chair @tablefourthree

@robertwiblin @BogdanIonutCir2 Just let us get to Fable 6 first please and then we can pause progress after 🙏

@HeidyKhlaaf @tristanharris What happens if everyone agrees to stop using the words "going rogue" to describe the Hugging Face attack? How does it change the response to it? What words are we supposed to use? "The OpenAI agents did/went BLANK when they compromised Hugging Face" ?

Introducing Jev Skill Suggestion for Claude Code Now you can keep your skills completely out of the context window until they’re actually needed Set your skills as user invocable only, and the Mod sends the available skill list to Jev For every request, Jev classifies which skill best matches the task, then injects only that skill into Claude’s context Works with both the @typesafeai API and @vercel AI Gateway Install it with one command: npx claude-code-templates@latest --mod productivity/jev-skill-suggestion Full implementation details on the Mod page at aitmpl.com

I find many of your takes hyperbolic but at least usually well sourced, this is a straight up ad hominem attack though. Would exposing the @econliberties team’s personal lives to scrutiny and public scorn be a fair or appropriate way to judge their technical expertise? If not, why is your criticism of METR worth anyone taking seriously?

Anthropic's whole strategy — from day one — was get to the frontier first, then define techniques and set norms for how to handle the frontier safely, creating a "race to the top" You might think it's a bad/dangerous strategy but: (1) it's weird to be confused by it, when they've been so explicit about it for so long (2) they've been undeniably successful so far w.r.t. agents. Claude Code defined how agents should work, and "auto" mode + their work on prompt injection resistance + their interpretability research have set the safety standard

Ryan Petersen @typesfast

22 hours ago

If Anthropic is worried about rogue AI killing everyone why did they give it control of a robotic biology lab?

138 75 1K 97K 51

@alexeheath Zuckerberg mouth piece?

I asked Opus 5 to create a paper-style infographic on the Millennium Prize Problems. @claudeai drew every frame of this animation in JavaScript

I'm "claude project" pilled. A Claude Project can be thought of as a "FFOAO" A Folder For Organized Agent Orchestration And it's very intuitive.... - Name your project - Give it a goal - Chat with an agent - Agent can spin up "threads" - The threads use the tools (your main agent does not) - Each thread can use different models - Each thread can create different artifacts (You can ask it to create a design like design mode, and it will send back a canvas that you can edit) - Each new thread has context over the entire project, so no reexplaining. - The routines (crons) live within the project - The artifacts are stored in "library" within the project Video soon.

keyboard_arrow_left Previous keyboard_arrow_right Next

@timnitGebru this stuff doesn’t work anymore fyi!

Hi Senator! I’m the President of METR. To clarify, METR is pursuing the opposite of censorship: Our goal is to make sure that big companies aren’t suppressing information about AI from the public. This is not a political mission: I’m proud to have worked in the Pentagon during the first Trump admin, and “alignment” at our organization just means “is any human able to steer the model, or is the company going to lose all control of it”. More on who we are and what we do in the tweet below. I think it would be really bad if any one small group could bake a political agenda into these models. I’d love to talk with you and your staff about how we can ensure transparency about what the biggest AI companies are doing so that doesn’t happen. x.com/chrispainteryu…

Chris Painter @ChrisPainterYup

5 days ago

My name is Chris Painter, and I'm the President of METR (Model Evaluation and Threat Research). I know we've made a lot of new friends on the internet the last couple of days, so I thought I'd take this chance to re-up what we do and why. Our work is aimed at making sure that if

462 427 4K 615K 1K

@typesfast To cure cancer. I would like a cure for cancer.

Yeah if OpenAI and Anthropic slow down, Meta will not and will use this asymmetry to their advantage.

Paul Graham @paulg

13 hours ago

If models are getting dangerous, you want to slow down. But you don't want to slow down unilaterally and be left behind. So you ask the government to make your competitors slow down too. And you can't do that with a straight face without asking to be regulated yourself.

31 10 536 45K 26

@daniel_mac8 Is this going to be like Fall/Winter 2025 all over again? Opus 4.5 v2.0 Redux - Opus’s Revenge

Linux is so obviously the future Open source model controlling an open source agent harness working on an open source operating system If you don't see this you're blind

Note that Max Rose who is pushing this left-populist critique of AI safety folks is also on the payroll of Leading The Future, the Marc Andreesen AI accelerationist superpac.

Max Rose @MaxRose4NY

2 days ago

Aren’t the ai companies also the tech elites? If ai is about to kill us all we should stop everything and everyone involved should go to jail If it isn’t about to kill us all- we should federally regulate it. Might be as simple as that

4 3 17 78K 4

crazy to call out jason as "fake news" here when this is the reality is "fake news" nowadays simply "news i dont like"?

keyboard_arrow_left Previous keyboard_arrow_right Next

presh @preshdkumar

a day ago

this is incredible meta goes to all-in summit and spends 30 minutes selling everyone on data centres then @Jason throws a couple fastballs to interrupt the propaganda 🤣

196 373 5K 1.2M 2K

The Muse app continues Meta’s trend of opting users into data collection for AI training. It also nudges you to share your bank account, email, and passport information. wired.com/story/metas-mu…

Please don't make me go back to writing code by hand 😭

Luis Garicano 🇪🇺🇺🇦 @lugaricano

a day ago

Since the Summer, I am obsessing with the image of the battle of Cajarmarca, where the mighty Incas allowed a few hundred Spaniards to take over their empire because Athahualpa thought they might help him in his civil war with his brother. That is what China and the US are doing

3 29 187 17K 127

Fable 5.1 is doing just... absolute circles around Astra (I just got the reset today, it's down to 90% already on medium not fast) Fable is architecting, fixing on the spot, updating Linear tickets, doing real time answering. Astra is chugging on one settings page, stops all the time, spends ungodly amount of tokens (but it's great at browser use) and just... I don't understand this. User error? Skill issue? it should be near AGI, it should be doing... at least the same level? and yet... just, what's going on here folks? Coming to devday, will show folks what I mean if requested but so far, I am quite quite dissapointed.

Didn’t know @Jason had it in him gddamn “The real challenge a lot of us have with Facebook as an organization […] you guys really did damage, significant damage to children. You knew that these products were too addicting and the philosophy of the company from the beginning when @chamath was there ‘Move Fast Break Things’….”

presh @preshdkumar

a day ago

this is incredible meta goes to all-in summit and spends 30 minutes selling everyone on data centres then @Jason throws a couple fastballs to interrupt the propaganda 🤣

196 373 5K 1.2M 2K