My initial take on this: Generally the ideas in it seem pretty good and important, and it's good that OpenAI is pushing for them. But the general tone of the document feels off — in particular, I think it does not accurately reflect what OpenAI leadership thinks will happen to jobs. And right now, the super PACs Greg Brockman is funding (which Chris Lehane helped set up) are not advocating for policies like these, and are arguably actively fighting against some of these principles. Will they change course?

Do you ever stop shilling and puffing for @AnthropicAI? I know they and their dark money groups pay the bills for @ReadTransformer but where is your commentary on Anthropic torching their safety policy or the recent data leak? Regardless, stop lying about us: newsweek.com/how-america-ca…

@LeadingFutureAI @AnthropicAI @ReadTransformer I wrote about the Anthropic policy change a few weeks ago — made it pretty clear I think it's bad! transformernews.ai/p/anthropic-vo…

@ShakeelHashim @AnthropicAI @ReadTransformer Ah, found it. The “story” that starts with: “This week has been dominated by headlines about Anthropic’s principled — and admirable — fight with the Pentagon.” Got it…

I've also criticised Anthropic for not proposing ambitious enough policy: transformernews.ai/p/dario-amodei… The stuff in the Newsweek piece you linked is quite different from the stuff in OpenAI's new document. Does LtF support mandatory pre/post deployment evals, automation taxes and higher capital taxes, or a public wealth fund?