[thw_9a83c]

I think we can call LLMs artificial intelligence. They don't represent real intelligence. LLMs lack real-life experience, and so they cannot verify any information or claim by experiencing it with their own senses. However, "artificial intelligence" is a good name. Just as artificial grass is not real grass, it still makes sense to include "grass" in its name.

[foxrider]

I'm going to get pedantic. That's why I've been using this exact term "LLM" instead of "AI" for the last few years. The "artificial" implies that it has man-made origins, yet serves the *same* purpose. Richard here argues that it doesn't fit because it does not, in fact, serve the same purpose. You can argue that for certain people it can serve the same purpose, like if you're a CEO who replaces low-level support staff with chatbots - but then it's an artificial support staff, not artificial intelligence.

[thw_9a83c]

> ...it doesn't fit because it does not, in fact, serve the same purpose.

For many people and purposes, it does indeed serve the same purpose. I use it all the time for coding, which is still very tricky, and for writing emails. For writing emails in particular, it is already a life-changing technology. I have always wanted my own secretary to dictate and finalize various letters. But, for some reason, companies don't provide secretaries anymore. Now, I can finally have an LLM instead. I guess there's no discussion that a good secretary must have always been quite intelligent.

[gaganyaan]

LLMs lack real-life experience in the same way that humans lack real-quark experience.

Does only direct sensory input count, or does experience mediated through tools count? How much do you or I really know about the Oort Cloud?