming for your job. It's coming for your illusions.
The illusion that following instructions is enough. The illusion that someone else will provide the direction. The illusion that the world has low fault tolerance and one mistake means oblivion.
The people who thrive in the AI era aren't the ones with the best prompts. They're the ones with the strongest cores.
They know what they want. They aren't afraid to iterate. They treat failure as fuel, not as final judgment.
Everything else is just tooling. And tooling is cheap.
The question isn't whether AI will replace you. The question is: when the amplifier turns on, what does it amplify?
If the answer is clarity, judgment, and resilience — you're going to be unstoppable.
If the answer is confusion, fear, and passivity — you were already gone. AI just made it visible faster.
Mercury Technology Solutions: Accelerate Digitality.