Transcription: OpenAgents Episode 124 - Magic AI = Deep State

Source: https://x.com/OpenAgentsInc/status/1829238237925458015 Wiki source: https://raw.githubusercontent.com/wiki/OpenAgentsInc/openagents/Video-Series.md Media title: OpenAgents - Episode 124: Magic AI = Deep State AI coding startup @magicailabs ra... Upload date: 20240829 Transcription model: gpt-4o-mini-transcribe Generated at: 2026-06-01T16:45:20Z

Machine-generated transcript. Review speaker labels and wording before using this as quote-grade source material.

[00:00] Speaker: I was thinking of making a video just to make fun of this company, Magic, also building an AI coder, because they raised $110 million a long time ago, didn't ship anything, just raised another $300 million, and they say that their goal is to, imagine if you could spend $100 in 10 minutes on one task and reliably get a great pull request for an entire feature. And yeah, we shipped that two days ago, two PRs in 12 minutes for $5, and we didn't need to raise $465 million. What are they doing? If we've already shipped what they're trying to build, what are they doing? I read the blog post closer. Oh, you can see at the bottom, sufficiently advanced AI should be treated with the same sensitivity as the nuclear industry. Where have I heard this talking point before? In addition to our commitment to standard safety tests, we want Magic to be great at cybersecurity and push for higher regulatory standards. And they say something in here about, oh, what we're doing is building on no open source foundations. Okay, these people are not advancing open AI at all. And where was it also? One of these people associated with Magic, it might've been the founder, even used the phrase safe superintelligence. Which, you know, if you've been following the Open Agents Inc. Twitter account, is also funded by the same people funding Leopold, who is called for nationalizing AI labs. So, it is clear that Magic is on team closed AI, regulatory capture, should have been obvious based on their investors, who are the same people funding Leopold's VC fund, who's loudly called to nationalize the AI labs. And these people now have one big lab, SSI, with Ilya, based in Israel. And they now have one big lab, Magic, based in the U.S., talking about state superintelligence, funded by who? Eric Schmidt, AKA the U.S. deep state, and the common actors behind Leopold and SSI, and now, and Magic's initial round, Nat Friedman and Daniel Gross. Why are these people funding regulatory capture and their goal, stated goal, of people that they're funding, is to nationalize AI labs? So I'm gonna repeat the same message that I gave in episode 101 of this series. We are not locking down AI labs. We're gonna open AI labs. We're gonna open new ones, open source, open science, truly open AI. And if you wanna stop us, you can mow on fucking Lab A. Look at this. Look at this on the side. Not even this topic, a different topic, but related. OpenAI Anthropic AI safety pact with the U.S. Guys, they're coming after OpenAI, and we're not gonna let them. See you soon.