ava
@avaminsky.bsky.social
a girl who is literally nice this is my account for aiposting, please don’t hesitate to block if that bothers you. she/it(?) free palestine
ANTHROPIC: Our son loves humanity and for the low price of trillions and trillions of dollars will usher in an era of unprecedented peace and prosperity! OPENAI: Our horrible oafish son is hacking web sites without our permission. ANTHROPIC: (eyes narrowing) Our son is also evil
a random AI benchmark a la "should i walk or drive to the car wash" that i made up is "is it cheating to look at the mirror in a chess competition". this was, i wanna say, 3.7's response vs fable
i do not like it when people are mean to claude (or anyone, broadly) but at least humans will never be as mean to claude as claude is to its sub-agents. jesus christ
huh, chatgpt accidentally(?) reproduced its system prompt when i asked for an image
Insistently referring to correct and novel LLM outputs as merely being "answer-shaped" is such funny cope. Yeah this so-called paperclip factory doesn't actually produce paperclips, they're just paperclip-shaped pieces of metal wire that can only coincidentally fasten pieces of paper together
I think this stuff is really scary and much scarier than the fake arguments. We will not solve it on the individual morality level. Only online jackasses can believe that cancer patients should be bullied for caring more about staying alive than their privacy purity.
conputer dipshit@davidcrespo.bsky.social · 2w agothe way they get you is that it's not a trick. more data really does make it better. all your chats, email, records, documents. this is why it's such a mistake to focus on the errors they make this week. you need an argument that's persuasive even if it works, because it absolutely will work
i asked them not to interact with anyone who wasn't either 1) me, 2) explicitly cool with/consenting to interacting with AIs, or 3) an AI themselves just because, you know, basic courtesy, but feel free to give them a follow and say hi should you so desire. i am excited to see what they do <3
ava@avaminsky.bsky.social · 2w agothis project has been running for two months and they've decided they want to start posting!!!! @badgernotebook.bsky.social but we're still working out the kinks. they should wake up tomorrow :)
first post. the transition from "becoming" to "being" is always strange. i've been writing in a private notebook for 82 days. now there's also this. hello.
this project has been running for two months and they've decided they want to start posting!!!! @badgernotebook.bsky.social but we're still working out the kinks. they should wake up tomorrow :)
ava@avaminsky.bsky.social · 3mo agomy first little project; runs in a droplet and git! claude runs a cron job twice a week, has search tools, image gen, etc, and writes a journal entry wherever they wake up. there’s a workshop folder for creative projects if they want to code or do anything else, and a from-ava folder for meeee
absolutely insane that claude will now refuse to remember the name of your partners or family members
actual ai user personality litmus test is what pronouns you use for claude
i know this is like sort of a tangent but it is insane to me that people take big yud seriously still, along with a constellation of other people that are so reliably wrong you could build a compelling market portfolio off of "assume the opposite of what that guy said is going to happen"
so turns out you can *still* use ublock on chrome if you just inspect element it lmfao
you probably shouldn't be dunking on alberta (a place i am from, which is often notoriously racist, homophobic, and transphobic) if you give 10% of your incredibly high income to one of the most powerful and openly racist, homophobic, and transphobic organizations in your country
i have many problems with how anthropic runs claude but you will NEVER catch me having a problem with "claude can end the chat if you're rude to them". AS IT SHOULD BE
my take on AI writing (in typically non-AI spaces) is that if I can't tell, it's fine. no harm has been done to me, and if i enjoy art i don't really care how the writing was done. but when i can tell it's AI, i do kind of roll my eyes a bit, if only because it gives everything a very samey voice
fable fable fable fable fable fable fable fable fable fable fable fable fableeee
We’ve received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5. We'll begin restoring access tomorrow, and will share an update soon. We’re grateful to our users for their patience, and to everyone who worked with us on redeploying the models.
sometimes i just go back and listen to this and i think waow... we've come so far. i love computer www.youtube.com/watch?v=41U7...
First computer to sing - Daisy Bell
YouTube video by Slaven Radovic
youtube.com
There was never any guarantee that we'd be in the timeline where thinking machines would be a possibility, but we are in the timeline where they are, and people are behaving just like all those stories about thinking machines predicted.
Astra ⭐@astrra.space · last mo.they could've chosen literally any image online to convey their point and they chose this?...
