Hacker News Reader

Comment by phh | Hacker News Reader

I think I agree with you (I even upvoted), but this might be an anthropomorphism.

Back like 3-5 years ago, we already thought that about LLMs: They couldn't answer questions about what would fall when stuff are attached together in some non-obvious way, and the argument back then was that you had to /experience/ it to realize it. But LLMs have long fixed those kind of issues.

The way LLMs "resolve" questions is very different from us. At this point, I think that if we want to prove that LLMs need to be rooted in the real world to achieve intelligence, we need to find some real-world phenomenon that is so obvious that noone ever wrote about it... but then we'd have written about it?

View on HN

Think of it this way:

Intelligent beings in the real world have a very complex built-in biological error function rooted in real world experiences: sensory inputs, feelings, physical and temporal limitations and so on. You feel pain, joy, fear, have a limited lifetime, etc.

"AI" on the other hand only have an external error function, usually roughly designed to minimize the difference of the output from that of an actually intelligent real world being.

Yeah, that's exactly my thinking man. We have to root intelligence in the real world, otherwise it will endlessly spin in these abstract loops.

Akin to how logic -- untethered by emotion, intuition and experience (wisdom, maybe if you want? Understanding? Sure) -- can justify any obscene conclusion, and can not discriminate between morality.

Reward functions or a system of values -- these things are rooted in real world experience. Logic is required, sure, but insufficient. At least alone! Haha :)

It's not anthropomorphism, it's just life.

The exploring of an environment, learning and surviving is not unique to humans, but all life on this planet.

Sure, some of us may not see them (our alive brothers and sisters on this planet) as "intelligent", but undeniably their learned, and hereditary behaviors (learned through evolution, I guess), are very intelligent, especially for survival, and are tethered to the real world. They are all part of an intelligence which we all share.

They are (these behaviors), in fact, from a certain point of view -- reflections of the real world, or models of it. In a way that is even closer than language, or at least, orthogonal. We need that orthogonal source of real-world data (which is even more enriched with reality-info than language is), to bootstrap these AIs to a higher level of utility. :) hahaha :)

Regarding your point on degrees of fidelity with reality, the way I explain that is that they (OpenAI et al, or the AI models) have extracted world-information from semantic mining of enormous data. That's good, but only up to a point, as we see.

I think we need real world experience to get the rest of the way. Or, put it differently, to make it a helluva lot easier to get that juicy world info haha! :)

I don't think we to "prove" it. It's not math. It's a blackbox. We just have to try it and see. :)

Reply on news.ycombinator.com