mei y. 🍡

愛 || zillenial || he/her || fiction writer and literary editor || 🇨🇳🇯🇵 🇺🇸 || 🌈

[mei y. 🍡 @aiywrites]

did i miss it? 😓
hi #writersoftwitter ~ i'm ai! i'm an asian american in my mid 20's writing:

🌈 queer stories/romance
✨ contemp. fantasy/magic
🌏 bipoc characters

i like:

🎶 kpop
⌨️ mechanical keyboards
🫖 tea
🐚 marine life

よろしくお願いします 🙇🏻‍♀️

[freya @boxingraccoon]

also i posted this on tumblr for an anon asking about my list of books for improving on prose. i talked moreso about "reading with intention" but it does also include an incomplete list of works that i felt have taught me something about writing. happy to chat about it.

[tai 🦦 @itsjusttai_]

A reminder that if you shop on https://t.co/E5E94Nv2ua you can select your favorite indie bookstore and all of the profits of your purchase will go directly to them! They frequently run different sales (including free shipping) with a massive selection of books to choose from !!

[anna °❀.࿔ @M00NLIGHTFAE]

this is so weird ???

[sett's murder victim (RIP kiriko nananan) @PumpkinNMayo]

i love these comments by the editor making the accusations

[moni @iycsts]

my favorite books by and about women for women’s month!

[sam 🌸 @kindlestuck]

wait this is insanely good what do you mean it was written abt d20 omfg

[Tumisha(🤡)'s Stud Sheriff Boiwife 🫡🤠 @disrupthehuman]

who do we have to blame for the belief that plot hole means "something that isn't explained or resolved"

[Nav Toor @heynavtoor]

🚨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?

[Pop Base @PopBase]

Black-owned bookstores in the U.S. now have their own directory for the first time ever, launched by the nonprofit National Association of Black Bookstores — who found that:

• There are 306 Black-owned bookstores nationwide (8% of independent bookstores)
• 14 states have no Black-owned bookstore
• Nearly 90% report annual revenue at under $250,000
• Sales of books by Black authors declined by 14%, despite an increase in overall print sales

Check out the directory: https://t.co/1AuCvzBThm

[smidgeon 🪺🇵🇭 @smidgeoni]

Generative AI should be nowhere near children's media. These people are willing to stunt a child's creative and cognitive development to earn themselves a quick buck.

If you use AI to make anything for kids, I do not believe you truly care about children.

[raphlecia @aformofreverie]

sites such as internet archive, medici tv, operaonvideo, and sadlerswells also have full productions. be a pirate. i also highly suggest watching peking opera, shastriya nritya, and greek tragedies, and researching classical arts of countries whose cultures interest you.

[Tiny Rodo. 🌶️💜 @adesolaaaaa]

https://t.co/8oJnvpBqAy has almost every book in the public domain for free.

As we’re debating on the ethics of piracy, let’s simultaneously find solutions, okay? 🤍

[ꨄ︎ ✧･ﾟ* ★ @likeafantasy__]

free access pdfs of some palestinian historical books to read in your free time to educate yourself, friends, loved ones, coworkers, etc 📚🇵🇸

a thread:

[Pagebound @pageboundapp]

Human curation over algorithmic/AI curation any day! Our brains are much more creative 🌱

Pagebound is an anti-AI Goodreads alternative. You can think of our Lists feature like making an OG Spotify playlist for the community—the more niche & unique, the better

[Ellipsus @ellipsus_writes]

New to Ellipsus? Let’s take a quick look around. We’ll walk you through the dashboard, show you how to start writing, import your existing work, set themes, and lots more! More tutorials on the way.

[kz.ᐟ skzayn ambassador @zaynjjong]

he's really into buff men he's not even trying to hide it 😭

[ᴀʀɪ🌻 @stayclown0325]

🐺

#Straykidsfanart

[cam(ille chong)! 🐍 瑷薇 add LOVE GODS SINNERS ON GR @camiscribbles]

talking to non writers is so funny,, told a girl today I'm finishing up my manuscript and she was like "ah...do you use chat at least?" and I was like No Absolutely Not and she was like how do you write so much with no chat?? um like every author does 😭 god forbid we use chat??

[Kitty-Bit Games & Cosplay~ @kitty_bit_games]

Keep at it!
-Uninstall/turn off AI features.
-Respond to surveys and say no to AI.
-Downvote and speak your mind in reviews.
-Buy hardware from before AI integrations.
-Reply to corpos and big accts and let them know we don't want it.
-GenAI needs to go and pressure is the key.