Grok Cannot Reason. AI Is Not Intelligent.
I’m not in the habit of ‘arguing’ with machines, even cuddly machines called ‘Grok’, labelled intelligent by the world’s richest man. But I came upon a thread on X where this machine was apparently attempting to counter the ‘conspiracy theory’ that Starmer, Merz and Macron were snorting coke on a train - or at least trying to conceal the evidence of coke snorting paraphernalia on display at the table where they were sitting. I just had to reply to this multi-million dollar artificial super-brain.
First, Grok claimed that the 'cocaine spoon’ was a stirrer for tea or coffee. But then it changed its mind and said that it was either a stirrer or a toothpick. This is what piqued my curiosity - because there’s not a lot of similarity between a stirrer and a toothpick.
So I quizzed the machine:
It did not address my question - twice. So I had to spell it out.
The logic being, that if the ‘high quality images’ being referred to by Grok cannot discriminate between a toothpick and a drink stirrer, then they cannot be of sufficiently high quality to rule out the possibility that the object is in fact a coke spoon. But poor Grok just couldn’t comprehend what I was getting at. This was its reply:
It was able to comprehend that there was a significant difference between a toothpick and a stirrer, but it could not apply this knowledge to logically conclude that a coke spoon was probably equally different from either of those objects, therefore a ‘high quality image’ which could not differentiate between the first two objects would not be able to positively discriminate against the third object. But Grok insisted that the object was either a stirrer or a toothpick, but definitely not a coke spoon. That is logically inconsistent. Grok is unable to reason logically. It is not intelligent.
Of course it can’t reason, for reasoning is far more than just a series of arithmetic operations, which is all a computer can do, despite the cleverness of the software.
In the words of Edsger Dijkstra (if you don’t know who that is, you really should look him up):
“The question of whether a computer can think is no more interesting than the question of whether a submarine can swim.”
“AI”, despite the hype, simply guesses sequences of words. It’s a mathematical model and mathematical models can only say what the modeler tells them to say.
Belief that “AI” will ever become “sentient” is little more than superstition. It won’t.
The tissue definitely appears to be a tissue, despite a fake AI generated image doing the rounds, which makes it look like a plastic bag. This removes the context for supposing that the toothpick or stirrer is a coke spoon. End of controversy I think, other than the fact that Macron still looks totally wired and still grabs a tissue and hides it with a very guilty look on his face. He is obviously whacko and Starmer is still obviously an arsehole who sounds like a broken down Dalek.