Anthropic's Safety Nonprofits Are a Self-Amplifying Regulatory Capture Machine

Anthropic is in regulatory-capture financial loop

Kevin Bass argues that AI safety nonprofits aligned with Anthropic — from Tarbell Fellows in media to METR in evaluations — all depend on Moskovitz's $7 billion in Anthropic stock, which grows as Anthropic grows. Though they claim independence, their funding lineage ties them to the lab. Bass contends this creates a self-amplifying loop: Anthropic's success funds more AI Doom messaging, which drives bigger IPOs and regulatory capture, with Anthropic positioned as both problem and solution.

Anthropic promotes the problem -- AI Doom -- and promotes itself as the solution -- AI Safety -- which means control of regulation by Anthropic.
  1. kazinator

    I can't help but be reminded of how Edison scaremongered around electrocution: the War of the Currents.

    Edison was a proponent of distribution of DC, and so he used scare tactics to try to convince people that the competing AC, which could be stepped to high voltages for long-distance transmission, was dangerous. He staged public electrocutions of animals to spread fear.

    This is much the same thing: AI is dangerous, except our brand of AI.

  2. Havoc

    I guess with xcancel being down and twitter website being literally unusable I guess I’ll just not know what this post is about

  3. ElProlactin

    Whether or not Dario, Sam and Elon really believe the worst-case scenarios are realistic, it's obvious that they'd like to form a cartel that they think would benefit their companies.

    But this guy looks...

    https://case.kevinnbass.com/

    https://x.com/kevinnbass/status/2098493154781446283

    https://x.com/kevinnbass/status/2098293338646421691

More from this day

2026-09-15