[suncaptor]

linked tree (includes options to donate to Ghanaian projects)

petition to show support

all out donation link

[hometoursandotherstuff]

Too bad they don't really make this. They're so cute.

[douche-canoe-regatta]

So Obvious Plant *does* actually make everything you see in the photos, they're just super limited runs. You can buy things directly from their website or sign up for their email list to get notified of new releases. This is incredibly dangerous and damaging to the wallet.

[chongoblog]

ITS APRIL 13 YOU KNOW WHAT THAT MEANS

FETCH ME NEIL

[chongoblog]

[chongoblog]

HAPPY BIG TWENTY NEIL

[itsapmseymour]

They got new photos of the moon,

I knew she had colors hiding in there 🥹

[wizardimpersonator]

Correction to the link sourced by OP! The Artemis II team did not take these photographs. These were shot by a Ukrainian student named Ildar Ibatullin sometime in the last year or so. Furthermore, the Moon does not have these actual colors in reality; Ildar does a lot of post processing and compositing to saturate the colors in order to artistically represent the mineral composition of the lunar surface.

If you would like to see photos taken by the astronauts aboard Artemis II (and more!), here is the repository that NASA is keeping.

[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?

[i-like-to-cry]

GO FERB GO!!!🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪🔪

[gulyas069]

[gulyas069]

hmmm

[lovehael]

One of their last Tweets before the account got suspended:

With an update since getting it back:

[vi-stri]

PLEASE talk about Congo, Haiti, and Sudan, with the same fervor you give Palestine. PLEASE care about black struggle and suffering.

[flutterbvnny]

this fund gives menstrual kits to women in sudan and helps keep community kitchens running ! it’s only $10USD to donate for a menstrual kit!

https://www.globalgiving.org/projects/padsforpeace/

[a-rebellious-waffle]

oh cool the UK government covered up a spike in trans youth suicides after they restricted trans care.

prosecute everyone in that fucking government for murder.

[humanfist]

Trans Youth Suicides Skyrocketed In UK After Care Drawdown; Government Covers It Up

A Freedom of Information request to the National Child Mortality Database reveals that 22 trans children under 18 died by suicide in England

[sophie-frm-mars]

Not meaning to be snarky here but british trans people have been reporting on this for ages and it's only really making a splash in the US now that erin reported it. I'm glad that she's taking on the morally necessary international perspective but we've been pretty fucking loud about it. I just sometimes feel like americans don't appreciate that everyone else keeps up with your news out of necessity but the reverse is much rarer.

Trans Kids Deserve Better, the under 18s direct action group that do legitimately more than the vast majority of cis or trans adults, vandalised the NHS headquarters with bloody handprints and "release the trans suicide youth report" because they knew just from being trans kids that there was a massive spike that wasn't coming through in the numbers

TransActual and the Good Law Project have reported on the suicides quite extensively, doing FOIA requests and doggedly pursuing the people responsible who have repeatedly tried to cover it up with no real answers for even the incomplete data that could be accessed which disproved the claim of no rise in suicides. I don't remember which weeks it was now because it was so long ago but on Red Planet we've talked about this several times and keep an ongoing update on Wes Streeting's war on trans kids.

There has been extensive reporting in outlets even as liberal and wet as Pink News about clinic wait times, bolstered by the data accessing and reporting by QueerAF, WhatTheTrans and others.

Even Erin's article is largely just repeating what the Good Law Project's press release said. She hasn't done anything wrong there to be clear, it's just the lack of people broadly paying attention to this story has been a little frustrating is all. The UK getting treated like the 51st state is ingrained by reactionary political currents like the anti-trans movement and it needs to be called out very loudly especially when children are literally being killed by it.

I guess it would've been cool if more people believed the kids from TKDB and reported on it a lot sooner is all. I hope the links are helpful to people

[hannabec2]

Fábio Gomes Trindade, a Brazilian street artist, is celebrated for blending street art with nature, incorporating branches and flowers as living parts of his murals. His renowned works often depict women and children, where the natural greenery becomes their “hair,” resulting in a powerful and organic balance. Artist: @fabiogomestrindade

[kilsie]

So it turns out I actually have to build the life I want to live

[eroscestlavie]

how it feels after you do that btw

[theconcealedweapon]

[queer-as-city-folk]

Hope for the democratic party?

[animentality]

[orbitalmirror]

reblog if you're an actual person IRL who gets genuinely upset if you catch someone using ChatGPT and view it as a serious moral failing

[evalempire]

As an 80s kid, adults were worried calculators would make people unable to do arithmetic. Are we doing the same thing? Here's a few ways this is different.

Calculators did not think for you. You still needed to know basic math to formulate the problem.

Calculators did not lie to you. If you got the wrong answer it's because you gave it bad inputs. Garbage in, garbage out. Not because it made something up. That would be a broken calculator. People can and are being hurt and dying because ChatGPT will confidently feed you bullshit.

Calculators would give you the same answer for the same inputs. LLMs cannot because their answers are based on probability and randomness. Sure, you can run an LLM locally and control all the random seeds to get repeatable results as long as you keep everything exactly the same, same workflow and same versions and same parameters, but most people don't do that.

Calculators weren't spying on you. Can you imagine? Nobody would use that calculator.

Calculators did not promote a secret corporate agenda or worldview. No rich executive was tweaking the calculator's answers to appease their shareholders or advertisers or because the answer was "too woke". That would be a broken calculator.

Calculators did not pretend to be your friend and confidant. There were friendly talking calculators for kids, but you knew what you were buying. Nobody claimed it was a person, that would be creepy.

Calculators always worked. Provided you've got batteries. You didn't have to worry about having an internet connection or the site being down. Sure, you can run LLMs locally, but most people don't.

We owned our calculators. Nobody was going to change it out from under us. It would not suddenly start behaving differently because of an update. It would not get enshittified. It would not get arbitraility shut down. Calculators would continue to behave the same way until it broke.

We understood exactly how calculators work. Even the people who work on LLMs can't fully explain how they make decisions.

Does that make it a moral failing to use ChatGPT? "Moral failings" are super problematic. I prefer to simply check if I'm just being an old fart. I don't think I am.

I can confidently say this: ChatGPT, or any LLM, should not be a load bearing part of your life. If you meet someone who is dependent on ChatGPT or Grok or whatever, keep some distance.

[br1ghtestlight]

spent MONTHS looking for this stupid tumblr post bcuz i constantly want to reference it and it wouldn't come up no matter what i searched despite it being (what i thought) was a popular well-known tumblr post only to find that the original blog turned off reblogs and deactivated and that it only got 12k notes total. but im posting it anyway to preserve its legacy

[nyantara]

Protein Powders and Shakes Contain High Levels of Lead - Consumer Reports

CR tests of 23 popular protein powders and shakes found that most contain high levels of lead.

supplement loophole strikes again

[feralbnuuy]

"Before 1994, manufacturers had to prove herbal products were safe before selling them. That changed after Congress passed the Dietary Supplement Health and Education Act. The law sharply limited the FDA’s authority, leaving supplements far less regulated than drugs."

i have once again been harmed by william clinton

[imakeficrequestsandthendisappear]

if cis people are so great then why don’t they have a siberian orchestra