kt

kt

@kaiite.bsky.social

shitposting, mostly

92 followers125 following191 posts

kt reposted

Read this thread, and then remember everything I've been trying to tell y'all about the ways that "Generative AI" depends on violating consent and is being used to undermine the foundations of knowledge, consensus reality, and democracy.

Karen Attiah@karenattiah.bsky.social

So here is what concerns me. Yes LLMs are text generators, we know, we know. But text and output deletion? To delete inputs or questions without consent and then to be programmed to *lie* and gaslight about it.. is… something. Has anyone experienced this?

kt reposted

Nice reply "owning" me, unfortunately you parsed 2 levels of irony but missed completely the third, wherein it's clear I'm making fun of the persona your reply intended to critique. That being said I'm still hurt to an almost unimaginable degree

kt reposted

For the first time in history, DSA has 80,000 members in good standing! Tens of thousands of new members have joined the last year alone. In the face of rising fascism, a US-backed genocide, and an escalating climate crisis, socialism is the only path forward. 1/3

kt reposted

I've been telling people for years that there is a difference between *reducing cost* and *moving cost around* - and LLMs are only doing the latter. Maybe now that there is a study, they will listen...oh, who am I kidding.

hbr.orgAI-Generated “Workslop” Is Destroying ProductivityDespite a surge in generative AI use across workplaces, most companies are seeing little measurable ROI. One possible reason is because AI tools are being used to produce “workslop”—content that appea...

kt reposted

Just a reminder to check for your name in this list of books that OpenAI trained from. If your name is there, they probably owe you several thousand dollars. OpenAI cried that if everyone eligible author files, the company will go bankrupt, so I'm alerting every author I have ever spoken to.

theatlantic.comSearch LibGen, the Pirated-Books Database That Meta Used to Train AIMillions of books and scientific papers are captured in the collection’s current iteration.

kt reposted

“That’s all you can do, is learn and adapt,” said freelance writer Kiesha Richardson. “And I have some colleagues who are adamant about not working with AI. But I’m like, ‘I need money. I’m taking it.’” The WGA literally had a strike to prevent this & you should too www.nbcnews.com/tech/tech-ne...

nbcnews.comHumans are being hired to make AI slop look less sloppyIn the age of automation, human workers are now being brought in to fix what artificial intelligence gets wrong.

kt reposted

i think about the “do not split” principle from hong kong often

Dan Nexon@dhnexon.net

We have a decades of scholarship on how pro-democracy forces win in competitive authoritarian systems: they form a united opposition that brings together people with otherwise profound political disagreements.

kt reposted

regardless of what happens to the tariffs, they're doomed from the start because the average conservative holds all three of these positions at once: — virtue signals about supporting US manufacturing — against increasing the minimum wage — buys foreign imports because they're cheap

kt reposted

The reason the dems are so opposed to Mamdani's winning the primary is that he is direct proof that running a progressive campaign actually does work. They just don't want to do it.

kt reposted

Premium Newsletter: The enshittification of AI has begun. With GPT-5, OpenAI has cut access to all other models, reducing the value of even their $200-a-month subscriptions. They join Anthropic in screwing their customers over by making their products worse. www.wheresyoured.at/the-enshitti...

wheresyoured.atThe Enshittification of Generative AIThanks for subscribing to Where’s Your Ed At Premium, please shoot me an email at ez@betteroffline.com if you ever have any questions. Yesterday, OpenAI launched GPT-5, a new “flagship” model of some...

kt reposted

Annnd @xai just deleted Musk's new prompt which told the Grok LLM to "not shy away from making claims which are politically incorrect, so long as they are well substantiated." That prompt seemingly caused the LLM to become Nazi 4Chan, and has now been deleted, but other recent changes remain.

Aaron Reichlin-Melnick@reichlinmelnick.bsky.social

My non-data-scientist guess is that by telling it not to be "shy away" from "politically incorrect" comments that are not "supported," the model has metaphorically unlocked 4Chan from the box marked "do not emulate." After all, it was trained on the internet, and 4chan produced a lot of "support."

kt reposted

Replying to @volts.wtf

We really need to make this point over & over again. When they say, "we're going to try to be more fair & represent the right's perspective," it *sounds* (especially to them) like trying to do better journalism, but in practice, it is the opposite, because the right is built on lies.

my bluesky feed is like funny joke, death toll in gaza, cute picture of bunny, news about fascism in america, funny joke about fascism in america

kt reposted

I’d have to see the full list to know if Jeffries gets more from AIPAC than he does from *everyone else on this list put together*. But I don’t need to see more to be mathematically certain he gets more from AIPAC than *the rest of the Top 10 put together*, as that is how orders of magnitude work.

Elizabeth Spiers@espiers.bsky.social

His top donor is AIPAC. By a LOT.

kt reposted

The problem here isn't just the answer but the question. Mamdani has denounced antisemitism and clarified his position on this random slogan *other people* chanted at a protest numerous times. This is only a "tough question" if you think it's your job to repeat bad faith Republican attacks.

Aaron Rupar@atrupar.com

Hakeem Jeffries on Mamdani: "'Globalizing the intifada' by way of example is not an acceptable phrasing. He's gonna have to clarify his position on that as he moves forward."

kt reposted

1. Fuck those fucking fucks so hard, seriously. 2. Reminder: if the mRNA vaccines make you feel like death, prophylactic antihistamines will help make you feel significantly less like death.

This post is unavailable.

kt reposted

With new tech, "people need some time acclimating to what ‘private’ means to them, and if they haven’t looked at that (Meta AI) firehose discovery feed, they might not think through what ‘public’ means in this context,” EFF’s @thorinklosowski.com told @huffpost.com.

huffpost.comWith Meta AI App, You Can 'Discover' People's Wildest Thoughts — But Are You Unknowingly Sharing Yours?Its "Discover" feed can be all too revealing.

kt reposted

I know this is not a new thing to express, but it still strikes me as pretty bad that AI has both been sold as an ultimate arbiter of unbiased truth through aggregating all knowledge *and* programmed to hook users by telling them “you’re not paranoid, everyone is out to get you”

kt reposted

kt reposted

AI is not useful.

Anil Dash@anildash.com

I don’t think there’s a single serious critic who says AI’s not useful? The concerns I see are about content rights, worker impacts, environmental effects, epistemological threats, or media manipulation. But that’s all predicated on the idea that it *does* work, often too well for people to manage…