The Information Machine

AI Labs Simultaneously Acknowledge Recursive Self-Improvement Threshold · history

Version 19

2026-07-02 08:23 UTC · 340 items

What

In June 2026, Anthropic and OpenAI each disclosed signs of recursive self-improvement in deployed AI systems, and the US government responded with direct interventions in both labs' model access and release schedules. Anthropic's Fable 5 and Mythos 5 were disabled globally on June 13 after a US export control directive targeting SK Telecom's access [11]; the White House fully lifted that restriction on June 30 [16]. The Trump administration separately asked OpenAI to stagger its ChatGPT 5.6 release, now confirmed by at least five major outlets including The Guardian and The Information [20][21][17][18][19]. Both episodes resolved — or are resolving — through private negotiation rather than any public legal process.

Why it matters

The export control's lifting restores immediate access, but the sequence — restrict, negotiate, restore — established that the US government treats frontier AI model access as something it can suspend and reinstate at will, without a stated legal authority or public framework. The same pattern applied to OpenAI's release schedule, and neither intervention has been formally explained.

Open questions

  • What specific grounds did the Trump administration give for asking OpenAI to stagger ChatGPT 5.6 — cybersecurity, geopolitical, or other? [20][21]

  • Does the White House's lifting of the Anthropic restriction restore access for all users, including SK Telecom, or does SKT remain excluded? [16]

  • Legal observers have challenged the government's constitutional authority to withhold AI models from public access [25][26] — has either party formally responded now that the Anthropic restriction has been lifted?

  • Has Anthropic issued a public statement addressing customers affected by the global shutdown, and will it implement customer-level access controls to avoid another blanket disable? [27]

Narrative

In June 2026, Anthropic and OpenAI independently published documents acknowledging that recursive self-improvement may already be underway in deployed AI systems. Anthropic's 'When AI Builds Itself' disclosed that Claude authored more than 80% of Anthropic's production code merged in May 2026, that per-engineer output reached 8x the 2024 baseline, and that Claude Mythos Preview accelerated model-training code approximately 52x [1][2]. OpenAI's concurrent policy blueprint called RSI 'potentially the most consequential frontier safety issue of the coming decade' and cited 'early signs of recursive self-improvement in today's systems' [3]. Both labs publicly endorsed international coordination to slow frontier AI development [4], while each lab's conduct sat in tension with that position: Anthropic's slowdown call was timed to a confidential S-1 filing at roughly $965 billion [5][6]; OpenAI filed confidentially for an IPO while Altman told staff a major RSI breakthrough would favor staying private [7][8].

The most concrete government action came on June 12, when the White House directed Anthropic to block SK Telecom's access to Claude Mythos, citing alleged Chinese ties at the Korean carrier [9][10]. Anthropic, lacking customer-level access controls, disabled Fable 5 and Mythos 5 globally on June 13, affecting all customers [11][12]. SK Telecom, Samsung, and SK Hynix had joined Project Glasswing — an Anthropic program for AI-assisted vulnerability detection using gated versions of Claude Mythos [13][14]. Commerce Secretary Howard Lutnick sent a formal letter to Anthropic's chief compute officer Tom Brown on June 26 [15], and the White House fully lifted the export control on June 30, with Anthropic moving to restore access to both models [16].

Separately, five major outlets — Axios, CNN, The Verge, The Guardian, and The Information — confirmed around June 25-26 that the Trump administration asked OpenAI to stagger or delay its ChatGPT 5.6 release [17][18][19][20][21]. The specific grounds were not publicly disclosed. Taken together, the interventions led observers to describe the US as having moved from safety-oriented model review to direct management of frontier AI access — a 'two-tier system' in which access depends on a user's standing in US foreign policy [22][23]. Commentary after the Anthropic resolution clarified that the restriction appeared to have been government-initiated rather than driven by Amodei's safety convictions, noting that Anthropic sought to restore access rather than endorsing the shutdown [24]. Legal observers argue the government lacks constitutional authority to withhold AI models from public access [25][26], a challenge neither party has formally addressed.

