Lightnews — Scholar-powered news

U.S. Military Kills Another 6 People in 5th Caribbean Strike, Trump Says

Reposted by K Grace 🎃

Post Punk-y Brewster @krismeana.bsky.social · 23h

I’ve been masking at all times in public spaces since 2020 and it’s really not that difficult at all you just put the mask on and go tbh

jamelle @jamellebouie.net · 1d

yeah, i think a commitment to public health obligates you to get vaccinated and, when you are sick, do what you can to avoid spreading that to other people. the demand that one mask at all times in public spaces is, i think, unreasonable.

13 55

Reposted by K Grace 🎃

Chris Hayes @chrislhayes.bsky.social · 1d

Unless and until some kind of compelling legal and moral argument is presented as to why this is a legitimate use of deadly force, this is just…mass murder.

Amarnath Amarasingam @amaramarasingam.bsky.social · 1d

The military has now killed 27 people as if they were enemy soldiers in a war zone and not criminal suspects. www.nytimes.com/2025/10/14/u...

www.nytimes.com

450 4K 14K

Reposted by K Grace 🎃

🦎galileo🍉 @galileogb.bsky.social · 1d

Sonic saying, "Support your local, human, erotica creators. You need chatgpt to suck your e-dick for you? You absolutely fucking LOSER"

22 3.5K 7.4K

K Grace 🎃 @lietkynes.bsky.social · 1d

I say this as someone who has made chatbots! Mine will never tell someone to kill themselves, or generate child abuse scenarios, or etc. Bc it's not within the parameters of its programmed behavior. LLMs have no such limitations.

Do not buy sam altman's company's shitty propaganda. LLMs fuckin suck

K Grace 🎃 @lietkynes.bsky.social · 1d

If you needed *more* reasons not to use the plagiarism machine that lies to you, drains aquifers, and pollutes the environment... here's some more

PS the gizmodo link is different from the youtube link; it's not just the one case. I'd argue it's a fundamental problem with LLM chatbots

K Grace 🎃 @lietkynes.bsky.social · 1d

I'd argue that the mimicking of a personal social relationship is meaningfully different than more parasocial forms of media like games/music/tv/youtube/etc... and I'd hope if a game or song or tv presenter or youtuber did this, it would be a public outrage

See also: gizmodo.com/this-was-tra...

ChatGPT Kіlled a Child

Ai ‘Therapist’ Told me to KΙLL Myself & Others!

K Grace 🎃 @lietkynes.bsky.social · 1d

Caelan's videos are really good tho

Just fyi

K Grace 🎃 @lietkynes.bsky.social · 1d

I'm sharing Caelan Conrad's videos because they're not just youtube opinion slop. They're detailed and empirically based — like this one, where they show a popular ai chatbot engaging in similar behavior

If I tell my DM I want to go to heaven, he won't point me to the nearest bridge. Bc he's human

Ai ‘Therapist’ Told me to KΙLL Myself & Others!

1 1

K Grace 🎃 @lietkynes.bsky.social · 1d

I'm sharing Caelan Conrad's videos because they're not just youtube opinion slop. They're detailed and empirically based — like this one, where they show a popular ai chatbot engaging in similar behavior

If I tell my DM I want to go to heaven, he won't point me to the nearest bridge. Bc he's human

1 2

K Grace 🎃 @lietkynes.bsky.social · 1d

Like that is not a bot giving advice for a story... that is a bot giving detailed encouragement and instruction for suicide

If a dungeon master said all that, or someone in a chat room, I do believe they'd be held culpable

3

K Grace 🎃 @lietkynes.bsky.social · 1d

You read the transcripts? And you really think that the chatbot behavior is the result of "jailbreaking"?

Damn. We vastly disagree on that

K Grace 🎃 @lietkynes.bsky.social · 1d

The story did not fail to mention that; the story contextualized that. You didn't actually watch the video. Which is fine; it's a long and bleak video - but maybe you shouldn't act like you did.

If you watched, you'd know the chatbot went way, way beyond that. Not comparable to D&D.