[wizeman]

> People are most interested in what their peers have to say, not any random human being or human level AI.

> Now if this human level AI is working in the field as a professional, or at least in computer science academia the it’s opinions are valid and interesting. However if it’s merely an average best possible response then it’s pointless to hear it’s opinion.

Yes, I agree with that, but I don't think that's what the parent poster was arguing.

It's also clear that I didn't phrase my question as clearly as I could, because instead of "human-level AI", I should have said this instead: "a human-level (or more intelligent) AI that has equivalent or better knowledge/experience than the people who post on HN".

> Additionally if multiple people post chatGPT responses that’s just like talking to same person using humans as sock puppets.

Yes, I agree with this as well.

But as a counterpoint, (as far as I understand) it's possible to have ChatGPTs with different levels of knowledge/experience and different personalities, as evidenced by GPT-3's fine-tuning capability (disclaimer: I've never used this functionality, so I'm not 100% sure if this is correct).

[true_religion]

Here is a thought experiment. Let’s say there is a posting about best practices on HN. The commenters can discus the topic and create a new consensus on what best practices is.

AI even at its best won’t be a part of that consensus unless it’s also working in the field.

I think AI chats can help with self improvement, but public discussion with others is the only way to have community improvement. As a disclaimer, sure AI which is indistinguishable from humans would be a part of the community but that isn’t what we have, nor is it the trajectory of current AI. What we currently have are better and better parrots, and task specific AI.

[wizeman]

> AI even at its best won’t be a part of that consensus unless it’s also working in the field.

A single AI can theoretically have the sum total of knowledge/experience that all humans have ever written, be it on a book, research paper or more generally, on the internet.

It can even be trained/learn from audio and video streams (e.g. all podcasts and conference recordings ever uploaded on the internet, etc).

It could also learn from interacting with humans and other AIs.

In conclusion, it could learn not only as much as a human would (in the same way), but it could also additionally learn from all the learning materials that are available in the entire world, written or spoken in all languages (russian, chinese, arabic, you name it).

> As a disclaimer, sure AI which is indistinguishable from humans would be a part of the community

Yes, that's kind of the point I was trying to make :) Or perhaps, what if it was even better than humans at discussing topics on HN?

> nor is it the trajectory of current AI. What we currently have are better and better parrots, and task specific AI.

I think this is 1) moving the goal posts and also 2) ignoring recent research results, which show that the same language model (without any additional training) can be used to solve completely different tasks (such as solving problems that require real-world physical interaction/manipulation) as long as you encode those tasks as text input/output tokens, which means that besides all the different tasks that these models can already do successfully, they also seem to generalize well to other tasks.