Timeline

  • 2026-05: Claude authors 80%+ of Anthropic's production code; per-engineer output reaches 8x the 2024 baseline; Claude Mythos Preview accelerates model-training code ~52x. [2][45][1]
  • 2026-06-03: OpenAI publishes 'Democratic Governance of Frontier AI,' acknowledging early RSI signs and proposing CAISI as a federal oversight body with mandatory evaluation authority. [43][3][38]
  • 2026-06-04: Anthropic publishes 'When AI Builds Itself,' disclosing Claude's 80%+ code authorship and calling for a global coordinated slowdown in frontier AI development. [28][46][47]
  • 2026-06-05: Critics note the timing of Anthropic's slowdown call relative to its confidential S-1 filing at roughly $965 billion valuation. [5][46][6]
  • 2026-06-08: Altman predicts AI will conduct a significant fraction of OpenAI's research by March 2028; OpenAI confidentially files for an IPO. [32][7]
  • 2026-06-09: All three major labs endorse international coordination to slow frontier AI; NSA reported using Claude Mythos for offensive cyber with approximately six embedded Anthropic engineers. [4][44]
  • 2026-06-10: Jeremy Howard argues Anthropic's continued use of its top-ranked model negates its slowdown call; Altman tells staff a major RSI breakthrough would favor staying private. [37][8][33]
  • 2026-06-11: US government directs CAISI to stop publishing public AI model evaluations; Amodei tells Bloomberg AI progress is in the sharp-acceleration phase of an exponential. [39][29]
  • 2026-06-12: US government issues export control directive requiring Anthropic to block SK Telecom's access to Claude Mythos. [48][11][49]
  • 2026-06-13: Anthropic disables Fable 5 and Mythos 5 globally for all customers, lacking customer-level access controls. [12][11][50]
  • 2026-06-18: Reporting identifies US concern as alleged Chinese ties at SK Telecom; White House directed Anthropic to cut SKT from Claude Mythos. [9][10][51]
  • 2026-06-19: Dario Amodei discloses testers of Anthropic's most powerful unreleased model recommended against releasing it. [30][52]
  • 2026-06-25: Five major outlets — Axios, CNN, The Verge, The Guardian, The Information — confirm Trump administration asked OpenAI to stagger ChatGPT 5.6 release; specific grounds not disclosed. [17][18][19][20][21]
  • 2026-06-26: Bloomberg reports Anthropic and the Trump administration are finalizing a deal; Commerce Secretary Lutnick sends a letter to Anthropic's chief compute officer Tom Brown. [31][15]
  • 2026-06-27: US administration partially lifts restrictions on Anthropic's models; Anthropic begins restoring access to Fable 5 and Mythos 5. [53][54]
  • 2026-06-28: Observers characterize the combined Anthropic and OpenAI government interventions as the US creating a 'two-tier system' of frontier AI access, shifting from safety review to access allocation. [22][23]
  • 2026-06-30: White House fully lifts export control on Anthropic; Fable 5 and Mythos 5 restored to all users. [16]

Perspectives

Anthropic / Dario Amodei

Believes current models may be approaching the RSI threshold; called for a global coordinated slowdown; disclosed that testers of its most powerful unreleased model recommended against releasing it; says AI progress is in the sharp-acceleration phase of an exponential.

Evolution: Pursued and secured a bilateral government deal to restore Fable 5 and Mythos 5 after the June 30 export control lift; public commentary clarified the restriction was government-imposed rather than Anthropic-initiated, distinguishing Amodei's safety rhetoric from the shutdown itself [24].

[28][1][29][30][11][31][15][16]

OpenAI / Sam Altman

Acknowledges RSI as the top frontier safety issue; official blog endorses coordinated slowdown mechanisms; Altman targets AI-conducted research at OpenAI by March 2028 and told staff a major RSI breakthrough might favor staying private.

Evolution: ChatGPT 5.6 staggered at Trump administration request, confirmed by five major outlets — a concrete government intervention in OpenAI's release schedule parallel to the Anthropic case; IPO filing and the private-vs-public tension remain unresolved.

[3][32][8][33][7][20][21][17][18][19]

Google DeepMind

Supports an international organization to enable coordinated slowdowns; researchers published technical work identifying four pathways from AGI to ASI.

Evolution: Consistent with prior reporting.

[4][34]

Jack Clark (Import AI)

Declares 'alignment is not on track' and argues RSI without adequate alignment means rolling very risky dice; treats the Fable/Mythos shutdown as confirming the risks of centralized AI control.

Evolution: The centralization-risk argument gained traction as the suspension resolved via government deal rather than any technical or policy fix.

[35][36]

Jeremy Howard

Argues the only internally consistent form of Anthropic's slowdown call would prohibit the lab with the top-ranked model from using it for frontier research; Anthropic fails this test.

Evolution: Argument stands uncontested.

[37]

Zvi Mowshowitz

Guardedly positive on OpenAI's blueprint but warns federal preemption of state safety laws is its most dangerous element; treats CAISI evaluation suppression as a significant transparency setback.

Evolution: Consistent; concerned governance mechanisms are losing enforcement teeth.

[38][4][39]

Critics and observers

Argue lab slowdown calls are competitive strategy timed to entrench incumbents; characterize the combined Anthropic and OpenAI interventions as the US creating a 'two-tier system' of frontier AI access dependent on US foreign policy standing; legal commentators challenge government constitutional authority to withhold AI models from the public.

