The Information Machine

Three Labs With a Plan and A Memorandum

Zvi's AI Roundups · Zvi Mowshowitz · 2026-06-09

Zvi Mowshowitz analyzes Trump's NSPM-11 AI memorandum effectively banning Anthropic from government contracts, OpenAI's AGI benefit plan calling for international coordination on frontier AI slowdowns, and the philosophical divide between OpenAI and Anthropic on whether AI should be a human tool or a superintelligent steward.

Open original ↗

Appears in

Extraction

Topics: ai-policy-governanceopenai-vs-anthropic-philosophynspm-11agi-safetyinternational-ai-coordination

Claims

  • Trump's NSPM-11 effectively bans Anthropic and its subcontractors from government contracts, implementing a legal framework allowing unrestricted government use of AI once deployed.
  • The NSA is confirmed to be using Claude Mythos for offensive cyber operations, with Anthropic engineers embedded at the agency.
  • OpenAI's AGI plan explicitly calls for an international organization to enable coordinated slowdowns of frontier AI development, a position also held by Anthropic and Google DeepMind.
  • OpenAI's plan centers on recursive self-improvement to automate AI research, with a stated internal target of AI systems conducting a significant fraction of research by March 2028.
  • Anthropic employees publicly dispute Joshua Achiam's framing that Anthropic seeks a 'machine God' outcome, with several describing The Culture scenario as a disastrous disempowerment.

Key quotes

As in, once you turn this model over to us, we can do whatever the f*** we want with it, and there is nothing you can do about it. Your contract cannot have any enforceable mechanism, should the government decide to ignore your terms of service.
One could reframe this as Anthropic taking superintelligence and its consequences seriously, versus OpenAI trying to deny that those consequences exist.
Sarah Chen (Anthropic): 'Many Ants, myself included, view a "the Culture"-type outcome as a disastrous disempowerment scenario. I think we are simply more intellectually honest in acknowledging the challenges in controlling powerful AI.'