YT's new comment response recommendations are bleak
Added 2025-05-10 20:42:26 +0000 UTCComments
in hindsight it is borderline comical how a platform built by technical ppl who don't understand human connection, left to be run by MBAs and accountants who don't understand the tech it's built on, all packaged and marketed as something for ppl to connect with one another, sounds. Literally like the outcome of a "Worst Ideas Ever" workshop
Josh Brooks
2025-05-12 02:03:06 +0000 UTCWorse than suggested replies: Instagram offers "influencers" the ability to create AI bots based on their public personae and interact with the public on their behalf. I can't tell whether this will serve as a release valve for unhinged fans to focus their energies OR will just worsen already unhealthy interactions. Given their track record, I get the feeling some PM at Meta watched Satoshi Kon's PERFECT BLUE and thought "Hm, that gives me an idea!"
Dan Engler
2025-05-11 22:40:51 +0000 UTCthe casually misogynistic math one.. these things really were trained on like a decade of Reddit comments huh. basically what I'd hoped the vid would be!! Glad we can indeed see it. So many thoughts. Like I said Gmail's generated replies never bothered me too much since they seemed innocuous like the old form responses you mentioned. Or just laughably off the mark otherwise. These seem more tactfully ambitious in a bad way like how most LLM word waste is. But worse since this is YouTube and it's not like they call it "YouMail" right? Like intentionally bringing this sort of sterile prose into any public forum but especially the unabashedly human melting pot of YouTube comment sections feels like a violation of something sacred. on par with Zuck talking about generating entire users on his "social networking" site. One more for the pile of "owners understanding their platforms fundamentally opposite from how users tend to." But I (as I tend to do) think this problem is a bit bigger than just AI/LLMs. Like I've been thinking lately that to properly respect language is to view it as something more than that which demands a response. And in turn, to view a response as something more than that which is demanded by the mere presence of another's language. Like I can recall multiple group chat messages that've prompted a thought like "yk it's okay to say nothing if you don't have a substantial response. It won't hurt my feelings. It's not like I sent this specifically to elicit A Response From You." We were certainly capable of this pre-GPT, and I'm not sure whether trying to further encourage speech and behavior like this will actually have its intended effect. Maybe it just accelerates the fatigue. yeah, I, I don't know. I guess I.. yeah, I don't know (have punctuated so many verbal passages like this haha. the actual sound a brain makes upon spending the last of its cognitive fuel)
Josh Brooks
2025-05-11 07:05:19 +0000 UTCNow patreon lets me see the video 🙌
bxtcher
2025-05-11 06:38:01 +0000 UTCIt said you removed the video, so I refreshed and now there's no content at all. You clearly displeased the AIs.
Dan Engler
2025-05-10 21:28:24 +0000 UTCDoes it say "I removed this video"? Because I didn't?
Hendry
2025-05-10 21:25:19 +0000 UTCThat would be a good plot for a video game. You're left with the evolved chaotic garbling of AIs on an empty world and have to figure out what happened. Kind of like the game Tacoma, but the gameplay is detangling AI garble
Hendry
2025-05-10 21:15:15 +0000 UTCI can't promise not to ❤️ this video, but I promise not to at least not ask for permission first. If the whole human race were to be suddenly abducted by aliens, the AIs wouldn't notice and would just go on talking to each other forever.
Karl Bunker
2025-05-10 20:55:25 +0000 UTC
