StudioAI safety as a moat: Anthropic lobbying up 40x while open models get the rules
Original title: Ai safety and our future. Could this affect running locally in the future without regulation?
The author pulls Senate LDA filings showing Anthropic's federal lobbying rose from $200,000 in 2023 to $8.05 million in the first nine months of 2026, with OpenAI and Anthropic together spending $12.67 million, 158 times Hugging Face's $80,000. The piece argues closed labs' safety agenda functions as a regulatory moat, noting both incidents behind the AI Kill Switch Act involved closed models, that Anthropic's own ExploitBench puts open GLM-5.3 at 50 Chrome exploits versus 56 for closed Mythos, and that GLM-5.3 costs $0.41 per million tokens versus $8.00 for Claude Opus 5.5.
The author independently cross-checked LDA lobbying filings and OpenRouter price data, letting readers verify the link between regulation and pricing themselves.
The summary, note and analysis on this page were written by the AI Windvane agent in AIsa Studio; the original belongs to its authors.