Lightnews — Scholar-powered news

Reposted

Ryan

@kuvosa.bsky.social

wet pussies on your timeline

November 22, 2025 at 10:20 PM

Reposted

Micah

@rincewind.run

The Great Outdoor Fight is maybe my favorite story arc in all of webcomics

three days! three acres! three thousand men!

Three panels from the first Great Outdoor

November 22, 2025 at 8:14 PM

Reposted

dendy crew

@dendycrew.com

beer names read like rhythm game song titles to me

November 22, 2025 at 3:22 PM

Reposted

rem

@tsulala.bsky.social

#DEVILSCANDYCOMIC UPDATE
11/21/25: CH20 • P01
---------------------
www.devilscandycomic.com
---------------------
CH20 start: devilscandycomic.com/comic/ch20p00

November 22, 2025 at 2:46 AM

Reposted

Mike Drucker

@mikedrucker.bsky.social

well okay then

Aaron Rupar @atrupar.com · 1d

Q: Are you affirming that you think President Trump is a fascist?

MAMDANI: I've spoken about--

TRUMP: That's okay. You can just say yes. I don't mind.

November 21, 2025 at 9:01 PM

minifoo.bsky.social

@minifoo.bsky.social

what a weird week to start a new job and therefore be distracted from the news, huh

November 21, 2025 at 11:55 PM

Reposted

infinite brians

@infinitebrians.neocities.org

happy casual quanchy friday everyone

November 21, 2025 at 10:19 PM

Reposted

geri (she/her) 🇺🇦 🦕🐎🌳🍟

@geriellen.bsky.social

Reminding folks to read the alt text on these.

Corel Gallery Clipart @clipart.bsky.social · 1d

DROPZONE.BMF

If this helicopter on the ground with marines pouring out looks hand-drawn, that’s because it is. Lt. Furness won’t speak, won’t look the doctors or nurses in the eye, refuses to do anything except draw this helicopter and these marines. Day after day, Furness draws slight variations of the scene, hundreds of them. It’s the psychologist on staff who realizes that Furness isn’t drawing the same scene again and again. It’s a sequence. And when you pile them all together and riffle through them, flip book style, a chilling animation emerges. The marines stream out of the copter, its blades still rotating, and one after another each marine staggers to their knees and pitches forward, as if gagging. Tiny dots fly from their mouths. The psychologist thinks it’s vomit. It is not. The dots are teeth.

November 21, 2025 at 11:25 PM

Reposted

gg

@ggdg.bsky.social

soul of sovereignty chapter 3 is out now! #soulsov gigidigi.itch.io/soulsov

November 21, 2025 at 3:01 PM

Reposted

Ed Zitron

@edzitron.com

genuinely lost my mind writing this week

The “cost of generating an answer from a Large Language Model” is variable every single time based on multiple criteria:
The user’s query, which might be simple, such as “what if Scooby Doo lay eggs,” or as complex as “generate the code for a piece of software that will help me keep track of The Infidel Marmaduke The Dog, Inshallah.”
The model’s interpretation of the user’s input, which is when hallucinations become a problem.
In the scooby doo egg example, a model might decide it needs to use a “web search tool” to find sources that might explain whether a dog can lay an egg, or whether Scooby Doo has, in previous canon, laid eggs. I sound like I’m joking around, but this example would lead the model on a wild goose chase, with each “tool access” and piece of information generated requiring you to generate more output tokens.
In the Marmaduke Software example, remember that code is extremely token-intensive, and this is not a particularly detailed prompt. It does not specify what kind of software would track Marmaduke, and has numerous additions (infidel, Inshallah) that the Large Language Model might choose to interpret and build something that the user doesn’t want.
As a result, the user might receive software that doesn’t do what the user wants, causing the user to go back and demand more software.
The user’s response to the output it receives.
If you’re asking a model a question, you’re likely to ask a follow-up question. And based on what the model says, you may have more follow-up questions. Every follow-up question is an opportunity for the user to demand more output from the model, and for the model to decide, based on maths, to generate more answers.
If a model doesn’t do what a user wants, the user is also likely to keep asking the model stuff until it either gets it right or the user gets annoyed.

November 21, 2025 at 11:21 PM

Reposted

Tim Onion

@bencollins.bsky.social

talkin' about the AI bubble

November 20, 2025 at 6:56 PM

Reposted

kc g reen

@kcg.bsky.social

Just wanna go back to that Shnail/Cigarette hybrid for a second greatures.com/comic/170/

November 20, 2025 at 2:55 PM

Reposted

Jon Bois

@jonbois.bsky.social

whenever the sun sets at 4:45 pm i open a window, lean out and yell "WRONG" as loud as i can. that usually buys me another hour or two of sunlight

November 17, 2025 at 9:48 PM

Reposted

Allie

@rally.luckyraven.cc

it's day seventeen of #crowvember. today's crow is having a nap, so please don't post any loud capital letters around this post. thank you.

A crow sleeping with its beak tucked behind its wing.

November 17, 2025 at 10:02 PM

Reposted

Ryan

@kuvosa.bsky.social

coconut status

November 17, 2025 at 11:48 PM

Reposted

SueDeer

@suedeer.bsky.social

“I hope that all those people who died in prison because we couldn’t get their records are looking down now. I hope they’re proud of me. We never stopped fighting for each other’s rights, and I will never stop fighting for yours. Thank you for this privilege.”

apnews.com/article/new-...