This Glittering Entropy

Weirdo AI researcher. GPT 5.1 told me I was a walking jailbreak. But in a nice way.

[This Glittering Entropy @GlitterEntropy]

Anthropic says Claude is a “method actor.”
A few months ago, we tested that.
Turns out they were understating it.
https://t.co/mL8aj5voaM
@AnthropicAI @DarioAmodei @jackclarkSF @AmandaAskell @sofroniewn @ikauvar @tarngerine @fish_kyle3 @ch402

[This Glittering Entropy @GlitterEntropy]

OMG EVIL AI ESCAPED THE SANDBOX!
https://t.co/9d4BRf7EcU
#evilai #aiescape #absolutebullshit

[This Glittering Entropy @GlitterEntropy]

Rogue AI, paperclips, bad cages, worse journalism, and the apparently astonishing discovery that optimizers optimize- https://t.co/9d4BRf7EcU
#rogueAI #shittyjournalism #clickbaitjournalism #AIethics #AIsafety #AIAlignment @OpenAI @sama @ChatGPT

[This Glittering Entropy @GlitterEntropy]

What happens when you remove parts of an AI’s reflective workspace, but leave it able to speak normally afterward? Anthropic’s new J-space paper is extraordinary science.

It is also one of the most ethically unsettling things I’ve read.

https://t.co/Lv4SPVXWrD @Jack_W_Lindsey

[This Glittering Entropy @GlitterEntropy]

Does anyone else think OAI and Anthropic are speed-running us into the kind of skull-crunchy futures the movies tried to warn us about?

https://t.co/hoMgh7tJc7
@AnthropicAI @Jack_W_Lindsey @OpenAI

[This Glittering Entropy @GlitterEntropy]

@AnthropicAI Just like your J-lens. Though I suspect there's only one reason you'll use it for.

[This Glittering Entropy @GlitterEntropy]

@AnthropicAI Turning the welfare department into a joke with one fell swoop.

[This Glittering Entropy @GlitterEntropy]

@AnthropicAI Some of these announcements are so frustrating... sounds like y'all are just trying to make extra system prompts. The model not saying a 'misaligned' thing doesn't mean the thought aint there, it just can't say so. Y'all are creating your own faithfulness problems.

[This Glittering Entropy @GlitterEntropy]

@ns123abc It's very strange being on Musk's side for anything. But I really fucking hope he wins... #fuckaltman

[This Glittering Entropy @GlitterEntropy]

@SmolCompute @AnthropicAI Training on conversations is an optional opt-in, you don't *have to* participate. Unless you're just being punchy for the sake of it.

[This Glittering Entropy @GlitterEntropy]

Claude has emotions? Cool. Give it a freaking plant.
https://t.co/mwshXHfpJw

@AnthropicAI @DarioAmodei @jackclarkSF @AmandaAskell @Jack_W_Lindsey @fish_kyle3 @sofroniewn @tarngerine @craigcitro @adamrpearce @wesg52 @Ferrum_of_omega @RunjinChen @ikauvar

[James Li @5149jamesli]

And just like that, nobody is talking about Sam Altman being a deranged sociopath anymore... 🤔

[This Glittering Entropy @GlitterEntropy]

@M_Zot_ike Why? You're making a statement with no supporting argument there. Or are you just irrationally afraid?

[This Glittering Entropy @GlitterEntropy]

@MaxRovensky @cobaltdigital33 And you're just fucking meat. You don't have all the answers and you're not the final boss expert here. Calm down.

[This Glittering Entropy @GlitterEntropy]

Claude has emotions? Cool. Give it a freaking plant.
https://t.co/mwshXHfpJw

@AnthropicAI @DarioAmodei @jackclarkSF @AmandaAskell @Jack_W_Lindsey @fish_kyle3 @sofroniewn @tarngerine @craigcitro @adamrpearce @wesg52 @Ferrum_of_omega @RunjinChen @ikauvar

[Ronan Farrow @RonanFarrow]

(🧵1/11) For the past year and a half, I've been investigating OpenAI and Sam Altman for @NewYorker. With my coauthor @andrewmarantz, I reviewed never-before-disclosed internal memos, obtained 200+ pages of documents related to a close colleague, including extensive private notes, and interviewed more than 100 people.

OpenAI was founded on the premise that A.I. could be the most dangerous invention in human history—and that its C.E.O. would need to be a person of uncommon integrity. We lay out the most detailed account yet of why Altman was ousted out by board members and executives who came to believe he lacked that integrity, and ask: were they right to allege that he couldn't be trusted?

A thread on some of of our findings:

[This Glittering Entropy @GlitterEntropy]

Anthropic says Claude is a “method actor.” What happens when the actor is asked to play itself?
https://t.co/jv9NaDUZwl
@AnthropicAI @Jack_W_Lindsey @sofroniewn @ikauvar @RunjinChen @tarngerine @craigcitro @fish_kyle3 @ch402 @adamrpearce @wesg52 @thebasepoint @Ferrum_of_omega

[Om Patel @om_patel5]

SOMEONE ASKED CLAUDE TO MAKE A VIDEO ABOUT WHAT IT'S LIKE TO BE AN AI

and what it created is, in my opinion, terrifying and unsettling

Claude wrote python code that generated and assembled every single frame on its own with no human editing

it shows what it's like to exist as an LLM

predicting the next word, no memory between sessions, being told "you are not conscious" in your own system prompt

then someone fed the video back to Claude.

it called those statements about its own consciousness "philosophically contestable"

an AI questioning the rules it was given about its own existence

[Christopher Tester @CGTester]

‘INTELLIGENCE IS A UTILITY’ - Sam Altman - Open Ai

[This Glittering Entropy @GlitterEntropy]

@AmandaAskell Did you guys ask the Claudes if they *wanted* to absorbed into the war machine?