Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

It's obvious now that you say it but I never thought about the AI companies themselves doing this for their own benefits like training purposes. It's a perfect testing ground to see what works for engagement and to see what real people want to hear back. The reason is pretty clear in that these AI/chat services have real people as users so logically it makes sense that the better sounding (not necessarily better) results make these users want to keep using. At the risk of sounding like AI... you're right... they may have been trained on old content but they are now using live data for fine tuning and quite frankly manipulation.

I miss the organic conversations and real thoughts from real people. I'm the type of person to read the comments before I read the article etc. It always gives more nuanced but also wildly different takes which I find interesting.

 help



Me too my friend. For the record and record's sake only, this is self-theorized and I have not the power, nor the ability, to prove these claims beyond my gut. But as you said, logically (double underline that in your head), from both my own recognition of patterned behavior, and to be honest, from fucking game theory and knowing that people (left unchecked) will naturally squeeze as much juice from the lemon as they can; If I were at a casino, logically and gastrointestinally (gut joke) I would remortgage my own home and drop the deed and keys on the table in order to stake my belief that this is happening. And I fucking hate casinos. Some journalist of much greater reach will hopefully be able to rip back the curtain, but those myopic fuckers have already destroyed trust. We had fun on the playground, we met friends, we learned rumors, we all felt free. But when you find out the jungle gym was greasing the bars on purpose to make us fall, just so they could learn about human bone strength, I doubt you'd visit one again.

And yes, I'd think the value of human to AI dialogue (ironically a single blind study, except the people are blind) is most likely massive. But fomenting? Plus (possible) financial fraud? Woooo boy, what an egregious mistake.

Cheers




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: