Just a weirdo

Hi, I'm an adhd Trans girl, aspiring voice actor, and an absolute nerd

[greygloverdraws]

She also goes by Gramma, gran, Granny Belnak, Grammy Bea, Bea, Bell, and Butcher Belnak.

Her Battle Axe is called Gertrude after her beloved.

[ferretukas]

Thank you, for whatever that’s worth.

[beardedmrbean]

[theman]

I GOT A FUCKING RAISE THE POTATO WORKED WTF

[tattooedzombigirl]

This potato works. Every. Fucking. Time.

[secondlina]

Reblogging because it’s a damn potato and I want to encourage people to assume potatoes are magical.

[sinksanksockie2]

w-what if potato is actually lucky

[yorgen-sh]

[monamoni]

For a few minutes each year, sunlight makes this Yosemite waterfall look like a river of fire.

[princess-valentineee]

Flowery said a lot of interesting things huh.....Wait a moment

....GUYS I HAVE A CRACK THEORY

[the-haiku-bot]

Flowery said a

lot of interesting things

huh…..Wait a moment

Beep boop! I look for accidental haiku posts. Sometimes I mess up.

[totally-cis-offical]

That would explain how he goes between dark worlds! Because he’s already there!

[maple-salad]

op says we can repost

[prettyponyprincess-prancespiety]

This deserves not to rot on a Reddit comment section

[finnwolfhards]

this is superior humor

[duke-chainsaw]

I have some to add:

[pyros-random-thoughts]

M O R E

[oh-great-autistic-gay]

I hope you enjoy

[bravelycowering]

[no-one-offical]

‎

[theoriginal-offical]

THE ORIGINAL??

[algrolo]

[algrolo]

[algrolo]

[blurchur]

Never let this die

[stevishabitat]

[the-everything-remover]

­

[i-send-stuff-to-gimmicks]

@loss-detector get em

[loss-detector]

[evilfuckignskeleton]

[viralfrog]

[dragongirlbunny]

[derinthescarletpescatarian]

Sam Handwich.

[sufficientlylargen]

If Pikiwedia says it it must be true.

[zooperdoopers]

I made Pikiwedia real. Works for any Wikipedia page. Use this wisely :)‌

[sufficientlylargen]

Oh, wary dell vone

[thememedaddy]

[status-quo-hater]

This version of the progress flag legitimately looks so nice

Gilbert baker rainbow, huge intersex circle, the design is cluttered but in a good way 10/10

[ID: A version of the progress pride flag with a large purple intersex ring outlined in gold, looping through pink, blue, brown, and black chevrons on the side, which have a base of white. The horizontal stripes are: pink, red, orange, yellow, green, light blue, dark blue, and purple. End ID.]

[kira-serialfaggot]

It's happening the morphing into Ohio

[the-lumpfish-queen]

[yuriwinged]

[autistic-af]

[jam-does-audio]

Girls. That's the original image. We've come full circle. It's always just been the state of Ohio

[14-opossums-in-a-trenchcoat]

Love when people accidentally recreate the meme like with the "fuck this post and happy birthday Sonic" one

[noze2222]

Love these two goobers

[bprecioso]

My first and probably only contribution to the fandom

[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?

[undeadcorvid]

Let me save anyone who's curious a step: Here's the paper on ResearchGate