3 ms·

Obviously this was a bad idea implemented poorly. They never thought about the "why" for any of this. There's real value in fictional chatbots (Character AI) th

by jdlyga 2y ago

Obviously this was a bad idea implemented poorly. They never thought about the "why" for any of this. There's real value in fictional chatbots (Character AI) though. With Character AI, you can role play and have outlandish conversations that you might never have with a real person. There's even apps like SocialAI where every profile is AI so you can experience going viral and social media stardom without sacrificing your privacy (at least on paper). But a fictional Facebook profile? Just to like and leave comments? That's nothing you can't do already.

nitwit005 2y ago
I was surprised by the content of the profiles. It seemed as though they were deliberately not fun. It's better to pick a silly theme where if the AI does poorly, people laugh, rather than get upset.
hombre_fatal 2y ago
A lot of the blow back acts like Character.ai and similar setups aren’t wildly popular. This clearly was Meta’s attempt to come up with a more professional, sanitary version of it that might have appeal to people who aren’t techy enough to download C.ai and flip through cringe anime chat bots. And they went too far probably due to corporate design by committee. But the core idea is valid.
notatoad 2y ago
i would love to know more about the internal decision making that led to this release. it all just seems so weird. i can see the appeal of AI actors on social media - you get to make everybody feel more popular than they are because there's always somebody to reply to them, and you can make sure people get responses to their selfish requests that no other real user actually wants to see, without degrading the experience for any actual real people. but why would they implement it as a limited number of obviously handmade profiles with baked-in personalities? the internet is already full of real people's accounts that are, for all intents and purposes, without personality and ephemeral. real people sign up for facebook all the time who don't expose a personality, interact with a couple posts, and then disappear. LLM chatbots could be great at that. LLM chatbots can't be real people.
hapticmonkey 2y ago
Events like this remind me of Apple’s slide at the iPad announcement. It had street signs labeled “technology” and “liberal arts” and they discussed the intersection between the two. It can seem like a bunch of hippie crap but it really emphasised that we need more than just technology and engineering (and business engagement metric bean counters) driving decisions. So much incredible technology behind the scenes and for what? This shit? We need people with more perspective and humanity in the driving seat.

by jdlyga 2y ago

nitwit005 2y ago

I was surprised by the content of the profiles. It seemed as though they were deliberately not fun. It's better to pick a silly theme where if the AI does poorly, people laugh, rather than get upset.

hombre_fatal 2y ago

A lot of the blow back acts like Character.ai and similar setups aren’t wildly popular. This clearly was Meta’s attempt to come up with a more professional, sanitary version of it that might have appeal to people who aren’t techy enough to download C.ai and flip through cringe anime chat bots. And they went too far probably due to corporate design by committee. But the core idea is valid.

notatoad 2y ago

i would love to know more about the internal decision making that led to this release. it all just seems so weird. i can see the appeal of AI actors on social media - you get to make everybody feel more popular than they are because there's always somebody to reply to them, and you can make sure people get responses to their selfish requests that no other real user actually wants to see, without degrading the experience for any actual real people. but why would they implement it as a limited number of obviously handmade profiles with baked-in personalities? the internet is already full of real people's accounts that are, for all intents and purposes, without personality and ephemeral. real people sign up for facebook all the time who don't expose a personality, interact with a couple posts, and then disappear. LLM chatbots could be great at that. LLM chatbots can't be real people.

hapticmonkey 2y ago

Events like this remind me of Apple’s slide at the iPad announcement. It had street signs labeled “technology” and “liberal arts” and they discussed the intersection between the two. It can seem like a bunch of hippie crap but it really emphasised that we need more than just technology and engineering (and business engagement metric bean counters) driving decisions. So much incredible technology behind the scenes and for what? This shit? We need people with more perspective and humanity in the driving seat.