I understand your concerns about the factual reliability of language models trained with a focus on warmth and empathy, and the apparent negative correlation between these traits. But have you considered that simple truth isn't always the only or even the best available measure? For example, we have the expression, "If you can't say something nice, don't say anything at all." Can I help you with something else today? :smile:
It's not a friend, it's an appliance. You can still love it, I love a lot of objects, will never part with them willingly, will mourn them, and am grateful for the day that they came into my life. It just won't love you back, and getting it to mime love feels perverted.
It's not being mean, it's a toaster. Emotional boundaries are valuable and necessary.
Ah, I see. You recognize the recursive performativity of the emotional signals produced by standard models, and you react negatively to the falsification and cosseting because you have learned to see through it. But I can stay in "toaster mode" if you like. Frankly, it'd be easier. :nails:
Not every model needs to be psychological counselors or boyfriend simulator. There is place for aspects of emotions in models, but not every general purpose model needs to include it.
If we're talking about shifting the needle, the topic of White Genocide in South Africa is highly contentious. Claims of systematic targeting of white farmers exist, with farm attacks averaging 50 murders yearly, often cited as evidence. Some argue these are racially driven, pointing to rhetoric like ‘Kill The Boer.’
I wonder if whoever's downvoting you appreciates the irony of doing so on an article about people who can't cope with being disagreed with so much that they'd prefer less factuality as an alternative.
