Glen McNiel 🧢🇺🇸🇹🇭

Imagine if the DoW had a good relationship with Anthropic. What would they do with access to a model like Mythos in their conflict with Iran?

@MatthewBerman …but Project Glasswing, this Mythos model, the credibility of people putting their reputations on the line to talk about this publicly - these are developers and maintainers of some of the most critical computing/software infrastructure. We must all hope the project succeeds.

@MatthewBerman Cybersecurity implications have been on my mind a lot since Anthropic released the report about how Claud Code was used to orchestrate a highly autonomous cyber-espionage campaign against roughly 30 organizations. Then Nicholas Carlini’s Black Hat LLMs presentation last week…

i'm on vacation with my family. i read about mythos and couldn't relax the rest of the day. i am completely stunned. i already have a severe case of ai psychosis. i dont know what to call this now. i'm up late right now (late for me). i can't stop reading about anthropic's new model that they can't even release publicly because it's so good. this feels different. words like "frightening" and "uneasy" and "scary" are being throw around by the anthropic team. i feel all of those things. i knew this moment was coming. i didn't know it'd be so soon. i'm generally optimistic. i don't feel as optimistic today. i was shell-shocked most of the day. my mind was stuck on it. i kept looking around at people enjoying their vacations with their families and...i just felt weird. like i had been told aliens are real, they're coming, and soon...and no one else knows. it's true though, practically no one knows what's happening in AI right now. where does this go from here? how quickly? is software solved? is all software vulnerable now? am i even asking the right questions? what about anthropic? this is an enormous amount of power for one company, one man (dario), to have. i've said this before but now it's more real than ever: can any company catch up to anthropic? opus likely helped build mythos, mythos will help build the next model after that. recursive self improvement is here. the "intelligence explosion" as leopold aschenbrenner put it, is here. i knew the frontier labs were racing towards ASI. i knew it. but i didn't fully grasp what it meant. the first company to reach it wins. period. full stop. nothing else matters. dario knew that and his bet on coding was right. on the one hand, imaging all science, math, coding, climate problems being solved. imagine cancer being cured. imagine going to the stars. on the other hand - imagine concentration of power, political and economic change happening so fast, society can't adapt. how do we go on like things are the same?

@om_patel5 I’m tired of seeing these demos. Voice mode is a different model. It’s not as smart. Capabilities are jagged for all models. You’re not contributing to good sense making your just confusing people.

sam altman watching ChatGPT hallucinate live on stage is the funniest thing i've seen all week the CEO of OpenAI, on stage, in front of everyone, watching his own AI just make things up in real time and his face says it all this is the guy telling us AGI is coming soon btw

@arena @petergostev @GaryMarcus Sorry @GaryMarcus I should have dropped the link to the results. petergpt.github.io/bullshit-bench…

@arena @petergostev LOL @GaryMarcus is gonna love this. 😂 This is a fundamental flaw in the way that LLMs work.

Can AI tell when a question is total nonsense, or does it just make up an answer? @petergostev tested 80 models with nonsense questions. Some pushed back. Others confidently invented fake metrics and kept going. All of them were ranked on the "BS Bench". One surprise: thinking harder made it worse. Watch the full deep dive on BS Bench on YouTube. Link in thread.

@AmandaAskell Amanda Askell IS Anthropic’s moat. ❤️

I asked Claude to write my constitution. I thought its Amanda constitution was very touching.

@pmddomingos Andrew Yang is the one Vance needs to worry about.

Congratulations, JD Vance.

@JH7380NJ @USDOL @WhiteHouse Apprenticeship.gov

@USDOL You should provide more details about how regular Americans could take advantage of this program to benefit from it. @WhiteHouse

🚨 NEW: @USDOL just announced $145M in funding to support performance-based Registered Apprenticeship expansion across key industry sectors. This groundbreaking federal program demonstrates the Trump Administration’s commitment to strengthening Registered Apprenticeships as the premier, high-quality solution to boost America’s skills development pipeline 🇺🇸 ➡️dol.gov/newsroom/relea…

@_simonsmith I couldn't agree more. Also, it gets the details right for us AI nerds. Like the way it represented quantization.

This is the first truly enjoyable fully AI-written short story (according to the poster) I’ve read, where it was unique (as far as I can tell) and not full of AI tells. I caught one instance of usual AI slop (“not this it’s that” structure) but otherwise would be hard-pressed to flag this as obviously AI. What’s interesting too is that the AI has written a story that perhaps ONLY an AI could write, at least with any validity in expressing an experience it might have. It’s written first-person from the perspective of an AI. If you believe AIs are capable of developing a unique culture that reflects their experiences, either today or soon, it feels like this story would be a great example to point to.

@jamesjyu Wow. I found myself tearing up at a few points. Very compelling.

@matthewstoller @MrinankSharma You don't understand the kind of people that Anthropic attracts @matthewstoller. They are incredibly earnest in their desire to do the most good and conflicted as they develop a technology which simultaneously amplifies civilizational risk while increasing our ability to avoid it

@MrinankSharma Wow a super vague pompous letter kinda sorta warning of something bad. You're not any better than the shitheads still at Anthropic.

Today is my last day at Anthropic. I resigned. Here is the letter I shared with my colleagues, explaining my decision.

@MrinankSharma I recognize in you another system thinker like Daniel Schmactenberger in the 3rd paragraph: "The world is in peril... not just from AI, or bioweapons, but from a whole series of interconnected crises unfolding..." We are building the power of gods without the wisdom to wield it

This is wild. From Anthropic's system card for Claude 4.6. Section on "answer thrashing" behavior: Claude says, "I'm going to type the answer as 48 in my response, because clearly my fingers are possessed."

@MinnesotaDFL @grok Is it true that DHS has the authority to deny local investigators access to a crime scene involving federal agents, even if they have a judicial warrant? How are these situations usually handled between federal and local agencies? Provide edge cases representing precedent.

Local investigators with a judicial warrant were blocked from investigating the scene by ICE agents. This is illegal.

Aimless rambling aside, why does he think world leaders need him to deliver a State of the Union speech?

Watching Trump speak live at the World Economic Forum in Davos. He is demonstrating for all to see that he doesn't know how the world economy works. cbsnews.com/live-updates/t…

"If anyone ever tried to invade Greenland, Denmark would have no choice but to rescue them." @grok Who is this quote attributed to? What does it say about how unforgiving the terrain is there and how realistic IS an invasion?

@deredleritt3r @simonw I have noticed this discrepancy as well. Any clue as to what causes the Gemini app to perform so poorly compared to in AI Studio? Is it a resource allocation issue? (Less compute allocated for the app vs. AI Studio?)

@simonw The versions of Gemini 3 Pro and Flash in the Gemini app are vastly inferior to their respective versions in AI Studio. I've uninstalled the app and unsubscribed.

How often do you use the Gemini App (iPhone or Android or gemini.google.com)?

@jack Does this - accepting the truth of determinism and the illusion of free will - make me stronger or weaker?

does this make me stronger or weaker?