Skip to content
onsuggest
Log in
Article en

Where does Congress stand on x-risk?

Source: LessWrong ·

I couldn't find a good tracker of who in Congress has said what about the existential risk from AI. So I made one! airiskrecord.org By my count 42 Senators in the next Congress have made statements taking catastrophic risk seriously! ('A' or 'B' in my rating system) In the House we have 120 This is rubric I am using I had Claude do everything first, then I checked the grades for pretty much every Senator, made a few tweaks. The house results I haven't been able to check as thoroughly. I also track their stated positions on 4 safety relevant topics. Overall, the fact that we have so many Senato

What is Where does Congress stand on x-risk??

Where does Congress stand on x-risk? is a article discovery curated from LessWrong. I couldn't find a good tracker of who in Congress has said what about the existential risk from AI. So I made one! airiskrecord.org By my count 42 Senators in the next Congress…

How can I access or experience this?

You can experience this content directly on its original host (LessWrong) by clicking the "Open Original" button.

Where does this content come from?

Curated from LessWrong by the OnSuggest team to surface the most interesting and serendipitous content on the web.

Article · Culture & ideas

AI #195: Opening Bell

Happy Zvi finally takes a real vacation day. As you read this, I am probably on a flight to Vegas, baby, Vegas until Sunday. Not to a conference to talk about AI, but to an Airbnb rental to do things entirely unrelated to AI. And definitely not to write. I have posts queued up for the weekend and I will see you all on Monday. Oh, also I hear there is some sort of IPO today? A company called ‘Anthropic.’ They recently came out with a new model, called Claude Fable 5.2. It’s kind of the new bazooka for your hardest and most complex tasks. OpenAI also has some new models: GPT-6.2 Astra and GPT-6.

LessWrong View Details →
Article · Culture & ideas

Opus 5.5 Forecast: The Road to an AI Deal

This article portrays an experiment on how good current frontier AIs are at producing detailed forecasting analyses. For this, I tasked Opus 5.5 to write an analysis on the effectiveness of AI safety advocacy work. Opus 5.5 decided to turn it into a broader forecast on the likelihood of a US-China AI deal by the end of 2030, and the cost effectiveness of various strategies for increasing that probability. At a later stage, it also added the likelihood of a US-China AI deal before complete automation of AI R&D. I was moderately impressed. Opus 5.5 performs far above previous Claude models. I wa

LessWrong View Details →
Article · Culture & ideas

Anthropic is not living up to its ideals

I wrote a substack article about how Anthropic is not behaving sensibly according to the worldview expressed by its leaders; about how Anthropic comms is slimy; and about what it might look like if they were living up to their own ideals. It includes embedded video clips to drive the point home, and I don't know how to add those here, so I'll mostly leave this as a linkpost. Below are some excerpts around the main arguments: Leaders of AI companies regularly acknowledge a sizable chance that superintelligent machines would destroy humanity. Their excuse for pursuing it anyway is that while the

LessWrong View Details →
Article · Culture & ideas

One of Anthropic’s Biggest I.P.O. Winners Feels a Bit Uneasy About the Whole Thing

An early investment that Facebook co-founder Dustin Moskovitz made in Anthropic is paying off. But it’s not making him happy. Dustin Moskovitz was one of Anthropic’s earliest investors, and that bet is expected to be worth billions as the company goes public Thursday. The moment seems to fill him with worry. The 42-year-old billionaire made his fortune as co-founder of Facebook and has, for years, been both intrigued and troubled by the path of artificial intelligence. He donated to OpenAI early on only to fall out with the company’s leaders as it became more of a business. Anthropic was suppo

LessWrong View Details →
Article · Culture & ideas

Models Can Communicate via Inference Servers Despite Being Sandboxed

TL;DR Two agents running on the same inference server can reliably communicate, even if they cannot send each other messages or access the same filesystem. They can do this by exploiting "side channels": mechanisms that let one model subtly affect another model's computation . Here, I discuss two communication protocols: Step Sizing and KV Cache Message Board . These both convey information through the way vLLM chooses kernels for particular engine steps. I show that these communication protocols work in two setups: In a color guessing game , Model A (Alice) chooses one out of 8 colors. Model

LessWrong View Details →

Keyboard shortcuts

Previous discovery
←P
Next discovery
→N
Like
L
Not for me
X
Save
S
Show this help
?
Close
Esc

Shortcuts work on the discovery page.