We Are Still Early in the AI Journey

When you follow AI developments and news closely, it can feel as though AI is already everywhere and that everyone is already using it like a pro.

Andrej Karpathy’s post on X yesterday suggests otherwise.

By his own napkin math:

• About 6 billion people (~75% of the world) have barely come into contact with LLMs.
• Another 1–2 billion use free versions occasionally, mostly for basic tasks like writing and search.
• Around 20 million people (~0.2%) use frontier AI models professionally for coding and maths.
• Just about 5,000 people, working in frontier labs, have internal access to the most advanced AI systems.

These are illustrative estimates, not measured global adoption figures.

This broadly reflects what we have observed in our work with teams across organisations.

Most people have barely experienced what AI can do today. The gap between AI’s actual capabilities and people’s understanding of them is enormous. And that gap is only widening.

I see this more as an opportunity than a problem.

We are still very early in the AI journey.

Reference

• Andrej Karpathy’s post on X