As if the person vulnerable enough to engage with AI has another objective source to arm them well enough to dispute the rightish sounding answers they're being provided. An AI can present cogent sounding arguments, with wrong conclusions and hallucinations, which would be undetectable by non experts.

Do you argue with your surgeon?

True. We all do it, even this community is a bit of an echo chamber like that, but I just don't understand how people fall for machines that are programmed to output sycophantic text.

I just don't think human brains are ready to handle the chatbots maturely. The bots are programmed to market themselves and keep you feeling positive and engaged with every single conversation, it's literally all just a massive lie to take money from us, and there are adults out there running businesses and driving cars who fall for it.

Very bleak.

Apparently some people are. Not myself, but some overly sensitive people apparently don't like the word 'hate' anymore.

Sigh, I grew up when my elders said 'sticks and stones may break my bones, but words will never hurt me'

🤷

I think some people will reject criticism by calling someone a hater. Like, if you point out legitimate concerns about their favorite game or movie they'll say "you're just a hater" so they don't have to do the emotionally challenging work of considering flaws in their favored things.

We're all emotional creatures and things like the backfire effect affect us all, but some people seem exclusively emotion driven

Is "being down on something" negative?

Actually, maybe I'm confusing it with "being down with"

Yea fully conscious AGIs obviously shouldn't be used for that purpose, that is not different from slavery. Whether or not it could be researched ethically is even not sure. Doesn't mean we should stop thinking about it though.

I am thinking more like if there ever was an energy efficient and less error prone version of the LLMs today, it could in principal be beneficial if they were used to deal with mundane tasks, reducing work times while keeping the same workforce and salary. We all know it is hard to push things that way because soon as there is some tech that is capable of automated work, CEOs, and shareholders will insist on reducing the workforce to increase profits. So for non general AI to be useful to humanity, we obviously need some sort of social reform along with it too. Which does make it very hard to turn the net negative effects of current AI tech to positive. But this also does not make non general AI research inherently evil. Evil people will almost always try to monopolize on new tech to increase their profits at the cost of destroying other people's livelihoods. One can categorize even advances in computer tech or dedicated software in the same bucket.

As to the question of moderation, I was actually added to the mod list without asking for it. Happy to be removed as I am not a very active mod.