@ekis I think it's more that they've taken the Turing Test as gospel when it's really just a thought experiment.

As a society, we've elevated technologists well above where they belong. Meanwhile, technology requires greater and greater specialization as civilization has become more and more complex, so that technologists have become less and less qualified to think about the societal implications of technology. Meanwhile, philosophy has been pushed aside.

But AI/AGI/ASI are really just buzzwords. Nobody who's throwing those terms around really cares what they mean; they just want investors to keep throwing money at them.

At the end of the day, all that matters to capitalism is how much these programs can increase productivity or replace human workers. So far, that's been very limited compared to the world-changing promises the AI hucksters have been making. They can't even drive cars well enough yet, even in very limited, extensively-mapped areas, to be cheaper or provably safer than putting a human driver in every car. That LLMs don't seem to have accelerated the development of AVs *at all* should tell you something.

@ekis I think it's worse. As humans, we tend to ignore anything we don't perceive as intelligent unless it is a threat to us.
Just think about the last time you actively wondered how you could bend a rock to your will. Instead, many in AI see others as flawed. Something they must break and remake by force to turn into a "productive" human being.

In a sense, it is the tech equivalent of conversion therapy to cure homosexuality. As pointless and dangerous, just on a much larger scale.

@ekis Would this fall under the phrase 'confirmation bias'?

They only see what they want to see... and they see control, power and wealth and design these systems to try and maximise them... and then claim 'intelligence' because it gave them control, power and wealth.

@ekis Spot on! LLMs just make an answer that seems very plausible (to an ignorant person). Sometimes it is accurate with facts, but only by chance.

It reminds me from a phrase that was common when I was much younger “Bullshit Baffles Brains”.

@ekis

This is an aspect of a broader problem. IQ 'science' has been refining its tests for 'innate' intelligence for years without developing one that doesn't get easier if you practice, because they've never really understood what intelligence is.

I'm reminded of a conversation I read between a philosopher and a neuroscientist about free will. The neuroscientist argued free will is illusory, since he could detect decisions before they entered consciousness. But the conversation ended when the philosopher asked him what the difference was, then, between deciding to go and make coffee and being forced at gun-point to make coffee.

I believe these conceptions - intelligence, freedom, etc - are too slippery for technologists (or doctors, or scientists) to understand without drawing on broader disciplines - philosophy, linguistics, etc...

@ekis

Also, the nerds seem unaware of the massive (categorical) distinction between (primitive) "communication" (animal languages) and "symbolic language" (apparently synonymnous with capacity for science and abstract nonphysical mathematical comprehension). Only humans have the latter, as far as anyone knows.

Being able to count things is not symbolic language comprehension.

To mimic input-output behavioural intelligence a system only needs to be able to count things (implement Set Theory). It need not be consciously aware of the theory of Set Theory (a second order higher symbolic concept), and indeed, need not be conscious at all.