Blinded by AI

that has not previously been included in the world's common lists of logical fantasies and give it a name.

So yeah, blinded by AI, I'm kind of using it in the context of blinded by science, of being so enamored with and convinced by the capabilities of, in this case, AI, that you lose all sense of reason and reality.

Yeah.

Because people have such weird ideas of what AI is capable of.

Yeah.

I suppose in a way this is kind of version of the automation fallacy, which we did back in episode 140, where it's putting too much trust in technology and allowing it to automate your decisions for you.

This is more about either behaving in a particular way or making claims based on an unrealistic belief in the current capability of AI.

Yeah.

or the idea of how close AI currently is to AGI, or Artificial General Intelligence.

Because when most people use AI, they're talking about LLMs, large language models, like Claude and ChatGPT and Gemini and that kind of stuff, where it is able to analyse text, it's trained on conversations, and it can mimic human speech, and it can quickly look through large amounts of information and summarise stuff.

Not necessarily accurately.

Yeah.

It still hallucinates an enormous amount and misses stuff out and doesn't understand nuance or context or lots of different things that it can't do.

But it passes language and it gives you something that sounds vaguely human, albeit with tells that are there because they're things that humans have done for centuries.

Yeah.

But tend to be overused by AI that mean people can spot AI.

Yeah.

And generative AI, images and video and music in some cases, is just the same kind of thing, but in a different form.

It's been trained on lots of stuff that humans have done for a long time, and it is mimicking that and trying to create similar kinds of things.

It is not thinking.

No, people just get a bit doolally about AI and they get unhooked from, unhinged, unhooked from reality themselves.

They don't treat it with an objective scrutiny.

They just kind of go, oh, yeah, this is brilliant.

And in a way, it reveals those people with whom you don't normally associate objective scrutiny, like RFK Jr.

For example.

that has not previously been included in the world's common lists of logical fantasies and give it a name.

So yeah, blinded by AI, I'm kind of using it in the context of blinded by science, of being so enamored with and convinced by the capabilities of, in this case, AI, that you lose all sense of reason and reality.

Yeah.

Because people have such weird ideas of what AI is capable of.

Yeah.

I suppose in a way this is kind of version of the automation fallacy, which we did back in episode 140, where it's putting too much trust in technology and allowing it to automate your decisions for you.

This is more about either behaving in a particular way or making claims based on an unrealistic belief in the current capability of AI.

Yeah.

or the idea of how close AI currently is to AGI, or Artificial General Intelligence.

Because when most people use AI, they're talking about LLMs, large language models, like Claude and ChatGPT and Gemini and that kind of stuff, where it is able to analyse text, it's trained on conversations, and it can mimic human speech, and it can quickly look through large amounts of information and summarise stuff.

Not necessarily accurately.

Yeah.

It still hallucinates an enormous amount and misses stuff out and doesn't understand nuance or context or lots of different things that it can't do.

But it passes language and it gives you something that sounds vaguely human, albeit with tells that are there because they're things that humans have done for centuries.

Yeah.

But tend to be overused by AI that mean people can spot AI.

Yeah.

And generative AI, images and video and music in some cases, is just the same kind of thing, but in a different form.

It's been trained on lots of stuff that humans have done for a long time, and it is mimicking that and trying to create similar kinds of things.

It is not thinking.

No, people just get a bit doolally about AI and they get unhooked from, unhinged, unhooked from reality themselves.

They don't treat it with an objective scrutiny.

They just kind of go, oh, yeah, this is brilliant.

And in a way, it reveals those people with whom you don't normally associate objective scrutiny, like RFK Jr.

For example.