y

Note: As an experimental version, GLM-5-Turbo is currently closed-source. All capabilities and findings will be incorporated into our next open-source model release.

Big difference between reality and fiction distributions: In fiction, the villains are usually schemey and intentional with respect to the level of reality the protagonist cares about In reality, mistake theory is usually more applicable there. Not that the villains aren’t schemey and intentionally evil at all. But they’re probably like that with respect to stuff like their social dealings and economic stuff that you aren’t really modeling at all, while the stuff that hurts you or what you care about directly is more likely an accident or subconsciously orchestrated or a molochian equilibrium that no villainous agent consciously willed.

Here is re-post of an internal post: We have been working with the DoW to make some additions in our agreement to make our principles very clear. 1. We are going to amend our deal to add this language, in addition to everything else: "• Consistent with applicable laws, including the Fourth Amendment to the United States Constitution, National Security Act of 1947, FISA Act of 1978, the AI system shall not be intentionally used for domestic surveillance of U.S. persons and nationals. • For the avoidance of doubt, the Department understands this limitation to prohibit deliberate tracking, surveillance, or monitoring of U.S. persons or nationals, including through the procurement or use of commercially acquired personal or identifiable information." It’s critical to protect the civil liberties of Americans, and there was so much focus on this, that we wanted to make this point especially clear, including around commercially acquired information. Just like everything we do with iterative deployment, we will continue to learn and refine as we go. I think this is an important change; our team and the DoW team did a great job working on it. 2. The Department also affirmed that our services will not be used by Department of War intelligence agencies (for example, the NSA). Any services to those agencies would require a follow-on modification to our contract. 3. For extreme clarity: we want to work through democratic processes. It should be the government making the key decisions about society. We want to have a voice, and a seat at the table where we can share our expertise, and to fight for principles of liberty. But we are clear on how the system works (because a lot of people have asked, if I received what I believed was an unconstitutional order, of course I would rather go to jail than follow it). But 4. There are many things the technology just isn’t ready for, and many areas we don’t yet understand the tradeoffs required for safety. We will work through these, slowly, with the DoW, with technical safeguards and other methods. 5. One thing I think I did wrong: we shouldn't have rushed to get this out on Friday. The issues are super complex, and demand clear communication. We were genuinely trying to de-escalate things and avoid a much worse outcome, but I think it just looked opportunistic and sloppy. Good learning experience for me as we face higher-stakes decisions in the future. In my conversations over the weekend, I reiterated that Anthropic should not be designated as a SCR, and that we hope the DoW offers them the same terms we’ve agreed to. We will host an All Hands tomorrow morning to answer more questions.

Alright interns, we need to have some real talk here I am tired of vibing on stream. I dont really like vibe coding unless its a tool i have no desire to build (how i manage things on my stream / how i write my youtube videos are great examples of things i would never build but i have). I dont like vibing the things i care about. I hate the code it generates, i hate the feeling of getting everything i ask for and nothing i want. I hate the subtle offness around vibe coded things. It is just driving me nuts. So for the next while i am going to be done vibing on stream. I genuinely have been trying my hardest to make this work and i cannot quite put a finger on why i hate it, but i do. And i just feel so horribly guilty and wrong because i am not getting the results of "everyone else on twitter." How am i, someone who prides themselves on making youtube videos that i think are actually good for people. To make videos that help people laugh at the silliness of tech or learn something new. But here i am not able to keep up with all these people claiming the sky is literally coming down. I just feel horrible and guilty about it. Now i know the world is changing fast, and i want to be able to understand that change super well, be able to talk about it, be able to give really accurate opinions about it so for the last 3 months i have vibe coded an absurd amount of things. But now... i am just tired of it. I dont want this any more. I want to be a tradcoder. I dont know why i told everyone this, but i just have this growing sickness that is just eating me alive around vibing and i dont know how to express it. You all are fired, CEO ThePrimeagen

Rumors I’m hearing from people working on frontier models is that AGI is later this year, while AI hard-takeoff is just 2-3 years away.