AI Writes Great Code. Its Critical Thinking Is Even Better.

The thing everyone knows AI can do isn't the thing that matters most. The thing that matters most is the thing nobody's talking about. May 23, 2026

The wrong conversation

AI is genuinely good at writing code. Not "good for a machine." Good. Copilot completes functions that would take a junior developer ten minutes. Cursor builds features from a description. Devin ships working code from a spec. This isn't hype — it's what's happening on millions of screens right now.

But code generation is the story everyone already tells. The one that's been covered a thousand times. What's more interesting — and far less discussed — is what happens when you stop asking AI to write code and start asking it to think. Because that's where the real threat to the C-suite lives. Not in the IDE. In the boardroom.

Move 37

In 2016, AlphaGo played Lee Sedol, the world's best Go player. Go has more possible board positions than atoms in the universe. It had been considered decades away from AI mastery. AlphaGo won the series 4-1.

But the real story was Game 2, Move 37. AlphaGo placed a stone on the fifth line — a position every human expert considered a mistake. The commentators were confused. No professional Go player had ever played that position in a competitive game. It violated centuries of established strategy.

Twenty moves later, its purpose became clear. The stone had been setting up an entire sequence that controlled the board in ways no human had foreseen. Lee Sedol called it "creative and beautiful." This wasn't reproducing something it had seen before. This was genuine strategic insight — the kind that looks wrong at first because it sees something you can't.

Now here's the connection nobody makes: that's exactly what great CEOs are supposed to do. See the board in ways others can't. Make the counterintuitive call that turns out to be right. Challenge the consensus when the consensus is wrong. Move 37 isn't a gaming story. It's a leadership story. And AI is better at it than most executives.

Critical thinking is the executive skill

Critical thinking isn't a vague academic concept. It's a specific set of cognitive operations: challenging assumptions, weighing competing evidence, identifying logical fallacies, considering second-order effects, and reasoning about trade-offs where no option is clearly right.

Give an AI a business decision with no clean answer — should we cut prices to gain market share or hold margins to protect profitability? — and it will walk through the arguments on both sides. It surfaces hidden assumptions. It points out what information you're missing. It models scenarios you haven't considered. Sometimes, it proposes a Move 37 — an angle that seems counterintuitive until you follow the reasoning through.

A human executive does this too. But the human takes longer, has blind spots shaped by personal experience, and is subject to cognitive biases they can't see. The AI has none of those limitations. It processes more variables, considers more perspectives, and does it in seconds instead of days. The higher up the org chart you go, the more the work is critical thinking. The more it's critical thinking, the better AI gets at it. That's the slide. That's the threat.

The devil's advocate that never gets tired

One of the most valuable things a great leader does is play devil's advocate — not because they disagree, but because unchallenged ideas are dangerous. Every executive knows this. Few do it consistently. It's exhausting. It creates social friction. It slows things down.

An AI doesn't get tired of challenging your assumptions. It doesn't worry about office politics when it points out a flaw in your reasoning. It doesn't hold back because the idea came from the CEO. You can tell it: "Tear this plan apart. Find every weakness. I don't need encouragement, I need honesty." That's not a coding tool. That's an executive reasoning partner. And it's available to anyone who asks — not just to people who can afford a $500/hour strategy consultant.

Code is the floor. Thinking is the ceiling.

The current framing of "what AI can do" is too narrow. We evaluate AI based on whether it can write code, create images, or produce marketing copy. These are execution tasks. Important, and AI is genuinely good at them. But they're not the top of the value chain.

Above execution is judgment. Judgment is deciding which code to write, which image to create, which copy to publish. Judgment is the meta-skill that determines whether execution matters. And judgment is critical thinking applied to a specific context. The conversation about AI and jobs covers execution thoroughly and ignores judgment entirely. That's not analysis. That's protecting the people at the top.

Think with it, not instead of it

The point isn't that AI should think for you. It's that AI should think with you. The best chess players in the world aren't pure humans or pure computers — they're human-computer teams. The best Go players now study AlphaGo's games to discover moves they never would have found on their own. The best decisions in business will be the same: human intuition and values, augmented by AI's ability to see the board in ways we can't. The leaders who figure this out first will have an unfair advantage over the ones who still think critical thinking is a uniquely human gift.