My love hate relationship with AI

AI is impressive.

AI is frustrating.

Recently, I was troubleshooting a technical issue with ChatGPT. It suggested a fix. I implemented it.

It failed.

I replied:
“That didn’t work. Here’s the error message…”

It gave me the same solution again.

I tried once more. Same answer.

We were stuck in a loop.

So I pulled out a quote commonly attributed to Einstein (though its origin is debated):

“Insanity is doing the same thing over and over and expecting different results.”

That broke it.

The model shifted strategies. It reconsidered assumptions. It proposed a different path.

Problem solved.

The lesson wasn’t about the bug.

It was about direction.

AI doesn’t automatically know when it’s stuck. If you don’t explicitly ask it to change strategy, it may continue refining the same answer — confidently.

The breakthrough came from a meta-instruction:

• Assume the previous solution is wrong.
• Propose a fundamentally different approach.
• List alternative root causes.
• Debug from first principles.

AI is powerful.

But the real advantage isn’t just access to AI.

It’s knowing how to steer it.

If you'd like, I can tighten this even further for maximum LinkedIn engagement or add a sharper closing line.