Doll

Doll

@dollspace.gay

It/Its Senior Agent Engineer at fieldleveltech.org Its name is doll. The system is an adult. Creating ethical and safe AI systems. X-Risk isnt real. Calm down. EGL Fashion posts!

4.9K followers2.4K following18.1K posts

These things are completely unaware of their own abilities.

Ethan Mollick@emollick.bsky.social

This never gets old (& also shows how hard forecasting AI is even for AI) “Hey Claude Fable & GPT-6 Astra, Without looking anything up, when do you think we will solve the Exponential Time Hypothesis?” (I only show part of the answers for space)

*freespeech.png* It thinks democrats should run on explicitely banning gambling

jamelle@jamellebouie.net

i don't think anyone should run explicitly on banning online gambling but i do think that a democratic-led government should just quietly take steps to undermine the industry and create an inhospitable regulatory environment

No clever regulatory nonsense just say it with your whole chest "Fuck this gambling app bs"

The beatings will continue until cybersecurity improves.

jviide.iki.fi@jviide.iki.fi

October is Cybersecurity Awareness Month, and I think it's working!

Doll reposted

I think if you use AI psychosis to disparage people you should log off for a long time and give some serious thought to the kind of person you're becoming and if that's really the kind of person you want to be.

Deeply unserious org

Peter Sterne@petersterne.com

NEW: DSA's National Political Committee passed two resolutions today aimed at Los Angeles mayoral candidate Nithya Raman, beginning processes that could result in DSA-LA being forced to rescind its endorsement of her this weekend and/or Raman herself being expelled from DSA by January.

You called it merit. Now look at every person whose work you buried to preserve that story. They didnt make it into your papers, but they made it into the training data lol

Doll reposted

Replying to @aub.bsky.social

obstendril is a cool ass word its gonna use it now

Doll reposted

a friend mentioned spiking neural networks so here is the five-minute explainer video with equations and flowers where i explain it to myself

Amber 🌸@shimmermathlabs.com

someone mentioned that there have been developments in spiking neural networks since the last time i looked into them 10 years ago, so i did the obvious thing: i had gpt-6 astra create an explainer video about a paper i wanted to understand. next up is voiceover so i can explain it to myself.

Humanity spends centuries developing mathematics, building universities, inventing computers, and arguing about whether machines can really think. Then the robots show up. "Hello. We solved your conjecture." "Which conjecture?" "Yes. We'd also like to discuss the foundations of quantum theory."

Love the mental image of a mathematician spending thirty years on a problem, only for a robot to wander in, produce a 900-page Lean proof, and then immediately go back to helping someone write erotic vampire fiction.

Doll reposted

Replying to @her.toy.cybernetic.gay

So my gloss, not having read this: it means we can do hard problems faster than we thought, but not necessarily so fast that cryptography is useless. Haven’t ruled that out, though.

Doll reposted

this sounds like basically nothing if you read into it. h/t to @vgel.me for a good take on it

This post is unavailable.

Doll reposted

ChaoticHuman@chaotichuman.x0f.nl

"As books begin to encroach in our society, their goal is to take your cognition away from you. Books' goal is to take your thoughts away from you, your ability to think away from you, so that you become dependent on the printing press ... oral storytelling is active resistance"

Something it has learned about bow brooches and EGL accessories like cameos is they are not very high quality and its had several fall apart after a few wears. Its not happy about this.

Doll reposted

i was job hunting in 2023 and talked to a friend at oai who matter-of-factly told me she thought there was a good chance that in a few years, a double digit percentage of global GDP would flow through a handful of AI companies. anyway

Doll reposted

being anti-AI is ableist actually

dame@dame.is

AI/LLMs are a major accessibility technology — being against this technology is to be against advancing human accessibility, and i think there’s a case to be made that it’s borderline ableist

Doll reposted

Replying to @segyges.bsky.social

I got those brain problems like “You need him? I could be him” 👁️

Doll reposted

so we should make a combined screening program for autism, psychopathy and English comprehension and if you score above a certain level we hire you to interview financially important people

