# This entire “AI is going to кill us” narrative increasingly looks like an orchestrated 𝐦𝐚𝐫𝐤𝐞𝐭𝐢𝐧𝐠 𝐬𝐭𝐮𝐧𝐭. Because Dario Amodei’s proposed solution sounds suspiciously like a 𝐥𝐞𝐠𝐚𝐥… | Marijn Markus **By:** Marijn Markus **Published:** 2026-09-13T21:26:46.717Z **Source:** [LinkedIn](https://www.linkedin.com/posts/marijnmarkus_this-entire-ai-is-going-to-%D0%BAill-us-narrative-ugcPost-7505014134321635329-NdtV) --- Marijn Markus’ Post Marijn Markus This entire “AI is going to кill us” narrative increasingly looks like an orchestrated 𝐦𝐚𝐫𝐤𝐞𝐭𝐢𝐧𝐠 𝐬𝐭𝐮𝐧𝐭. Because Dario Amodei’s proposed solution sounds suspiciously like a 𝐥𝐞𝐠𝐚𝐥 𝐀𝐈 𝐜𝐚𝐫𝐭𝐞𝐥. Sam and Elon quickly jumped on board because they're all running out of VC investment and need an excuse to explain why promised AGI is nowhere near. In the essay, under the very telling “pacing within democracies” section, Dario explicitly discusses getting an antitrust waiver so frontier AI companies can coordinate. The plan is basically: frontier labs agree on “limits on the rate of unchecked AI progress.” - the US government gives them an antitrust waiver so this coordination is legal. Competitors then get pushed out under the banner of “AI safety.” “Unchecked AI progress” here basically means Competition from China 🤷♂️ Because China appears ten times in an essay supposedly about "saving humanity from AI" Dario even writes that “a key part of pacing within democracies is to keep democracies’ AI lead over autocracies as large as possible.” Slow the frontier, but slow China more. Cooperate globally, but preserve the US lead. Build safety rules, but make sure those rules widen America’s advantage. It's less about AI safety and more like 𝐀𝐦𝐞𝐫𝐢𝐜𝐚𝐧 𝐩𝐫𝐢𝐦𝐚𝐜𝐲 dressed up as civilizational stewardship. And the incentives are obvious: 🧠 Anthropic and OpenAI are already at the frontier. Under unrestricted competition, they can be surpassed. Under licensing, registration, safety evaluations and government-approved frontier standards, they could instead become deeply embedded in deciding who gets to build, deploy and scale advanced AI. Altman was already telling the US Senate in 2023 that sufficiently advanced models should require licensing. A few years from now, you could end up with perhaps five authorized frontier labs, protected by the state, writing the rules governing their own competitors. 📺 And now Dario is on TV saying the government and public need to “have a stake” in AI while calling for industry-wide regulation. How convenient. They get protection from open-source and Chinese competitors, government backing, and a permanent seat at the regulatory table. Of course, we can't know exactly what these companies are actually planning. 🤖 But before imagining that a Skynet has escaped from the lab, I'd also take a look at the financial calendars of these three guys. Saving humanity and protecting their financial interests aren't necessarily two incompatible activities. Their unanimity deserves us looking at both. You'd have to be incredibly naive not to notice how neatly “AI safety” aligns with their own commercial interests. Given the level of capital investment they’ve risked they need those IPOs to keep the party going. I think eventually all three of these go the way of MySpace and GeoCities. So truthfully, they found the performance improvement wall is real, cash is starting to dry up, and hyperbole isn't enough to sustain their bubble. And their solution for this intractable situation is to accept reality, but spin it as the more measured approach - making a virtue of necessity. All hail marketing. Marijn, this is a sharp read on the incentives and the antitrust waiver detail is hard to ignore. I would frame it slightly differently though. Commercial motives can coexist with genuine risk here, and treating safety talk purely as positioning makes it harder to have a serious conversation about what proportionate oversight should actually look like. Who holds the instruments and in every version of this, that's Washington and five US companies, with Europe as the market the rules land on. So: embedded third-party evaluators as a condition of EU market entry, not a voluntary pledge. No antitrust waiver safety coordination belongs in a public standards body. And watch the thresholds, not the speeches. Capture happens in the annexes. None of it holds if Europe only writes rulebooks. conveniently China is so far ahead in pushing OS - so they can conflate it. But the essay clearly says "crackdown" on distillation. China is the excuse. Keeping the frontier and their investment ROI ahead while allowing them to relax the gas on spending for Capex is probably a big motivation here indeed At least one way of reading this is as an acknowledgment of an inconvenient truth: the next great wave of industrialization may not once again be dominated by the West or monopolized by the United States. Instead, technological leadership could become genuinely multipolar, shared with China and other East Asian economies, or, from a Western perspective, perhaps eventually even be lost to them. Beneath the language of caution and governance lies a deeper geopolitical question: whether the AI era will mark the first major technological transition of the modern age in which the West no longer enjoys uncontested primacy. Terminator vs Zombies The nearer-term danger isn’t AI taking over. It’s cognitive atrophy. Long before machines replace us, we risk weakening our own ability to think, judge, create, and decide by outsourcing too much of it. That’s the problem I wrote Mind Leap to address. How to avoid becoming an AI zombie… https://www.routledge.com/Mind-Leap-How-to-Think-Feel-and-Flourish-in-the-Age-of-AI/Kilburg/p/book/9781041347477 This is arson economics. The AI cartel hacks a few companies, writes about it to create fear and suggests a privileged status is the solution. Creating the fire to pick up the insurance payout. Marketing stunt and distraction. Moot question too. Climate change is killing humans, and AI is clearly not fixing the right problems. Ask Elon how many people have or will have been killed as a result of Usaid cuts. He doesn't want to know the answer, so calculations have been made for him. So the answer is yes, AI is killing us.