Member-only story
3 Things AI Still Can’t Do (And Why That’s Your Biggest Advantage)
Find your edge
For Non-Members link to the full story
Let me take you down memory lane. It’s March 2026. The ARC Prize Foundation releases a new test for AI. It’s called ARC-AGI-3.
It’s a set of puzzle games. There are no instructions, no rules. You’re supposed to figure out the goal on your own.
I know that sounds tricky. But guess what?
Humans solved 100% of them. While every top AI model scored below 1%.
The internet had a good laugh. See how bad AI is? It can’t replace us.
That laugh lasted 6 months.
On September 2, OpenAI’s GPT-6 took the same test. You know what happened next?
It scored 99.95%.
In 6 months, AI went from 1% to 99%. Wow.
And that leaves us with one important question. If AI is able to succeed in a test it failed 6 months ago, is there anything left that it can’t do? And is it even relevant anymore?
For this, I went through some of the latest research. And I came across three things that still stand a chance.
The best part? In each one of them, you can have your competitive edge.