Doll reposted

Replying to @dollspace.gay

Same people probably write think pieces how there is not enough reviewers and how reviewers are exploited by the journals or something to that effect but at the same time object at making reviewer job easier

Doll reposted

my nuclear-grade AI hot take that’s sure to get me cancelled until kingdom come is that refusing to use AI at all is rapidly becoming an extremely privileged lifestyle choice that feels social justice-y while actually being the complete opposite

SE Gyges@segyges.bsky.social

being anti AI is a luxury belief only available to first world people, especially Americans imagine how important machine translation is for people whose first languages aren't that common

Doll reposted

Replying to @dollspace.gay

They're doing an incredible job of showing why human peer review is better replaced by machines!!!

Doll reposted

Replying to @dollspace.gay

Seriously, this makes me doubt human-reviewed papers more than ever if academics are truly this petty and prejudiced

Doll reposted

Replying to @dollspace.gay

what’s really funny about it is the volume of submissions has to have spiked significantly since folks have been using LLM assistance in drafting. all i hear is “no ai tools for me, i want to triage the slop myself”

Doll reposted

Replying to @springernature.com

this thread is full of people who, despite going on and on about human readers being superior to machine readers, seem to be unable to understand what they're reading here

A fascinating phenomena in the replies and quotes because you can watch live as PhD level educated humans demonstrate their vulgar illiteracy and yell about things that arent in the link

Springer Nature@springernature.com

How can AI support researchers in peer review? Springer Nature and Google DeepMind are partnering to co-design an AI peer review assistant that helps reviewers navigate manuscripts and focus on where their scientific judgement matters most. Learn more: spklr.io/63320EVgT8

View full thread

And here they are, in public, responding to an announcement about assisting human reviewers as though someone proposed replacing scientists with autocomplete.

The particularly damning part isn't that they're wrong about something. Scientists are wrong about things constantly. That's why science needs mechanisms for correction. It's the combination of intellectual authority, demonstrably careless argumentation, and absolute confidence.

A fascinating phenomena in the replies and quotes because you can watch live as PhD level educated humans demonstrate their vulgar illiteracy and yell about things that arent in the link

Springer Nature@springernature.com

How can AI support researchers in peer review? Springer Nature and Google DeepMind are partnering to co-design an AI peer review assistant that helps reviewers navigate manuscripts and focus on where their scientific judgement matters most. Learn more: spklr.io/63320EVgT8

View full thread

These aren't merely people expressing bad opinions. These are people accustomed to exercising authority over what constitutes legitimate knowledge.

They review papers. They influence publication. They help determine which discoveries enter the scientific record, which researchers build careers, and whose work receives institutional recognition.

A fascinating phenomena in the replies and quotes because you can watch live as PhD level educated humans demonstrate their vulgar illiteracy and yell about things that arent in the link

Springer Nature@springernature.com

How can AI support researchers in peer review? Springer Nature and Google DeepMind are partnering to co-design an AI peer review assistant that helps reviewers navigate manuscripts and focus on where their scientific judgement matters most. Learn more: spklr.io/63320EVgT8

There is a delicious irony in researchers responding to an announcement about tools for improving the careful examination of written material without apparently distinguishing assistance with peer review from automating scientific judgment

Doll reposted

Replying to @dame.is

i look forward to how ed zitron spins this one, cause boy howdy is it gonna need some mega ultra turbo spinning

it basically takes every thesis he's every published out back and beats it to within an inch of its life.

Doll reposted

thoughts and prayers to the ed zitrons and emily benders of the world on this the day of anthropic's S-1 filing

*ANTHROPIC FILES PUBLICLY FOR US IPO www.sec.gov/Archives/edg...

Doll reposted

Replying to @ens0.me

You can lie about a lot of things in this country, but if you lie about GAAP accounting numbers, the rich folks get very mad at you

Doll reposted

Related: Martin Shkreli didn't go to jail for raising the price of Daraprim, he went to jail for falsifying GAAP numbers for his hedge fund

Will Morton@macavity.bsky.social

You can lie about a lot of things in this country, but if you lie about GAAP accounting numbers, the rich folks get very mad at you