404 MediaLonely young people are likely better off texting a random stranger than talking to a chatbot, according to a new study.
https://www.404media.co/chatgpt-loneliness-study-college-students-random-strangers-texting/
Jigen@404mediaco Pretty pseudo random. Most people aren't youthful with youthful peers in the same physical locality.
When you consider what online communities are like, there's more than a 50% chance (probably over 90%) you'll end up messaging someone that doesn't reply or hates your guts.
Banapana@404mediaco This reminds me of a kind of hair-brained scheme I had: what if people signed up for a penpal service, but the penpals were highly tuned LLMs and the conversation happened via paper letters on stationary. Is that just nutty? I feel like the absence of the screen and the medium of paper might be calming. And it's something you can count on because you get a letter every couple of weeks.
Hue StL@banapana @404mediaco Trying to say this in the kindest way I can think of, but that idea misses the point entirely, it's not "nutty". Chatbots did no better than journaling for reducing loneliness. But texting a human being, on a screen, did. Sorry, but the missing ingredient isn't a printer.
@HueSatLight @404mediaco Without justifying my idea further, I respectfully disagree. The immediacy of the response matters (fomo/instant dopamine hits) and waiting builds the mACC (a brain component to do with willpower). The **medium** definitely matters, too. Letters don't have flashing lights and ads. They reflect and don't emit light. I don't think it would be that much different than a serialized novel, and I find reading on paper much more calming than texting.
@banapana @404mediaco the weight of your guess is not the same as the weight of the evidence provided in the article I read. Did you read it too? If not, I think you should.
@HueSatLight @404mediaco Actually, I went and read the paper. (https://www.sciencedirect.com/science/article/pii/S0022103126000417). Both interventions reduced loneliness—one less so than the other. So, you're right my kooky idea is not "nutty".
As an aside, I think their evidence and analysis are not that strong. Typical for a psych paper—reliance on self-reporting and ordinal rankings is just not a good measure IMO. But I still might look into that "pen pail" business? 🙄😅
Apen Ipal?
Letterbuddy?
@HueSatLight @404mediaco
I could do what all the hip social media companies do and just misspell it:
AI Pnpl!