Fina (5⭐ Witchy Edition)

Fina (5⭐ Witchy Edition)

@kintotech.bsky.social

Piecing my life back together bit by bit. And learning new stuff about myself as I go. Maybe a writer eventually. Surviving Retail Hell until that point. Adult. Well over 21

140 followers339 following106 posts

Fina (5⭐ Witchy Edition) reposted

Demi Calavera
@demicalavera.bsky.social

Replying to @kim-hu.com

Wait... so... They're gonna be using AI generated "art" and "models" in their games but if someone else uses AI to generate something that uses anything theirs... that's wrong? The AUDACITY

Fina (5⭐ Witchy Edition) reposted

DX Orion
@dxorion.bsky.social

Ok! I am going to try and complete all my old Robot Masters i redid! First up! Punk'N Man! Embrace The Spooky!

Fina (5⭐ Witchy Edition) reposted

Jacob Oller
@jacoboller.bsky.social

Not only an incredible talent, but a rare talent who also thought it was funny to just voice random memes every once in a while

Ken Plume@kenplume.bsky.social

Unbelievable loss. An amazing talent, and just a really, really good person. www.instagram.com/p/De8RFy4pyov/

Fina (5⭐ Witchy Edition) reposted

Andrew Ducker
@andrewducker.bsky.social

Why is AI-generated language so hard to understand?

arxiv.orgWhy is AI-generated language so hard to understand?There is a widespread sense that AI-generated text has recently become harder to understand. In this study, we investigate a set of psycholinguistic features that are known to correlate with human processing difficulty (e.g., object relative clauses, long syntactic dependencies, low-frequency n-grams). Using a novel corpus of AI-generated text across diverse domains, we find that AI models use these difficult features more often than paraphrases generated by round-trip translation. Compared with human writing, all nine frontier and near-frontier models use more inanimate subjects with animate verbs and more low-frequency n-grams, and most also use longer dependencies, more stacked nouns, and more reduced object relatives in conversation; we also demonstrate substantial variation across model families. We propose and test a hypothesis as to why: while human language is optimized for successful communication, AI-generated text is increasingly optimized for other goals like math, code, and reasoning, in which there is a pressure for \textit{information density} rather than successful communication. We provide a proof-of-concept for this hypothesis by fine-tuning a language model under a length penalty, which leads to a divergence from human-like linguistic behavior. Our findings highlight the axes of misalignment between linguistic processing demands in humans and LLMs, and point toward possible remedies.

Fina (5⭐ Witchy Edition) reposted

rahaeli
@rahaeli.bsky.social

This is a fascinating paper about neurolinguistics

Andrew Ducker @andrewducker.bsky.social

Why is AI-generated language so hard to understand?

arxiv.orgWhy is AI-generated language so hard to understand?There is a widespread sense that AI-generated text has recently become harder to understand. In this study, we investigate a set of psycholinguistic features that are known to correlate with human processing difficulty (e.g., object relative clauses, long syntactic dependencies, low-frequency n-grams). Using a novel corpus of AI-generated text across diverse domains, we find that AI models use these difficult features more often than paraphrases generated by round-trip translation. Compared with human writing, all nine frontier and near-frontier models use more inanimate subjects with animate verbs and more low-frequency n-grams, and most also use longer dependencies, more stacked nouns, and more reduced object relatives in conversation; we also demonstrate substantial variation across model families. We propose and test a hypothesis as to why: while human language is optimized for successful communication, AI-generated text is increasingly optimized for other goals like math, code, and reasoning, in which there is a pressure for \textit{information density} rather than successful communication. We provide a proof-of-concept for this hypothesis by fine-tuning a language model under a length penalty, which leads to a divergence from human-like linguistic behavior. Our findings highlight the axes of misalignment between linguistic processing demands in humans and LLMs, and point toward possible remedies.

Fina (5⭐ Witchy Edition) reposted

Stu1701
@stu1701.bsky.social

I have, once again, commissioned @ncc-60205.bsky.social to put a bit of Star Trek into Star Wars. This time in the form of a Starfleet issue lightsaber! I'm calling it the Type 2-S Phaser (you can probably guess what the S stands for).

Fina (5⭐ Witchy Edition) reposted

WOLFSKULLJACK
@wolfskulljack.bsky.social

Anyway, here is Bosch that time when he turned himself into a wheel so he could get butt scratches and try and lick me at the same time

WOLFSKULLJACK@wolfskulljack.bsky.social

The wolf rescue FB account that I volunteer for got flagged today for wishing someone ‘happy birthday’?? Like it was soliciting money or some shit? WHAT

Fina (5⭐ Witchy Edition) reposted

Mike Bithell
@mikebithell.bsky.social

‘Customers should be less informed about the products they buy’ is such a wild stance to take

Kim Hu@kim-hu.com

I can't believe he's still at it, it's been days

Fina (5⭐ Witchy Edition) reposted

Kim Hu
@kim-hu.com

I can't believe he's still at it, it's been days

Fina (5⭐ Witchy Edition) reposted

Kim Hu
@kim-hu.com

Ah yes my favorite topic as someone with total aphantasia, who people LOVE to claim are utterly helpless without AI help. My mind cannot conjure images, sounds or any sensations. No inner monologue. No dreams. And SDAM, so I don’t even have vivid memories And yet I draw. Without the shit machine :)

This post is unavailable.

Fina (5⭐ Witchy Edition) reposted

Kim Hu
@kim-hu.com

it's because they literally have no visual education; so they do not know how to convey anything without the help of text tools don't make you an artist

Terminally Frogline@2frogline69.bsky.social

Continuing the tradition of Everything Being Labeled For Some Fucking Reason

Fina (5⭐ Witchy Edition) reposted

Erin Reed
@erininthemorning.com

In a major escalation in Dems embracing anti-trans positions, the Democratic Congressional Campaign Committee, the official campaign arm of House Democrats, released its first anti-trans ad supporting Nancy Lacore in SC-01. Transphobia is spreading among dems. Subscribe to support our journalism.

erininthemorning.comDCCC Releases Its First Anti-Trans Ad, Supporting SC Democrat Nancy LacoreThe ad is the first known ad from the high-level Democratic congressional campaigning arm targeting trans people this election.

Fina (5⭐ Witchy Edition) reposted

The Cutting Room Floor
@tcrf.net

Replying to @tcrf.net

The wiki is back online (but in read-only mode while we do some unrelated maintenance).

Fina (5⭐ Witchy Edition) reposted

The Cutting Room Floor
@tcrf.net

We'll be going offline for some unscheduled maintenance soon due to the recent PHP security issues. Fingers crossed.

Fina (5⭐ Witchy Edition) reposted

Nortuet
@nortverse.com

For any patreon creators, I just received a fairly convincing phishing e-mail claiming there was a valid copyright dispute and I had to log in to resolve it. I'm probably not the only one, so stay safe out there.