Why do so many people on HN say LLMs aren't "artificial intelligence"

I am guessing you mean AGI? They usually mean they’re not generally intelligent, i.e., displaying cognitive flexibility for different types of tasks without training or extensive fine-tuning. Think of a human intern, you don’t need to tell that human intern beyond a simple phrase when you need something. The intern will figure out how to do it, including figuring out what they don’t know, and what they need to learn to do that thing.

https://en.wikipedia.org/wiki/Artificial_general_intelligenc...

It's a mix of trying to look blasé, of believing that a superficial knowledge of how they work is enough to dismiss them as purely mechanical, and probably genuine fear and denial for the implications of their existence.

LLMs are not just intelligent, they're general intelligences in that they're not limited to a single task (such as a chess-playing AI or a voice-recognition AI) but they're capable of any task that can be performed with text as input and output (which doesn't mean it's just text manipulation, the internals are not limited to text).

While the definition of intelligence is speculative, the way that LLMs work is well known. They don't apply any thinking or reasoning at all. It's not even conceptualized or attempted. Instead, they predict the most likely response to the prompt given the training data. We can all agree that statistical predictions alone are not intelligence.

Same reason why it can never pass itself off as a Scotsman.