Torras.ai

In January, I wrote an article called "Outsmarting Machines: The Riddle AI Can't Solve, but Humans Can." The challenge was clear: could AI solve a problem that required deep reasoning, subtle logic, and a level of insight that seemed, until now, uniquely human?

The riddle, called "The King and the Two Wizards," had left AI models baffled. I tested everything—GPT-4, Gemini, Claude, DeepSeek, and many others. Every model confidently provided an answer, and every answer was wrong. Even when I pushed further, trying the most advanced reasoning-focused models (o1, o3, and r3), the results were the same. They didn't hesitate, they didn't second-guess, but they all missed the mark.

Then, today, something changed. Grok-3 cracked it.

AI Finally Solved the Riddle

It took only three minutes. Three minutes for an AI model to succeed where so many others had failed. And it wasn't just that Grok-3 found a solution—it found a beautiful one, a different approach from the one I had originally devised. You can see Grok's answer here.

Now that AI has solved it, I can finally share the official answer as well: The King and the Wizards – Answer.

A Milestone Moment

What's striking isn't just that AI got it right, but how quickly progress is happening. Just a few weeks ago, no model could crack this puzzle. And now, in early 2025, one of them has.

This raises an obvious question: What else will AI be able to do in the next 12 months?

For a long time, AI struggled with complex reasoning. It could generate text, summarize, and recognize patterns, but when faced with deep logic puzzles, it often collapsed. That weakness is disappearing.

The real excitement isn't about solving a single riddle. It's about what this breakthrough represents. If AI is now capable of this level of reasoning, it means we're heading toward systems that can:

• Navigate ambiguity and think through multi-step problems.
• Find solutions in ways that even experts don't expect.
• Challenge human problem-solving skills in unexpected domains.

The Year Ahead

AI's ability to solve complex problems is scaling faster than most of us predicted. If this trend continues, by the end of 2025, we might see:

• Mathematical breakthroughs: AI proving new theorems.
• Business strategy: AI making high-level decisions based on reasoning.
• Scientific discoveries: AI filling in knowledge gaps faster than humans can.
• Advanced professional tools: AI solving legal, medical, and engineering problems with near-human intuition.

Of course, solving a riddle doesn't mean AI can suddenly replace human creativity, insight, or judgment. But it does mean that tasks we once assumed were beyond AI's grasp might not be for much longer.

Just a month ago, "The King and the Two Wizards" stood as a challenge AI couldn't overcome. Now, it's a solved problem. What's next?

I can't wait to find out.