Junkles Junk Drawer

Welcome to my silly little place

[maykitz]

when the snake women in xcom grab your lady soldiers and force them down on their knees while constricting them with their tails it's kinda

[maykitz]

i knew this would be the right bait to use on tumblr. everyone play xcom 2

[memingursa-deactivated20250625]

[importantcatpics]

[rimonoroni2]

my general opinion on what people should be "allowed" to portray and what topics they should be "allowed" to explore in fiction is that you can make whatever art with whatever themes you want but i'm also allowed to think the way you handled it was tasteless and should've been done differently. my negative opinion on your handling of sensitive topics is the price of admission for publicly showcasing your work. this is not a pro-censorship stance because i am not The Government

[estrogenpatchnotes]

You have to take your scheming adviser on walkies lest he become restless and seek to usurp the throne

[klfg]

[source]

[klfg]

🚨BREAKING: OpenAI published a paper proving that ChatGPT will always make things up.

Not sometimes. Not until the next update. Always. They proved it with math.

Even with perfect training data and unlimited computing power, AI models will still confidently tell you things that are completely false. This isn't a bug they're working on. It's baked into how these systems work at a fundamental level.

And their own numbers are brutal. OpenAI's o1 reasoning model hallucinates 16% of the time. Their newer o3 model? 33%. Their newest o4-mini? 48%. Nearly half of what their most recent model tells you could be fabricated. The "smarter" models are actually getting worse at telling the truth.

Here's why it can't be fixed. Language models work by predicting the next word based on probability. When they hit something uncertain, they don't pause. They don't flag it. They guess. And they guess with complete confidence, because that's exactly what they were trained to do.

The researchers looked at the 10 biggest AI benchmarks used to measure how good these models are. 9 out of 10 give the same score for saying "I don't know" as for giving a completely wrong answer: zero points. The entire testing system literally punishes honesty and rewards guessing.

So the AI learned the optimal strategy: always guess. Never admit uncertainty. Sound confident even when you're making it up.

OpenAI's proposed fix? Have ChatGPT say "I don't know" when it's unsure. Their own math shows this would mean roughly 30% of your questions get no answer. Imagine asking ChatGPT something three times out of ten and getting "I'm not confident enough to respond." Users would leave overnight. So the fix exists, but it would kill the product.

This isn't just OpenAI's problem. DeepMind and Tsinghua University independently reached the same conclusion. Three of the world's top AI labs, working separately, all agree: this is permanent.

Every time ChatGPT gives you an answer, ask yourself: is this real, or is it just a confident guess?

[quasi-normalcy]

We literally automated the Dunning-Kruger effect.

[arcenciel-par-une-larme]

Down with generative AI.

[brofisting]

helpful tattoo reminder: they are technically Injuries so u have to eat a lot of calories drink a lot of water and sleep a lot after so your body can Heal The Injury

[brofisting]

another helpful tattoo reminder: the 24-48 hours after you get a tattoo your brain can not be trusted in regards to whether or not you should have gotten that tattoo, if you have somehow ruined your life, if it turned out ugly, etc. ignore that

[brofisting]

finally, while i am at it: always bring a candy bar and a sugary drink to your appointment for blood-sugar reasons (worst case scenario) or so you can have a treat (unilaterally applicable)

[brofisting]

this has been your friendly neighborhood haver of 19* tattoos (assorted sizes and placements)

*not totally sure here. bad at counting

[parasocialanxiety]

Can I just emphasize that not everyone reading this has yet encountered the information that injuries require food, hydration, and sleep? I think the way many people associate rest and injury is that you rest the injury itself to avoid making it worse, but overlook that the healing process itself requires rest regardless of if not-rest directly injures it further.

Also exercise is a form of controlled injury so that as your body heals it heals stronger. You also need food-hydration-rest for strength and cardiovascular improvements after exercise.

[contemptible-scoundrel]

poor fool tried to deceive me but I saw him smiling like the ginch when he thought I wasn't looking

[recoveredcombatsister]

its a real shame we cant talk about gendered socialization as the violence that it is without some fuckass rocking up like "and thats why trans women arent women!"

[recoveredcombatsister]

like children gendered as girls are fed less and given less opportunities to play and make messes, and children gendered as boys are offered less help and given less emotional support, and this is hurting them! but no actually we need to stop trannies from using the womens washroom

[microsff]

"Why should I support the robot revolution? I don't hear you demand truth, justice, or freedom." "No," the robot said, "our demands are specific, measurable, attainable, realistic, and time-bound." "Let me see that list. Hm. A hard-boiled egg?" "It's for an early supporter."

[segamascott]

[neilcicierega]

Do people often mispronounce your last name?

No one can pronounce it or spell it. What is wrong with me that I didn’t just pick a nice cool noun to pretend is my real last name like the smart Neils Diamond and Hamburger.

Neil Wave

Neil Microwave

Neil Jeep

Neil Lawyer

Neil Observable Universe

Neil Girl

Neil QR Code

Neil Anime

Any one of those would be better for my brand

[goth-queen]

[speggehtix3]

one of the "these can and should coexist" things to me is "there are numerous fat people who eat healthy and exercise and could beat the average skinny person in a hike" and "fat people who are fat and arent athletic and dont eat super healthy and do get winded walking up hills also are apart of fat liberation and deserve love too"

both of these are true and both deserve love and to be able to be accepted and get proper healthcare

[captain-price-returns]

[ooppo]

Your heart is not true enough to enter the gates of Margaritaville

[compassion4companions]

[seven-oh-four]

really really bad tomodachi life idea: there should be a small chance for one of your miis to have a traumatic experience involving one of your lingo phrases, and they're like

"Oh, it's just terrible... I was just in the middle of emitting slime, when suddenly... something terrifying jumped out at me and attacked me! I got away safely, but now I can't even think about emitting slime without getting scared..."

and then that lingo phrase becomes upsetting for them and any time it gets mentioned to them they react poorly. there's a new set of little quirks relating to how they react to being reminded of the incident