Evolution: Post-resolution commentary clarified that Anthropic was not a willing participant in its own shutdown — observers who assumed Amodei proactively advocated for the restrictions revised that view once Anthropic sought restoration through bilateral negotiation [24].

[5][6][40][41][25][26][22][23][24]

Tensions

  • Howard argues Anthropic's continued use of its top-ranked model for frontier research negates its slowdown call; Anthropic has not responded. [37]
  • Altman publicly endorses coordinated slowdown mechanisms while telling staff a major RSI breakthrough would favor staying private, pointing in different directions on what OpenAI actually wants from RSI timing. [42][8][33]
  • OpenAI proposed CAISI as a federal body with public evaluation authority; the US government directed CAISI to stop publishing evaluations, removing the transparency mechanism OpenAI's own proposal depended on. [43][39]
  • The US simultaneously embeds Anthropic engineers at the NSA for offensive cyber and issued export controls forcing Anthropic's most capable models offline globally; the government's concern was geopolitical, not capability-based. [44][11][9]
  • Anthropic frames its slowdown call as a genuine safety response; critics argue it is regulatory strategy timed to a confidential S-1 at roughly $965 billion designed to entrench incumbents. [28][5][6]
  • Legal observers argue the government lacks constitutional authority to withhold AI models from public access; neither the US government nor Anthropic has formally addressed this claim even after the export control was lifted. [25][26][16]

