Competition, Not Control, Will Decide the AI Future

Originally published on Substack. 

Have you noticed that every few weeks, a new AI panic cycle hits the headlines.

Artificial intelligence will wipe out jobs. AI is too powerful to be left to markets. AI must be slowed down “for our own good.”

And—what a coincidence—the solution is always the same: more regulation, more centralized control, more political oversight.

That’s not caution. That’s fear dressed up as expertise.

The real question in the AI and technology debate isn’t whether change is coming. It’s who gets to shape it. Competitive markets or political gatekeepers. Entrepreneurs or regulators. Consumers or committees.

History isn’t ambiguous here.

Every major technological breakthrough—from mechanization to computing to the internet—followed the same arc: uncertainty, disruption, experimentation, and then explosive gains in productivity and living standards.

The gains didn’t come from preemptive regulation. They came from competition doing what it always does—testing ideas, rewarding what works, and killing what doesn’t.

AI is no different.

What is different is how fast policymakers want to freeze the process. They talk about “guardrails,” but what they’re really building are barriers to entry. Compliance regimes don’t restrain power. They entrench it.

Large incumbents can afford legal teams and regulatory navigation. Startups can’t. Neither can small businesses that want to use AI to compete with bigger players.

That’s how innovation gets cartelized.

The job panic is equally recycled. We were told ATMs would eliminate bank tellers, spreadsheets would eliminate accountants, and the internet would eliminate work itself. Instead, tasks changed, productivity rose, and new industries emerged.

AI will replace some tasks. It will also create new roles, raise output per worker, and dramatically lower the cost of doing complex things.

The real risk isn’t job loss from AI. It’s productivity loss from blocking it.

Ironically, the same people warning about concentration are advocating the very policies that cause it. Markets don’t create durable monopolies—governments do, by raising entry costs and picking winners. When AI development becomes permission-based, innovation slows, risk concentrates, and mistakes scale.

Open, competitive systems are not reckless. They’re resilient.

And globally, this matters. Innovation doesn’t stop because Washington gets nervous. It just moves. Countries that allow experimentation will attract capital, talent, and breakthroughs. Countries that regulate first and think later will import technology instead of building it.

AI doesn’t need a master plan. It needs breathing room.

Key Takeaways

• AI panic is being used to justify centralized control, not solve real problems
• Competition disciplines power more effectively than regulation
• Job disruption is temporary; productivity gains are lasting
• Overconcentration is usually caused by policy barriers, not markets
• Heavy regulation favors incumbents and freezes innovation
• Open, competitive AI systems reduce risk and expand opportunity
• Innovation thrives where experimentation is allowed

Closing Thoughts

The best technology policy is boring—and that’s a compliment.

Protect property rights. Enforce contracts. Remove barriers to entry. Let competition work. That formula built prosperity before, and it still does.

AI will shape the future whether politicians like it or not. The only real choice is whether that future is open, dynamic, and widely shared—or slow, concentrated, and controlled by the few.

Fear is easy. Control is tempting. But competition works.

Let it. Let people prosper!