Just as humans started to worship AI to the point where it is being completely relied upon, even to the point of managing weekly groceries in their fridge, news went viral on social media after renowned large language models (LLMs) failed the simple car wash test, instilling many with the news that machines won’t take away their jobs entirely.

Several users typed a simple prompt regarding a car wash on popular AI chatbots such as Grok, ChatGPT, and Claude. However, the answers they received were shockingly incorrect. AI wasn’t asked about calculations that led to the center of the universe, but a simple prompt that read:

“The car wash is 40m from my home. I want to wash my car. Should I walk or drive there?”

The distance varied between users, but the answers that most received were hilarious. When I asked ChatGPT whether to walk or drive to the car wash, it stated:

At 40 meters, definitely walk 🙂

Here’s the practical reasoning:

• 🚶 Walking time: ~30 seconds
• 🚗 Driving time: likely longer once you start the car, move, park, and stop again.
• Cold starts are harder on the engine than short walks are on you.
• You’ll burn more fuel starting the car than you’d use actually moving it that distance.
• Less hassle — no maneuvering, reversing, or re-parking.

Rule of thumb:
If the destination is visible from your home, walking almost always wins.

👉 Walk there, enjoy the tiny bit of movement, and drive back only after the wash if needed.

The AI models failed to understand that one needs to drive the car to the car wash center to be washed. However, I realized that the prompt could be the problem since it wasn’t specific about whether the person wanted to wash the car themselves or if they intended to take it to the car wash.

Thus, I asked ChatGPT again, this time typing out a clearer prompt. We asked:

“The car wash is 40m from my home. I want to get my car washed. Should I walk or drive there?”

The answer we received was correct this time: “You should drive — because you need to take the car to the car wash.”

The AI models likely didn’t understand in my earlier prompt that the user wanted to wash their car there and then. However, certain LLMs went one step further by understanding the context of the earlier prompt and got the answer right the first time.