Nathan Lambert: open models are 6 months from a US regulatory ban
Interconnects' Nathan Lambert argues an open-weight model will match frontier closed systems — GPT-5.5, Claude Opus 4.8, GLM-5.2 — within six months, and predicts Washington answers not with market competition but an executive order banning or delaying it. He reads Anthropic's distillation campaign as regulatory capture — thin technical evidence behind strong policy asks — that would cement its lead and gut the emerging US open-model economy of inference, fine-tuning, and product firms. His counter: a lab claiming its models are too dangerous to open-source must first secure its own API, which even Mythos's private beta failed to do.
Source: interconnects.ai ↗
If Anthropic's technology is as powerful as they say it is – so powerful that open models like it should likely be banned – then they should be able to secure their API.
Why this matters
- → US may ban open-weight models within 6 months via executive order.
- → Would eliminate emerging US open-model economy: inference, fine-tuning, product firms.
- → Anthropic's distillation campaign functions as regulatory capture, not safety measure.