Josh Kerr just smashed the world record for fastest mile.

But it took until 1954 before the first human being, Roger Bannister, ever broke a 4-minute mile.

But soon after the barrier was broken, others did the same thing.

For some reason, we humans sometimes need to see the next leap forward before we ourselves are capable of achieving it.

In the AI world, a typical debate goes like this:

Person A: “I used AI to code this cool app!”

Person B: “But your code is bloated, you didn’t optimize images, and you didn’t bundle your JS to save space on load…”

Person A: “Ok, I’ll just copy your exact complaints into AI, and watch it fix it.”

Because if you can explain the problem with a codebase clearly, you can use AI to fix the problem. This capability will only grow.

So the “gotchas” that are used to say that people experimenting with AI software should stop what they’re doing are really just prompts that lead to the next round of fixes.

If you are experimenting with AI in your workflows, don’t let the experts convince you to stop, because your personal app isn’t as secure/robust/optimized as a mission-critical, Fortune 100 commercial app.

Some of us just need to see what’s possible first to believe it.