well there's a reason for that and that is... hear me out... Episode 1 was actually the best of the prequels...

*hides*

yeah, but who knew that would be the case? These AIs do, in fact, pass the Turing test. But they do so by being believable, but occasionally stupid. And full of corporate, "I am going to talk a long time, but essentially about very little" talk, if you let them.

Now we are really gronking that the ability to be intelligent involves learning from mistakes. Where is the feedback loop that changes the training database? Can you argue an AI out of ever making that mistake again?

I've never hoped for a well pointed comments under a joke post, but here it comes...

I'll be honest with you, I actually don't know whether there even is something as a widely accepted full definition of an intelligence. Probably there is.

But I am pretty sure it contains more than the ability to hold human believable conversation.

What I miss with the current evolution of the so called AI is the "comprehension" - which is the basic prerequisite for true learning (including the possibility to really learn from mistakes).

For me this is not groaning, it's just naming the reality - current so called AI is, indeed, impressive in many aspects, but at the end of the day, as it completely lacks the true comprehension, it's pretty "stupid".