A Turing Test for AI? Identifying Leading Indicators
Credit: John F. Carrier

A Turing Test for AI? Identifying Leading Indicators

For reference, this is a photograph the post/sunroof in my car. It can be seen that water has leaked through the joint (symptom), after a rainstorm. I initially interpreted the problem as a leaky seal from the car window, but was uncertain this could explain the position and amount of the leak.

Two weeks later, after a more intense rainstorm, we found 2" of water in the front seat footwalls, resulting in an $800 repair job.

The root cause? A clogged front sunroof drain. Here is a photo of the exact same problem from the internet:

No alt text provided for this image

Question: If the the Internet can recognize a billion faces, why couldn't it recognize the leaky photo?

Question: How many leading indicators did you only recognize in hindsight? What was the consequence? Finally, why in industry do we constantly see the same symptom - root cause - consequence cycle without fixing it?

Brian Perlstein

Customer Success Focused Digital Innovator and Technology Leader

2 年

I have been hit twice by a clogged sunroof drain, both on the passenger side. Poor design with a tight bend that kinks the drain tube. You’d think the design software would have been able to identify this before it went into production.

回复

要查看或添加评论,请登录

John Carrier的更多文章

社区洞察

其他会员也浏览了