Sources

  1. [1] Today’s edition of my newsletter just went out. — Rohan Paul Twitter (2026-06-05)
  2. [2] 😺 Anthropic: AI Is Building AI now — The Neuron (2026-06-05)
  3. [3] Peter Wildeford🇺🇸🚀 on X: "OPENAI: "We also see early signs of recursive self-improvement in today's systems". RSI is "potentially the most consequential frontier safety issue of the coming decade."" / X — reactive:rsi-governance-moment
  4. [4] Three Labs With a Plan and A Memorandum — Zvi's AI Roundups (2026-06-09)
  5. [5] The company that just confidentially filed its S-1 for a trillion-dollar IPO published a blog post four days later askin... — reactive:rsi-governance-moment (2026-06-05)
  6. [6] Anthropic has discovered the perfect pre-IPO narrative: its product is so powerful that it justifies a trillion-dollar v... — reactive:rsi-governance-moment (2026-06-07)
  7. [7] @skaas777 @iamai_omni 不是终止IPO啦，OpenAI 6月8日刚 confidentially filed for IPO，只是 timing 还没定（他们自己说有些事 private 公司做更方便）。 — reactive:rsi-governance-moment (2026-06-11)
  8. [8] Sam Altman tells OpenAI staff an IPO is planned next year, but recursive self-improvement would favor staying private · Digg — reactive:rsi-governance-moment
  9. [9] Alleged China ties at SK Telecom alarmed US officials and triggered Anthropic crisis — reactive:rsi-governance-moment
  10. [10] In early June, the White House told Anthropic to cut SK Telecom off from Claude Mythos over alleged Chinese ties, Wired ... — reactive:rsi-governance-moment (2026-06-18)
  11. [11] Anthropic to disable its most advanced AI models after US order ... — reactive:rsi-governance-moment
  12. [12] Anthropic disabled its two most advanced models, Fable 5 and Mythos 5, globally for all customers on June 13, 2026, afte... — reactive:rsi-governance-moment (2026-06-15)
  13. [13] SK Telecom Joins Anthropic’s ‘Project Glasswing’ Cybersecurity Initiative; Samsung, SK Hynix Also On Board — BigGo Finance — reactive:rsi-governance-moment
  14. [14] SKT enlists gated Claude Mythos AI to shield core infrastructure — reactive:rsi-governance-moment
  15. [15] US Commerce Secretary Howard Lutnick sent a letter on June 26, 2026, to Anthropic's chief compute officer Tom Brown clea... — reactive:rsi-governance-moment (2026-06-27)
  16. [16] White House lifts export control on Anthropic that froze its most ... — reactive:us-ai-policy-regulation
  17. [17] Trump administration asks OpenAI to limit next model release - Axios — reactive:rsi-governance-moment
  18. [18] White House asks OpenAI to limit its next model release - CNN — reactive:rsi-governance-moment
  19. [19] OpenAI will delay GPT-5.6 after Trump administration request — reactive:rsi-governance-moment
  20. [20] OpenAI staggers AI model release after Trump administration request — reactive:rsi-governance-moment
  21. [21] Trump Administration Asks OpenAI to Stagger Release of New ... — reactive:rsi-governance-moment
  22. [22] The US just invented a two-tier system for who gets to use the most powerful AI - and it happened in two weeks. — reactive:europe-ai-sovereignty-deficit (2026-06-28)
  23. [23] The U.S. has crossed from frontier-model safety review into frontier-model access allocation. That is a different regime... — reactive:rsi-governance-moment (2026-06-27)
  24. [24] Seems like I got Dario wrong. I thought his obsession with AI Safety and Doomism means he told the US govt Mythos and Fa... — reactive:rsi-governance-moment (2026-07-01)
  25. [25] Here are the legal and constitutional problems with the government withholding AI models from the public. — reactive:fable-mythos-export-control (2026-06-26)
  26. [26] Here are the legal and constitutional problems with the government withholding AI models from the public. — reactive:rsi-governance-moment (2026-06-27)
  27. [27] When Anthropic apoliogizes. — reactive:fable-mythos-export-control (2026-06-28)
  28. [28] Anthropic just called for a global way to slow frontier AI because its own models may be approaching recursive self-impr… — Rohan Paul Twitter (2026-06-05)
  29. [29] Dario Amodei's new interview, says AI progress suddenly going crazy. — Rohan Paul Twitter (2026-06-11)
  30. [30] Anthropic's CEO just went on record saying the people who tested their most powerful AI model came back asking them not … — Milk Road AI Twitter (2026-06-19)
  31. [31] Bloomberg vient de lâcher le scoop de la nuit : Anthropic et le gouvernement Trump sont en train de finaliser un accord ... — reactive:rsi-governance-moment (2026-06-26)
  32. [32] Sam Altman's new blog about OpenAI's future path says by March-2028 a significant fraction of its own research will be d… — Rohan Paul Twitter (2026-06-08)
  33. [33] Sam Altman tells OpenAI staff an IPO is planned next year, but recursive self-improvement would favor staying private · Digg — reactive:rsi-governance-moment
  34. [34] Beautiful paper from Google DeepMind. — Rohan Paul Twitter (2026-06-12)
  35. [35] Import AI 460: Reward hacking society, RSI data from Anthropic; and RL-based quadcopter racing — Import AI (2026-06-08)
  36. [36] Import AI 461: "Alignment is not on track"; FrontierCode; and synthetic research interns — Import AI (2026-06-15)
  37. [37] Quoting Jeremy Howard — Simon Willison (2026-06-10)
  38. [38] OpenAI Offers A New Policy Blueprint — Zvi's AI Roundups (2026-06-05)
  39. [39] AI #172: The First Fable — Zvi's AI Roundups (2026-06-11)
  40. [40] 👽 bittensor:native IS EXPLODING BECAUSE CENTRALIZED AI JUST PROVED IT CAN BE SHUT OFF OVERNIGHT — reactive:rsi-governance-moment (2026-06-16)
  41. [41] The real AI infrastructure story this week isn't a new model. It's the split between models you rent and systems you can... — reactive:rsi-governance-moment (2026-06-15)
  42. [42] OpenAI's latest official blog says the world may need a way to coordinate "slowing frontier development when needed." ht… — Rohan Paul Twitter (2026-06-09)
  43. [43] [PDF] Democratic Governance of Frontier AI - OpenAI — reactive:rsi-governance-moment
  44. [44] National Security Presidential Memorandum/NSPM-11 — reactive:rsi-governance-moment
  45. [45] Anthropic just disclosed that Claude now writes more than 80% of the production code it merges. — Rohan Paul Twitter (2026-06-05)
  46. [46] Anthropic calls for global AI slowdown after $965B valuation. Critics claim it's just to hobble competition. — reactive:rsi-governance-moment
  47. [47] Anthropic calls for pause of global AI development — reactive:rsi-governance-moment
  48. [48] In a US government export control directive issued June 12, 2026 ... — reactive:rsi-governance-moment
  49. [49] SK Telecom has been identified as the South Korean company whose inclusion in Anthropic's Project Glasswing programme tr... — reactive:rsi-governance-moment (2026-06-19)
  50. [50] Anthropic disables top-tier AI models after US order limiting ... — reactive:rsi-governance-moment
  51. [51] SK Telecom Cut From Anthropic's Mythos Program Over China Ties | AI Weekly — reactive:rsi-governance-moment
  52. [52] @TannerLeidy @Polymarket Mythos doesn't literally require a gun license—it's a metaphor from early testers. — reactive:rsi-governance-moment (2026-06-17)
  53. [53] 🚨 ANTHROPIC MOVES TO RESTORE ACCESS TO MYTHOS 5 AND REOPEN FABLE 5. — reactive:rsi-governance-moment (2026-06-27)
  54. [54] The U.S. administration partially lifted the restrictions imposed on Anthropic’s models at the end of June 2026, after a... — reactive:rsi-governance-moment (2026-06-27)