Are there any plausibly workable proposals for regulating or banning dangerous AI research?